Denis Dyack, known for his work on games including Blood Omen: Legacy of Kain and Eternal Darkness: Sanity's Requiem, says big game developers are "abusing early access" by using it as a marketing tool for highly polished games that are ready for full release anyway. That, Dyack believes, has shifted expectations for early access releases overall, and made life more difficult for indie developers who rely on the process to actually get their games finished and launched.

"They can put up a polished product and rake in cash," Dyack told GamesIndustry. "It's a marketing beat for them. So, that leaves players often expecting a near-perfect game that, in our opinion, defeats the purpose of early access, which is a collaborative development process based on player feedback.

"AAA publishers are abusing early access, skewing player expectations in such a way that it heavily disadvantages smaller independent developers."

Dyack said games like Subnautica 2 and Path of Exile are "drastically changing the landscape and players' expectations," which in turn has led to "significant barriers to entry" for less polished games like Dyack's current project, Deadhaus Sonata, which is now in early access on Steam.

Ethan McKinnon, who joined Deadhaus Sonata developer Apocalypse Studios as creative director after the game's rough early access launch earlier this year, added that "players naturally compare everything in the same marketplace, so smaller teams are being held to a much higher standard than they were in the past."

Interestingly, Subnautica 2 creative director Anthony Gallegos seemed to agree with at least part of Dyack's proposition, saying in the same report that early access is "not a few edge-case games participating anymore."

"Early access could mean a game like ours doing long-term open development, or it could be a game three months from a 1.0 release that's effectively doing a paid beta soft launch and using the Early Access term as a way to soften player expectations," Gallegos said. The result, he continued, is a "more diverse and confusing landscape."

I think that point is more salient here than any perceived 'abuse' by big publishers: It's tough to set meaningful broad expectations for "early access" when it's such a loosely defined term in the first place. But even though I'm not entirely convinced by Dyack's argument, he and McKinnon do have sensible advice for developers looking at early access launches: Ensure the game runs well, pay attention to Steam reviews, and connect with influencers who can get it in front of eyeballs.

Dyack admitted that Apocalypse Studios missed those points with the early access release of Deadhaus Sonata earlier this year. "We failed to prioritize the basics right from the get-go," Dyack said. "Also, we didn't handle Unreal Engine properly at the time and didn't do a good job explaining what made our game cool and different. In a lot of ways, we were too focused on what made our game so different and unique that it took our attention away from making sure we got the basics done right."

McKinnon put it more simply: "Developers need to treat early access like a real launch, because players do."

AI's a touchy subject in the development industry, but in other areas it's a lot less controversial—take Japan, for instance, which is embracing the technology a lot more than audiences in the West are. That's not to say there isn't any push back at all, mind.

Still, an example of how there's more of a positive spin on it overseas comes courtesy of a Famitsu interview (translated here by Automaton) where seasoned JRPG devs from Square Enix seem broadly positive on the tech—and are repeating some oft-heard talking points that I, at least, am personally very bored of, even if I wouldn't dismiss them out of hand.

Hiromichi Tanaka, who worked on many Final Fantasy games during his tenure (which ended in 2012), says that the creations of the future "will definitely come down to how well people can master the use of AI." Yoshinori Kitase (also a veteran, and current producer on Final Fantasy 7: Revelation) concurs, saying: "major game companies will need to rethink their development methods to avoid being swept away by this wave."

Kitase points out that we've already had an increased wave of indie games from the popularisation of accessible game engines, and "when you add support from new tools and technologies on top of that, even people with limited experience can more easily make games."

We have seen a touch of that in all of these decompiled games, but there's also the fact that people with limited experience are also able to just up and steal ideas easier, too. And I'm not talking about a corporate aversion to piracy, here. The developers of Moonlighter recently had to yank their ReVamp demo from Steam because it was getting cloned—that's an indie dev having its hard work copy-pasted.

Kitase does at least acknowledge that people don't tend to find purely AI-generated or cloned slop tasteful, though: "Humans need to be involved in the initial step of turning zero into one," a turn of phrase he uses to describe the human creative spark and original ideas. "That will never change. The next generation of creators absolutely needs the ability to turn zero into one."

Tanaka adds that this'll turn the industry into a "competition of ideas". Which is a nice fantasy on paper, though I'm not entirely sure it's panning out that way in practice—as mentioned above, indie developers, already on rocky ground in an unkind industry where discoverability is a nightmare, are at risk of either having their games cloned or drowned in a sea of AI slop.

Well, here's a headscratcher. Paolo Pedercini, developer of Future? No Thanks! recently admitted to using generative AI in his ecosocialist satire while also referring to the technology as "a technofascist, anti-human and anti-art project." Now, the developer has removed several AI generated voice lines from the game, following a backlash to his use of the very technology which the game lampoons.

You might remember Future? No Thanks! from another recent controversy, whereupon the game was briefly delayed after Steam held up its release while it reviewed some fleeting nudity in the game. Pedercini—who also goes by Mollieindustria—received considerable support during that debacle. But the reaction to Future? No Thanks' AI use has been rather less supportive, prompting Pedercini to rework the game and respond to the backlash.

Pedercini originally revealed his AI use in a blog post (via Rock, Paper, Shotgun). Apparently, the game contained several voice lines generated using text-to-speech, which Pedercini claimed, "used to be an uncontroversial technology and an early application of neural networks" before "AI bros started commercializing AI voices". Alongside this, the game also includes "a tiny fraction" of AI generated code, which Pedercini said he used for things like "fixing a shader I didn't write which broke after an update" as well as "finishing touches and minor debugging."

In this blog, Pedercini defended his AI use by pointing out that Future? No Thanks! "satirizes generative AI and the careless ways it can be used," adding that the "100% no-AI stance" adopted by many game developers "is very recent, and the game was being developed before a consensus emerged." He also cited his role as an academic and game design teacher, stating "I have to look into these tools which most of my students use, and I was trying to figure out if there are non-brain-rotting ways to do so."

Yet it seems these arguments failed to convince Pedercini's audience, as yesterday he published another blog post revealing he has now removed the offending TTS voice lines, replacing them with lines by human actors. Apparently, these lines were "recorded earlier this year," but Pedercini opted not to use them because he "thought the actual uncanny text-to-speech was funnier in that context."

Although Pedercini has removed the lines, he's clearly not happy about it. In the second blog, he stated that he believes that "the game is worse not, but only, like, 0.001% worse." He also included a "personal update" through which he responded more broadly to the backlash.

"Somebody from Bluesky (the Twitter/X clone) dragged me due to a post disclosing what I foolishly thought a minor use of generative-AI in Future? No Thanks!" he explained. "I gave versions of that statement multiple times before, whenever asked by festivals and events." He also pointed out that, per Valve's updated AI disclosure rules, "Steam doesn't require disclosure of generative AI for things like code assistance, so many people assume their favorite indie games don't use AI at all just because the developers stay mum".

The post then took a weirder turn. Pedercini admitted that his original blog "was not a good post," and that he thought "nobody would read it", before claiming that his AI disclosure resulted in "more insults than the time at the height of GamerGate when Adam Baldwin unleashed his half million followers on me." He followed this by linking to a list of game developers that reject generative AI, claiming that "all studios" who haven't disclosed genAI use are "using it for coding in some capacity" and began his signoff with "I drank from the chalice of blood. Nothing will ever redeem me and my cursed little game."

As for what to make of all this (at least, the bits that aren't about drinking blood), I can sort-of understand Pedercini attempt at deploying TTS voice lines for satirical purposes, and he's correct in saying that TTS has a history that goes back further than modern LLMs. If that was all the genAI in the game, it would be more defensible.

But the fact he's also used generative coding tools for clearly non-satirical purposes is much harder to defend. Also, if you're making a game that satirises generative AI, the primary audience for which is going to be people who are avowedly against generative AI, it's probably prudent to ensure that the creation of said game is as far removed from generative AI as possible.

Yet among the references to gamergate and accusing the entire industry of secretly using AI, Pedercini did make two other points worth noting. First, he said that players who bought Future? No Thanks and "feel betrayed" by his use of AI can request a refund by emailing their Paypal or Venmo information to him. Second, he promised that his next project "will be completely hand-coded", as he had already determined that "AI-assisted programming is bad for me."

The developer of an upcoming horror drinking game has cancelled plans for a public Steam demo in favour of private preview sessions, citing concerns that AI vibe coders might use the demo's content to clone the game.

Drunken Goddess Reflux is a "shot glass Russian roulette" drinking game where players square off against an anime demon girl who loves you but also kind of wants to murder you (which sounds like a standard relationship to me), with the goal of getting your girl-frenemy so blotto that she barfs up her soul. Frankly, it doesn't sound like my cup of tequila, although there is a mechanic where the rules of the game grow weirder and wilder the more alcohol you imbibe—which sounds neat.

Regardless, developer Alliance Arts planned to release a Steam demo on October 19, but in a post on X (via Rock, Paper, Shotgun), the studio explained this is no longer the case. Instead, the studio will host "closed online play sessions for a limited number of participants."

"We sincerely apologize for the short notice," Alliance Arts added, before explaining its reasoning behind the demo's cancellation. Apparently, the demo would have come with "some restrictions on posting and streaming gameplay videos", which Alliance Arts decided it wouldn't be able to properly manage.

On top of this, however, the devs are also worried the demo will be exploited to create AI-coded clones. "We're concerned that the demo's content could be used to quickly imitate or replicate the game's presentation and gameplay with AI or other tools, and that such works could be released before launch."

It isn't clear which of these was Alliance Arts' bigger worry, but game developers having their projects vibe-cloned is becoming a more frequent issue. Back in March, a developer shared her game prototype online only to have it cloned within five hours. This isn't the only example either, over on Bluesky, McPixel developer Sos Sosowski shared numerous examples of indie game developers having their works almost instantly copied by AI cloners and released on Steam.

Games being flagrantly copied is not a new problem—the mobile space has struggled with it for years. But the way AI tools have sped up the process is clearly a cause for concern among developers—especially since one of the best ways to improve a game's chance of success in today's crowded market is to share the idea early and build a community around it. As such, unless Valve and other game stores take a firmer stance against outright copycats, we may well see more developers adopt Alliance Arts' approach.

An unannounced Star Wars and Marvel shooter in the works at Epic Games has been delayed, and its creators instructed to take an "AI first" approach to development. Staff were advised of the pivot last week, according to a Bloomberg report, with the project delayed indefinitely from its (unannounced) March 2027 release.

That "AI first" approach seems pretty comprehensive: the instructions per Bloomberg are to "implement an 'AI first' production model using tools such as Anthropic's Claude". Art will be "generated with full AI workflows", according to the same set of instructions.

Shortly after Bloomberg's report was published, Epic Games CEO Tim Sweeney took to X to share his own thoughts:

• On AI use at Epic Games in general: "We have a huge pivot underway here at Epic, moving developers around to retool for maximum AI acceleration," he said. "The aim isn't to do the same stuff with fewer people, but to enable everyone here to build MORE and BETTER stuff, and to bring the benefits to all teams using Unreal."
• On Epic 'not reading the room' when it comes to consumer attitudes towards generative AI: "Much of the room wants it to be 2016 again, but it's not. Our job, as the provider of tools and technology to much of the industry, is to design for the present that is and the future that will soon be. Personally I'm optimistic about it and see it as empowering for creators."
• On whether Epic will "accelerate an IPO": "Our more modest aim is to survive the AI revolution, help other developers do the same, and come out on the other end making better code and content faster."
• On potential improvements to development time offered by AI: "Right now it's highly variable, from 20% in some areas (like the most technical areas of UE [Unreal Engine] dev) to 5x for other areas (like web dev). By this time next year I expect very little will be less than 3x of January 2026 efficiency."
• On AI's effects on game design in general: "The biggest change will be much faster time from having an idea to playtesting it. This will be great for gameplay, as most of the recent AAA failures can be attributed to game systems going through far too much of their development process without being able to playtest.

Epic Games laid off over 1,000 employees in March, citing the decline in console sales, Epic's ongoing legal disputes with Apple and Google, and difficulty in delivering "consistent Fortnite magic with every season".

Are projects like Epic Games' Star Wars x Marvel crossover inevitable? Will we just have to cope?

Let's be honest: a Marvel and Star Wars crossover shooter made by Epic Games, predominantly utilising generative AI and AI workflows, will be a pretty fascinating experiment to behold. It's such a rich intersection of exhaustingly ubiquitous pop culture touchstones that there's every possibility its AI origins will be less detectable than they might otherwise be.

Because AI slop already mostly looks like Fortnite, Star Wars and Marvel. Add Pixar (which Fortnite looks like), and you basically have slop Mount Rushmore. This is the abyss staring back into the abyss.

Sweeney's quote about audiences wishing they were still in 2016 is telling. The AI industry loves to frame uptake of the technology as an inevitability that detractors / luddites will just need to cope with, or at least, they do when they're not loudly worrying that the technology might kill us.

Still, terrible things that no one likes do have a way of bleeding inescapably into our everyday lives. Tech companies own platforms best equipped to make us tolerate things we innately abhor. People hated Facebook's rollout of "likes"—now most of us "like" things with our mouths firmly closed. People hated it when social media feeds stopped being feeds, so that, for example, weird videos about putting various objects through meat grinders are now mainstream. The transformation from "way to catch up with friends" into "ruinously addictive habit" wasn't appreciated, but it's the norm now anyway.

In other words, the tech industry knows that it doesn't matter whether you like it or not: you'll get what you're given.

Capcom has been on an historic run over the last few years, delivering hit after hit during a period when many other publishers are struggling to keep their heads above water. This year alone, Capcom has released Resident Evil: Requiem, Pragmata, and Onimusha: Way of the Sword, all of which landed to both critical acclaim and commercial success.

So now is obviously the perfect time to wade into the not-at-all controversial topic of AI-powered game development.

Yesterday, Capcom held its own tech conference titled "Capcom Open Conference RE: 2026" during which programmer Satoshi Ishida gave a talk detailing the company's future plans for its proprietary RE: Engine—which powers all the publisher's modern games. During this talk, Ishida revealed Capcom's intentions to gradually transform RE Engine into an "AI-generation game engine."

Now, it's important to stress here that what Capcom appears to mean here is an engine for the AI generation, rather than a generative AI engine. As reported by Gamebiz (via IGN), Capcom's plans are more nuanced and detailed than slapping ChatGPT into its game development workflows.

During the talk, Ishida stated that the project is called "REX", and it involves six different programs that will integrate AI and machine-learning technology into RE Engine. Some of these are designed for enhancing program performance during development. Re:Dox, for example, is a data handling program designed to translate different types of data into a common set of rules for faster processing, while RE:UI is built to provide AI-assisted optimisation the RE: engine's development tools—sort-of like adding DLSS support for Blender.

Alongside these are tools designed to make the process of game development itself faster. There's a program called RE: Runtime, which apparently allows the engine to animate large numbers of objects and characters in groups—rather than running those real-time animations individually. Most intriguing is RE: Flows. This is apparently designed to simplify the implementation of game mechanic ideas, then convert them into a standardised programming language. This sounds vibe coding-adjacent to me, though it's hard to say for certain without seeing it in action.

According to Gamebiz, Capcom says it wants to integrate all these ideas into RE Engine for several reasons. First, modern games are so massive and technically complex that data management is a huge challenge and making changes to that data takes a long time. Second, more than 2,000 people now use RE Engine, including people from many different nationalities, so Capcom wants to create a system that is universally understandable. Third, AI technology is rapidly evolving and is essential for streamlining development—a somewhat arbitrary justification that I'm not convinced by.

Broadly though, these plans fall in line with Capcom's previously declared stance on AI, where the publisher stated it "will not implement assets generated by AI into our game content", but would use it to "enhance efficiency and boost productivity" in its development pipeline. Certainly, none of these tools appear designed to create art or character models and so forth, though it seems like some of them would require AI generated or AI assisted code to function.

It's also worth noting that Capcom's approach to development has worked pretty well for it over the last ten years, earning a reputation as one of the most reliable developers around. While I can understand the desire to make game development faster and more efficient, as the old adage goes, "If it ain't broke, don't fix it".

To most, the highlight of this week's big Deadlock patch is the snazzy map makeover, new heroes, and tons of quality-of-life additions. To me, it's all about the patch notes.

I spend a lot of waking hours reading white-text-on-black-background changelogs in this line of work, so when a page comes along that was specifically designed to be pleasant and exciting, it deserves to be rewarded. Valve has a long history of designing beautiful, bespoke webpages to introduce meaningful changes to its games, but the graphics department is really going ballistic with Deadlock.

The City Never Sleeps patch notes are styled like a tourism brochure for Deadlock's refreshed New York City, walking readers through scenic stops across the four major districts (Chinatown, Theater, Haunted Lot, and Plaza). Descriptions weave between in-world storytelling and technical explanation where it makes sense, with Valve's splashes of humor reminding readers not to take any of this too seriously.

Dotted lines draw the reader to new points of interest as they scroll while furthering the map motif, and the page is stuffed with custom artwork for new items and mechanics that might only exist there. Valve did such an annoyingly good job with presenting the next era of Deadlock that I'm almost tempted to play a MOBA (but only almost).

Finally, some good food. Welcome back, Valve patch notes.

Valve's golden age of patch notes

It cannot be understated that Valve pioneered storytelling in multiplayer shooters with Team Fortress 2, arguably kicking off with the 2009 WAR! Update, aka "Soldier vs. Demo." This was before the concept of multiplayer games having seasons, before League of Legends had a TV show, before "live service" was even a thing—Valve had created a cast of nine lovable weirdos and new updates that weren't just dropping items, but telling stories that put the characters in conflict with one another and added context to the multiplayer maps we'd already been fighting in for two years.

It was weird in the first place for an FPS to receive so many free post-launch additions, and it was even weirder for a developer to treat the arrival of new weapons, modes, and hats as worthy of dedicated splash pages with accompanying comic books.

TF2 was the gold standard for post-launch support, and its patch notes drew massive interest, often dropping late on a Thursday or Friday, Pacific time, hinting that Valve may have been polishing until the last minute. Valve's blog posts were never the same month-to-month, often adopting new color schemes, layouts, and original elements for the occasion.

There was no single template they stuck to. The Engineer update introduced his advanced Gunslinger gauntlet and Wrangler as blueprints with hand-written notes, while Scout's update debuted the Force-A-Nature scattergun in a department store catalog. I especially love announcing the Mac version of TF2 with a white, sanitized home page complete with symmetrical boxes and tiny text in the style of Apple circa 2002.

Poring over the "history" section of the TF2 website in 2026 is like uncovering one of those lost civilizations that invented a ton of sophisticated things before everyone else.

We're in an era marked by formulaic seasons, videogames as advertisement platforms, and teams scrambling to alter plans when they're not immediately popular. And yet 15 years ago, Valve had so much confidence in its direction that it could decide to wage a war between Soldier and Demoman one day (live-tracking the kills between both classes—Soldier narrowly won with 6.40M kills to the collective Demomen's 6.37M), introduce a crafting system the next, and then rebrand as Scream Fortress for October.

It wasn't exclusively good vibes—Valve caught flak for its (now quaint) paid hats, paid loot boxes, and speculative marketplace that became a fixture of updates. In a move that I would be surprised to see a studio make today, Valve made fun of itself at the time with a subsequent "Hatless Update".

Deadlock's blogs are practically retro, it feels weird to say. I wonder what inspired Valve to go this hard again. Counter-Strike 2 and Dota 2's changelogs do not share the same sense of theater. Their patch notes are comparatively regular, functional bullet lists—perhaps in line with how often those games are updated. Valve does occasionally produce unique web pages for major CS2 announcements or Dota 2 events, but we're far from the days when an incremental Mann VS Machine update was announced with a lavish movie poster in the style of kaiju cinema.

TF2's exceptional, ever-changing patch notes were a reflection of the game they promoted. Valve leveraged the flexibility of an arena shooter and imagined a world where Red and Blu were real factions locked in stupid conflicts. You never really knew what elaborate scenarios Valve would think up next for its over-the-top action figures, and that filled TF2 fans with a sense of wonder that very few games have pulled off since. Maybe it's no coincidence that the only example of innovative multiplayer storytelling in recent years, Helldivers 2, is another over-the-top war game that finds comedy in the futility of endless violence.

Valve is tapping into that dormant TF2 energy in small ways with Deadlock, and it's great, though I wonder if, being a MOBA, the game's inherent sweat level will eventually lead to a regimented, low-nonsense future that's more focused on the minutiae of balance. Whatever Deadlock is up to years from now, I'll be keeping up with the patch notes.

Martin Klíma is an iconoclast. As the co-founder of Warhorse Studios, he's responsible for one of the crunchiest, most peculiar, and best RPGs of the last decade—Kingdom Come: Deliverance 2, a game few other studios would make. As a former student rabblerouser in the old eastern bloc, he's the only person I've ever spoken to for this job who can say they helped overthrow a government.

He's celebrating the Warhorse's 15th anniversary, a milestone moment in a life that has a lot of them already. "I'd swear it was just yesterday," he says, that he and his fellow co-founder—the eternally fractious Daniel Vávra—were two disaffected devs striking out on their own. But Klíma's career did not begin with the incorporation of Warhorse Studios. It began in the actually existing socialism of Czechoslovakia, in the 1980s.

History check

"I was learning English at school, but it was sort of difficult," recalls Klíma. English educational materials had to be imported from friendly countries, and the only friendly English-speaking country at the time was Ghana, whose materials were miles away from the things Klíma was actually interested in. "I mean no disrespect to the people of Ghana, but it's a bit of an unusual approach to teaching kids English, and so our vocabulary was more about combine harvesters and blast furnaces than swords and dwarves."

What Klíma wanted to read, more than anything, was The Lord of the Rings. But although The Hobbit had been officially published in Czechoslovakia, there was no such luck for LOTR. Around 1987, Klíma was able to get his hands on a copy that someone had brought back from a trip westwards, "but I didn't know enough English to read it, so I had to improve my English. One thing leads to another, and the very fact that we can talk to each other right now is because of JRR Tolkien."

The first movie adaptation of The Hobbit was, in fact, this 1967 animation produced in Prague.

Not long after, Klíma was lucky enough to obtain another imported fantasy treasure: a Dungeons & Dragons Basic Set. "I think for everybody who experiences a role-playing game for the first time, it's an amazing experience: this idea that you can pretend to be somebody else, and you can assume a role and you can live through these amazing adventures in your fantasy."

The game gave him a kind of missionary fervour, a desire to spread the good word of D&D. "I started translating it with my then-girlfriend, now wife, and we went through maybe 80% of the text." The thinking was to publish a translation of the game via Czechoslovakia's underground printing presses; "to officially translate it, officially license it… that was completely unthinkable." Klíma already had experience in this sort of thing—he ran a fanzine that quietly printed popular sci-fi by "[pretending] that they were contributions to our student magazine."

Klíma's dreams of clandestine TTRPG distribution were disrupted by the other part of his life. The group he played D&D with was, in fact, a subset of an independent student union that he and his allies were organising, Stuha. As a student at the University of Prague, Klíma had organised a petition against the demolition of campus buildings. Not, ordinarily, a revolutionary act, but to the flighty Czechoslovakian government it was a "borderline rebellion."

In decades past, people might have shied away from signing a petition like Klíma's, but things had shifted by the late '80s. "Gorbachev was in power [in the Soviet Union]. There was this talk about 'change' and 'opening' and so on, so people did sign my little petition. In the end, almost 1,000 people signed it, and this really helped me to reach other people who were interested in changing things." It was not Klíma alone—Stuha had other organisers, other borderline and not-so-borderline rebels out to overturn the system—but his was an important role.

Important enough, in fact, that he was chosen to speak for Stuha at a public demonstration on November 17, 1989—International Students' Day, an observance tied to Czech student resistance to Nazi occupation and its subsequent violent repression. Klíma summarises it like so: "There was a big demonstration against the German occupiers in 1939, and on November 17, all the Czech universities were closed [in retaliation] by the Nazi government.

"It seemed like, if you make a student demonstration on the 50th anniversary of these events, then you must be very safe, because no regime would be so stupid as to fight students, exposing itself to this very obvious comparison to Nazi Germany."

The Czechoslovakian government, it seemed, did not see things that way. "There was this confrontation with the police, and people were badly beaten… that really radicalized the students and radicalized the rest of the country. In a week, there was a general strike, and by the end of December, Václav Havel was president of Czechoslovakia." The Velvet Revolution—Czechoslovakia's chapter in the Soviet bloc's collapse—had taken place.

Rolling dice

As capitalism took root in Czechoslovakia (and, after 1992, the Czech Republic), Klíma's sci-fi and gaming ambitions started to look less like a hobby and more like a career. Though he and his partner had been forced to set aside their translation of D&D by the Velvet Revolution, he returned to it in the aftermath. "I wrote a letter to [D&D publisher] TSR, and they never answered me… I even persuaded my relatives in the US to send them letters and faxes… but we never got any response from them. Maybe they thought it was a crazy idea. 'Why would some communist kids want to play D&D?' Maybe it felt absurd to them."

Eventually, inspired by the German TTRPG Das Schwarze Auge, and convinced no one at TSR was going to pick up the phone, Klíma decided to adapt the work he and his partner had already done into something new. "We rewrote most of the text. We created our own classes and our own races. We created our own combat system." That last bit in part because Czechoslovakia did not really possess the means to manufacture 20-sided dice in the early '90s, "so we only had a six-sided die and a 10-sided die."

The result was Dračí doupě. "It became hugely successful, because it was the first fantasy RPG [in Czechoslovakia], and it was really like when it rains in a desert and green sprouts appear… and then I've worked in the games industry ever since." It wasn't long before he moved into computer games with titles like Fish Fillets, Original War, and UFO: Aftermath, developed under the Altar Interactive brand, itself spun off from Altar Publishing, whose name was on Dračí doupě.

Klíma's experience in the games industry between the 1989 revolution and Warhorse's 2011 founding could probably fill books by itself, but a recurring theme in what he told me about those days is frustration with publishers. This came to a head in the late 2000s with an RPG called Vision, led by Czech board and videogame designer Vlaada Chvátil. "This is going to be our big break," Klíma remembers thinking at the time, "This game is going to make history."

The game's publisher, CDV Software, was not quite as bullish. "CDV asked us to do some changes, and when we did these changes, they told us that the managing director who signed the original contract with us had left, and now they were no longer interested because of these changes that we made. And then six months later, they went bankrupt as well.

"We were sort of despondent and turned over the company to Bohemia Interactive," says Klíma, with the transfer happening in September 2010. After working on games like Arma for a while, followed by a stint at Codemasters, he was approached by Operation Flashpoint lead designer Viktor Bocan (now one of Warhorse's two creative directors, alongside Prokop Jirsa).

Bocan knew a guy with a vision: an achingly realistic RPG set in 15th-century Bohemia, he just needed someone "who is good with Excel, who will do this boring stuff about numbers and plans and so on." The game, of course, was Kingdom Come: Deliverance, and the guy with a vision was Daniel Vávra, its director. Bocan told Vávra to wait right there while he got Klíma on the phone. Warhorse Studios was born.

Thy will be done

One of my first questions to Klíma was, if he could travel back in time and tell his past self how KCD had panned out, would he have been surprised? His answer: the game's success wouldn't have shocked him a jot, but the fact that it actually got made and released would have amazed him.

"We were asking for quite a lot of money," Klíma recalls, and publishers, perhaps predictably, weren't champing at the bit to take a big risk on a studio making a historical RPG that had strong opinions about the order in which you equipped your armour. "It was really difficult at the time… It took us two years to find an investor."

The investor was Czech-American billionaire venture capitalist Zdeněk Bakala. Bakala's cheque gave Klíma and co the means to properly start Warhorse and begin work on KCD, but not enough to complete the game from beginning to end. "We had a really nice demo of the game. We hired an agency. We traveled the world. We talked to, I don't know, 50 different publishers on three continents. Nobody was interested."

The answer: Kickstarter, where it turned out that 35,384 backers were more willing to take a punt on Warhorse's weird medieval RPG than major publishers were. "I think [it's] the idea of a grown-up RPG," Klíma reflected. "I still am bored and tired of all the RPGs where you are this chosen one… Nobody is making a game that would say, 'Okay, this is a big RPG, and you are nobody, and you are not going to become a world saviour. This is a game about an ordinary random person that's doing ordinary, random-person stuff, and at the end, he is just moderately less of a nobody.'"

KCD scratched that itch for over 35,000 people, where "publishers, bless their hearts, are very conservative by the nature of their business… If you have this huge studio, like these established publishers do, it means that your next game is going to cost $100 million… and then you need to sell 4 or 5 million units just to break even, and that's difficult with a new IP.

"I've been in this business for 30 years now, and I only managed to sell one game to a publisher, and that was Original War back in 1997. This experience, trying to sell Kingdom Come to publishers, was really eye-opening in realizing how difficult, almost impossible it is. So it was a great relief and vindication when we saw that the players agreed with us."

All we have to decide is what to do

The first KCD did not light the world on fire—and in Klíma's estimation, is more like Kingdom Come "zero point eight" than a full realisation of the studio's ambitions (that would be KCD2)—but it hit hard enough to let Warhorse expand and to put it on the path to KCD2. "There are of course people who think that KCD1 was a better game than KCD2 and KCD2 is just a watered-down, mainstream, plays-by-itself game for the masses.

"I don't agree… but I can see why somebody thinks that. But still, I think the fact that [KCD1] crashes every three hours is not a good experience," he jokes, "even if you are the most hardcore player ever."

But lots of people, us included, really did like KCD2, and the game was successful enough to send Klíma full circle, from teaching himself English via LOTR books to negotiating the development of a Warhorse-driven LOTR game. "It's clear that because of the success of Kingdom Come, somebody would be willing to entrust us with this license."

Warhorse is now a two-game studio, its teams divided between its LOTR game and an upcoming Kingdom Come adventure. "Now you have people sitting in the same office that used to work together, were part of the same team even recently. And now we have to split them into two groups and focus them each on their own project, and that's quite a challenge."

But overall, probably one of those problems it's nice to have. So there's a little paradox: with Warhorse and its ambitions both growing, the wider games industry seems bleaker than it's been in decades, riven with layoffs and ignominious cancellations. For Klíma, it indicates one thing: the growth of videogames as a market has run into limits it can't transcend, even as the progress of technology continues to demand splashier games with bigger goals and larger teams. "You throw more people on the job, and that makes the games more expensive… If the game is more expensive, it needs to sell more units in order for this to be a viable business.

"How can you do it? The easiest [way] is you grow the market… But we reached the limit here long ago… If somebody is not playing the games, it's because they don't want to play the games, or they can't afford to play the games. So there's no way unless we teach dogs how to play games." Another way to squeeze more profit? Cut out the middlemen, but the industry's done that too: "We destroyed retail, and we took all their money," jokes Klíma.

"The last thing you can do is to increase the unit price. That would help a lot, but everybody's scared to death about it. Now, hopefully, GTA will blaze the trail…but I think we've run out of options." Meanwhile, technology keeps getting better (though it's maybe been arrested, a little, by the RAMpocalypse) and expectations keep getting higher, in Klíma's assessment.

"I don't see a way out of it other than moving the production to the countries that are cheaper… it's maybe Central and Eastern Europe, maybe China, maybe India, maybe East Asia in general." Good news, perhaps, for Czechia-based Warhorse, but a grim prognosis for North American and west-European development. "I take no pleasure in being a prophet of doom and desolation, but specifically in North America, I think it's a really, really huge problem right now."

The solution much of the executive class is keen on is, of course, AI. For Klíma, many of them have a "simplified mental model" of how it can be introduced that doesn't quite work in reality. Paraphrasing him, that model goes like this: if AI is a 5x force multiplier, he says, many people imagine you can cut a 200-person dev team down to 40. Even if each developer needs their own AI, and each AI costs the same as one dev, you can now do the work of a 200-person dev team while only paying for the cost of an 80-person team.

Even leaving aside questions of the quality of the AI's output, Klíma doesn't think it works like that. If you have 200 devs and a magical AI multiplies their output by five, players are just going to want the game that 1000-person-dev-team equivalent can make. "This is not a 'get out of jail free' card, the AI. It will not reduce the cost of the game, but it will allow us to create bigger games for the same cost."

I'm more than a little sceptical, as you might expect, but it's clearly a question Klíma, and Warhorse, has been thinking about. Will the force multiplier (as he describes it) of AI play a role in Warhorse's future games? For the first time in our long chat, Klíma demurs at the question. "I can't talk about it," he says. Something for us to find out, I suppose, on our next visit to Middle-earth or Bohemia.

In 2013, Valve head honcho Gabe Newell spoke to students at the LBJ School of Public Affairs in Austin about "Productivity, Economics, Political Institutions & The Future of Corporations." It was a much more interesting topic than that makes it sounds, and has been doing the rounds again thanks to being sliced up into shorts by the Webknower channel.

One clip that's been going around is Newell's response to being asked about how easy it is for hackers to get into game worlds and do things the creators don't want them to do: "One of the biggest mistakes game developers make is to mistake hacking, or to mistake a genuinely entertaining action by a player as hacking. So the example game developers tell each other is when Lord British was killed during his first speech in Ultima Online they rolled the world back rather than recognizing that it was the coolest thing that had ever happened in Ultima Online."

In the early singleplayer Ultima games, Lord British had huge amounts of hit points and other defenses to discourage players from attacking him, which of course made some of those players see it as a challenge. Blasting him from offshore with a shipboard cannon or using a magic artifact called the Skull of Mondain were a couple of the classic techniques.

In Ultima Online, where Lord British was controlled by Ultima creator Richard Garriott, his murder also took some doing. A player named Rainz waited for Lord British to give a speech, and used his high-level pickpocketing skill to steal a scroll with the fire field spell from someone in the crowd. He didn't know it would work, but after the server reset immediately before the event, Garriott had forgotten to toggle his character's invincibility back on. Lord British caught fire and died, and Rainz was banned from the game as if he'd cheated when actually what he'd done is contribute to the lore that gave Ultima Online its reputation.

Newell's point is that if Ultima Online had embraced that, rather than resetting the server and pretending it never happened, the player-led storytelling would have made for something more interesting than the ongoing plot of Ultima Online. Which is forgotten today, while the assassination of Lord British is a central text of MMO lore. Game designers will be telling that story to new hires for years to come.

The industry obsession with high-resolution graphics has haunted it for years, but given the dire state of things in regards to both layoffs and hardware prices? It's now starting to be downright incompatible.

Raph Koster, a veteran developer who worked on Ultima Online and was the creative director for Star Wars Galaxies—and is now making another MMO, Stars Reach—has been banging that particular drum for years.

"I think the first time I spoke about, ‘Hey, cost is going to kill us’ was 2005," he tells our friends at Edge Magazine in a massive interview inside Issue 428. "Myhrvold’s Law says that software is a gas—it expands to fill whatever computer you have … The place that shows up the most in games, for better or worse, is in content costs."

Reiterating similar thoughts spoken to me when I chatted to him back in June, Koster says that "what used to be a single 128x128 texture is now multiple 4K textures with shaders and all this stuff. And, yes, tools have improved along the way. You don’t handpaint that texture pixel by pixel any more—you can use procedural tools like Substance, or just purchase it from a texture library or whatever.

"But there’s still [a] downstream cost, even if it’s just testing how those umpteen textures and shaders interact. It all adds up." Combine that with the fact that game prices haven't gone up in parallel to these ballooning development costs, thanks to constant deals—"who buys something at $60, even, today?"—and you've got a recipe for disaster.

A cyclical one, mind. "Unless we hit the singularity," Koster says, "a platform will come along that changes things." Koster lays out a few key qualities this hypothetical platform would require in order to save the industry (or at least, the developers working within it) from this steadily-approaching iceberg:

"It needs to provide a new affordance, meaning there has to be gameplay you cannot get except on that platform … [it needs to be] worse for making traditional games—because that means the big incumbents cannot just transfer their expertise, come in and dominate.

And to top it all off? "The number-one biggest thing this new platform needs is to have shitty graphics." Referring to a few historical examples, like Flash games, social games on Facebook, the Wii, and mobile. That's not to say these platforms'll be immune to those problems, though.

"They eventually catch up, and then the costs rise in exactly the same way. But we haven’t had a new platform come along to reset that cost curve in a long time. Like, a long time."

Early access on Steam is a simple concept. You buy a game before it's finished, provide bug reports and feedback to developers, and help shape a project's direction and ambitions. Everybody wins: You get to play the game early, and developers earn money that helps them continue and finish their work. It's been a boon for numerous notable games over the years, including Baldur's Gate 3 and Palworld.

In an interview with GamesRadar, though, Pocketpair communications and publishing chief John Buckley said that proposition is becoming increasingly difficult for developers, because the whole idea of early access has "lost its meaning."

"Go back to the Steam Greenlight days, when Steam changed Greenlight to early access, there were a few years where it really was running and operating exactly the way I think Steam dreamed it," Buckley said.

"I think everyone knew what they were in for. The kind of people who were buying early access games knew what they were in for. They were the kind of gamers who want to be involved, they want to give feedback, they want to talk to devs. It was this really beautiful golden era."

Now, though, the rise of what Buckley called "hyper-casual mobile culture, and hyper-casual gaming culture, and live service gaming culture" has eroded that understanding, and the accord between developer and gamer that rose from it: "The culture has changed and early access, essentially, has lost its meaning. People just buy an early access game like they buy a normal game and treat it like that's the game."

There's no doubt some truth to that. Generally speaking, I subscribe to the principle that if a product is offered for sale, and people pay money for that product, they're well within their rights to expect that they're getting the product. But early access isn't quite so cut-and-dried: Steam Greenlight was a more cumbersome system, but it also clearly delineated between a work-in-progress and a final release, which early access releases often do not.

Dialed-in gamers, or those who are old enough to remember when beta testing was an actual phase of development rather than a promotional tool, may have a clearer idea of what to expect from early access games. But as gaming becomes more of a mainstream pastime it's reasonable to expect that an increasing percentage of the audience is going to neither know nor care about what early access: They just want to buy a game and play it.

Buckley doesn't believe that's necessarily a bad thing, but he did say "it's getting harder on Steam, sadly, because this culture has come through." And while he remains a believer, he also acknowledged that early access isn't ideal for every game, and that the system is "abused" by some developers.

"The biggest way it's abused now is just like, 'We'll release it now because we need money'," he said. "It's a way now to sidestep traditional funding means. It's almost become the new Kickstarter in some ways. Put out the build to get the money."

I'm not entirely sold on the idea that "put out the build to get the money" really qualifies as abuse—it is, after all, the whole basis for the early access system. But it can lead to trouble if early access sales aren't enough to support continued development, or simply through mismanaged expectations. The best thing developers who release into early access can do to avoid potential headaches, Buckley said, is to be clear about what's happening and, as much as possible, keep people fully informed.

"All we can do is try to softly educate people. I think we do an okay job. I think we try to be quite transparent. 'Hey guys, the next update's a while away. It's not going to be content-heavy, it's going to be QOL-heavy.' We try to balance it."

Pocketpair has indeed taken that approach with its fans: Buckley recently warned his followers not to expect too much from Palworld's 1.1 patch, saying it's "not planning to spend a year making a massive update."

Nobody likes game delays. Sure, a delay is temporary and suck is forever, as Gabe Newell once said, but it still chafes. But sometimes a little extra development time can go a very long way. In a new interview with the NoClip podcast, immersive sim maestro Harvey Smith shared one such story about Deus Ex, explaining how it leveraged some extra post-production time to give a much bigger role to one of the game's most famous NPCs—and make it a much better game in the process.

Smith's tale actually followed a remark by Ben Horne, his co-founder at the new Black Pony Immersive studio, who said, "You can elevate a good immersive sim into a great immersive sim in a very short amount of time, a matter of months, if you just protect a post-production time."

Publishers generally don't want to hear that, Smith added—although he obviously agrees with the sentiment—and so 20-some years ago, when Ion Storm Austin asked Eidos for more time to tighten up Deus Ex, "We geared up for this giant fight." Instead, Eidos just said yes.

"They were so distracted by what was happening in Dallas with Daikatana and such that they were like, 'Sure, sure, sure. Take five more months, take six more, whatever you need.' And we were like, 'Oh, cool. Okay'."

What was happening in Dallas, as Smith put it, was essentially chaos, which you can dig into via our history of Ion Storm. But, as Smith has previously related, it was a useful distraction for Ion Storm Austin, which did indeed get the extra time it wanted to really nail down Deus Ex.

A focus of that extra dev time was one of the game's foremost NPCs, "this German guy with a skull gun" named Gunther Hermann—who, for the record, very much wanted a skull gun but as far as I know never actually got one despite filing the required paperwork. "He's kind of a big presence in a couple of missions, you go on a mission with him, and then he just vanishes," Smith said. "He has no closure. He's not even in the game.

"So, I literally said, I can drop him into Paris, a big fight in the cathedral, which is an empty building right now, and I can hide some triggers on the roof with data links, and if you're crawling around on the roof because you have 'super jump' or whatever, I'll have him say, 'I'm watching you, Denton, like an insect crawl around on the roof. I am watching you through the satellite, which I am linked to right now, Denton.' You know, so we do this kind of thing, and players go, 'Well fuck, somebody's watching me from a satellite.' It's like smoke and mirrors, and then you get in there and you fight it out with him.

"If you've been paying attention, you know that these earlier models, the cyborgs, they have kill phrases, him and Anna Navarre. And this was an idea that we took from Neil Stephenson or somewhere, so Denton can say this like long crazy kill phrase and Gunther just fucking dies, right? And it's a powerful moment. And we never would have had that in the game had we not gotten that extra time to just look at our game."

This is fascinating for me personally because I had that exact experience: Skulking around a Paris cathedral, quietly working my way inside, when all of a sudden Gunther busts in on the radio to let me know I'm not fooling anyone. Giving it even more weight, I discovered before our final confrontation that Gunther had been waiting for me for some time—and that MJ12 troopers had heard him quietly crying, probably because I'd previously killed Navarre, his cyborg partner and someone for whom he apparently had deeper feelings than were immediately apparent.

Making it even better (and I swear this is true), Gunther bugged out near the end of our fight and ended up running into a corner of the cathedral basement, where he stood, silent and unmoving. He looked like what he was—an utterly beaten, broken man, who'd had everything taken away from him—and I couldn't bring myself to kill him. At the time, I thought the option to show him mercy was intentional, like other hidden NPC choices in the game. (It didn't happen in subsequent playthroughs, though, and I've smoked ol' Gunther's ass every time since.)

It was a powerful moment, as Smith said, and it's kind of wild to imagine a version of Deus Ex where Gunther doesn't get his big finish—and what else might have been left on the cutting room floor besides. Something to keep in mind the next time you find yourself grinding your teeth over a delayed game.

Smith reiterated the importance of that point at the end of his story. "If you are listening to this right now and you fund games, give your team a post-production," he said. "Give them three to six extra months and you will win some awards and you will make a bunch of money, instead of leaving a crater in the ground."

Generative AI in videogame development, despite the looming potentiality of a bubble popping, has continued to be a hot-button topic pelted at developers during earning calls, despite most of them recognising that there's not much use for it beyond mundane busywork.

Sure, it's being used commonly, but rarely for the actual making of the videogame itself. Filling in the roles that procedural generation might or, like, resizing shoulderpads—or for "user preference analysis". Not exactly the kind of stuff that revolutionises an industry, and certainly not enough to justify reducing labour en masse.

This also seems to be the view of joint CD Projekt Red CEO Michał Nowakowski, who was asked such a question during an earnings conference yesterday, and once more found himself required to emphasise that they're using people to make videogames like The Witcher 4: "AI is definitely becoming a thing in games development as well. We are still predominantly using people and this is our plan—to make games where we're not walking away from that."

That's not to say that AI tools aren't being experimented with, they're just nowhere near a focal or vital part of development for the studio: "AI, if anything, is a tool like many others that's helping in places that can maybe become more helpful in the future but we're not planning to rely on AI making complete games."

Part of that, Nowakowski emphasises, is because the kind of games they make are so complex that it wouldn't make a lick of sense to leave the careful craftsmanship of them up to an LLM or vibe coding: "especially with very complex games like ours, I don't even think this is the right or ethical path to follow.

"So whether this is true for the whole industry, I'm not the person to make those statements. I think everybody has to make those judgments on their own."

Even if you scrap the general distaste for AI slop, it's not like going all-in on the tech is risk free. From a legal standpoint, anti-AI clauses are quickly becoming boilerplate for videogame publishers, at least when it comes to the creative work you'd need to prove you own in a court of law.

Dwarf Fortress has an endless capacity for telling stories. It's a game built on procedural generation, and any longtime player'll be able to spin up a dozen yarns about some absurd thing that happened to their gaggle of miners.

Just look at a headline from our own Lincoln Carpenter: "Dwarf Fortress forensic specialists blame mysterious case of spontaneous human combustion on a historical artifact that suddenly remembered it's four times hotter than the surface of the sun." What a beauty.

Anyway, this sort of puts the lie to the idea that generative AI will be the first advent of something that can tell infinite stories, always and forever—because Dwarf Fortress has been doing that without a single LLM for its entire existence. Point being, if you want something that can tell infinite stories, you're probably best off giving us the tools to do that, rather than the finished product itself.

Speaking to PC Gamer's Joshua Wolens at Gamescom 2026, co-creator Tarn Adams gets a little into it while talking about how they decide whether a simulation's worth it to layer on top of the game: "If you feel like you've played through it yourself a bunch, and you're like, 'oh, that was cool. That's a good beat,' and you're seeing a little narrative forming? Then you know it's going to work."

Adams talks about a hypothetical in which a meeting of dwarven barristers is interrupted by a very rude dog that comes in and bites someone—not even particularly strange, by Dwarf Fortress standards. Asking him as to why we find these emergent, mechanics-based stories so endearing (as opposed to AI slop, which hasn't yet produced a notable story people seem to like), Adams responds:

"Well, I mean, [you like human stories] because you're a human, and you resonate with things that other humans have done because there's a common experience at just the barest basic level.

"You can't always tell," Adams admits. "And that's part of the trick, right? That's what a Turing style test is, right? Can you tell the human spark that's in there? And when you fail a Turing Test as a human, meaning that you got confused by a robot … maybe that's credit to the robot, or maybe that's discredit to you. I don't know which one, but we're all going to be very sorely tested in the coming years as we're surrounded by such matter."

It's an interesting way to look at a Turing Test—not whether a computer can act as intelligently as a human, but whether a computer can replicate our "human spark". Which, at the risk of sounding sentimental, does feel like a vital part to actually enjoying art.

Speaking personally, a large part of why AI art or stories don't interest me isn't necessarily in the technical proficiency of what's happening—it's because I'm interested in learning about a person's worldview, their tastes, their likes and dislikes, their hopes for the future, and having those hopes stirred in myself. The fact a person made something causes it to be more interesting to me.

Which I suppose is why the proc-gen stories of Dwarf Fortress are interesting, because there's a human touch in how the random, computer-generated narratives are described. If you just rattled off a list of events, you'd bore me half to death. But phrase it like Lincoln did above, and I'm having a sensible guffaw—these silly little dwarves shall make writers out of us all.

Generative AI is a hot-button topic in the industry, and for good reason—it's been kinda destructive, whether that's jacking up hardware prices or putting people out of a job—and while there have been some fairly mundane use cases that seem helpful, it's been hard to escape the sheer sense of mania that's built around its use. Or the bad PR, juiced out of the sheer exhaustion we've all had from executives promising the matrix or something.

So, naturally, it came up in a conversation with Tim Willits, the CCO of Saber Interactive and also the co-owner id Software from 1995-2019. Speaking to PC Gamer's Joshua Wolens at Gamescom 2026, he notes that: "As most people know, we are experimenting with AI in future games. We don't have it here right now, but we announced [Rideshare Simulator] and I want to clarify things because there's some misconceptions."

If you're wondering why we're talking about a Rideshare sim, here's some context: Saber landed in hot water earlier this month over it, because it seemingly delayed its AI disclosures on Steam for a while, and was exposed as using AI by a former writer who claims they were booted off the team in favour of it.

That same writer was badmouthed by Saber's CEO Matthew Karch, who later apologised. Willits says that Rideshare's main campaign mode and all of its side missions are written by humans, and the AI components are only used for an optional game mode called Free Ride: "We're going to use AI for the passengers, so we can deliver new stories.

"So that is only possible if we use AI for that—this is an experiment. We at Saber want to figure out how we can use this technology, so we're doing it at a small scale to try to learn. We're like everyone else, we're just trying to figure this out. But there is no generative AI in our big titles. No Hellraiser, no Turok, none in Stuntman."

I feel like an asterisk has to hang over that "only possible"—while, yes, churning an infinite stream of dialogue for an optional mode is only possible with generative AI, a free ride mode with enough conversations for it to not grow stale absolutely could've been cobbled together with actual writers and voice actors. Sure, you'd run out eventually, but I don't think anyone's hungry to make Rideshare Simulator their forever game.

Still, Willits emphasises: "We want to experiment, and I do believe that there will be a company—I don't know who—but some game development company will create a game that uses AI in such an amazing way that everyone will be like, 'That's it, that's what we need to do'." He then goes on to joke that "someone needs to come up with the 'Half Life 2 of AI', something that's so amazing that it just launches some type of killer app."

I remain characteristically unconvinced, though in fairness to Willits, I do at least think that—if generative AI does find a use as the core of a videogame—it'll probably be a new genre of its own, rather than trying to create an infinite RPG with infinite NPCs. An act which would just shake out to make something unappealing and meaningless.

The thing that makes me cynical is we keep watching money being burned and datacenters being built precisely because the tech doesn't really seem all that useful for anything other than busywork. When it takes centre-stage, it produces voice acting that's bland, writing that's terrible, and art devoid of direction or memorability, even as it gets more technically proficient.

I'll eat my artificially-generated hat with a side of humble pie if Willits' "Half-Life 2 of AI" game materialises, but a lot of people have swung at that ball and missed, and each time I feel like the bat is a little smaller. Still, Willits says that id Software probably would've used generative AI if it was a thing back in the day—because the studio was always at the cutting edge.

"Yes! Oh, of course. I mean, look at what [id Software] has done. In 1999, when we released Quake 3, it required a graphics card, and we got so much hate for that. People said we were going to ruin the industry. Nobody was going to spend $100 on a video card, and it was the dumbest thing that anyone ever did … I know for sure back in the day if it was available, John Carmack would be like making Skynet or something."

Still, to be fair to Willits, he seems to share my view that the game'll have to prove that generative AI can change the industry itself, rather than via the promises of a CEO somewhere: "The games will speak for themselves, even in our experiment with Rideshare, I encourage people to check it out when it comes out, play it themselves. They may be like 'oh, this is cool! I like this'. And if they don't like it? Fine. And but the games speak for themselves."

One tragedy about the fall of Arkane Austin—aside from Microsoft's seeming unwillingness to give studios a chance beyond the next quarter—is the pervasive feeling that the studio, whose devs were previously known for their immersive sims, was forced to make something way outside of their wheelhouse. Redfall was a live service shooter, and a pretty cruddy one at that.

So it might be buoying news to hear that Dishonored co-creator Harvey Smith and seasoned Arkane vet Ben Horne's new studio isn't going to try and chase down trends. In an interview with The Game Business, Horne and Smith were asked if they thought the immersive sim (which Black Pony Immersive has been strangely careful to talk around), is a dead or dying genre.

"If you look back through those games," says Smith, "There are a lot of titles that people still talk about 10 or 20 years later. They have a lot of cultural cachet. And there are a few, if you look at BioShock or Deus Ex or Dishonored, that make a lot of money."

Part of the issue, as many developers have outlined, is the general ballooning budgets in the industry itself: "If you look at Dishonored, for instance, the budget that we were working with was… I wouldn't say tiny, many triple-I developers would envy that budget, but it was very reasonable. And our [return on investment] was four or five times the total spend."

For Horne's part, one of the mission statements of Black Pony Immersive is to get away from tend chasing—which Arkane was clearly and famously bad at—and just make stuff its developers would like to play:

"The industry is growing in such a way that there are a large amount of players with a really varied appetite. We don't really know what turn the industry might take, and we're probably not going to be very good at trend chasing. And so, our best shot is to rally around a creative vision that excites us as developers. It excites the audience that we know wants to play this kind-of game. That's the best chance of making something great."

Smith concurs, outlining the (understandably cynical) perspective that, hey, he's heard a lot of these promises from big executives before: "Very well-paid people say things to me, going way back now, like, 'First-person games don't sell,' or, 'RPGs are a niche market,' 'If you're not working on a free-to-play game now, in five years you'll be out of a job,' 'MMOs are the future,' 'In the future, games will have live motion video sections with Hollywood actors in them."

If we skim through that list, we can see a lot of nonsense. First-person games have continued to sell, one of the most successful games of the decade was an RPG based on Dungeons & Dragons, and live-service free to play games have been an absolute trashfire for some publishers, even with deep wallets. And while Hollywood actors are in videogames now, they're not in live action.

"I've heard smart, good people," Smith continues, "friends of mine in executive positions, say these things over the years, and sometimes they're right. But what Ben said, going back to what you love with a strong vision, that often comes through for us, so that's what we have got to do."

Redfall is one of the biggest industry disasters in recent memory—coming from Arkane Austin, a studio known for an all-time classic in Prey, the game was panned both critically and commercially. Even we, at PC Gamer, dubbed it the "first Arkane game we don't like".

It also served as the death knell for the studio as, one year later, studio owner Microsoft announced it'd be shutting Arkane Austin down. The narrative paints itself, especially given Microsoft's utterly brutal behaviour in following years. The big bad M forced Arkane Austin, king of the immersive sim, to make a crummy live-service glob.

And that's sort of the truth—it is, after all, why the studio founder left—but it's not entirely Microsoft's fault. The studio director at the time, Harvey Smith, tells The Game Business in a recent interview that it wasn't just corporate which became swept up by the live service fugue.

"Arkane is very specialized, but the industry is growing, and there’s always more mouths to feed. And it’s great when you can reach a wider audience. We got swept up in the games-as-a-service thing, and I had the foolish chutzpah to assume that it always works out for us."

Smith describes a sort of naive optimism that Arkane could nail the same live service successes elsewhere in the industry: "We can do this. Let’s take what we love, and mix it with [games-as-a-service]. And do different mechanics by day and night. And have character classes.

"Then in the middle of all that, Covid happened. We also had the great resignation where everybody went remote and started changing jobs." Later in the chat, Smith describes a deeply chaotic development cycle, plagued by all sorts of baffling occurrences: "There were fires. The government announced aliens at some point. There was a swarm of murder hornets coming toward Texas. We had an ice storm that put power and water out for like two weeks for some people. It was crazy. That development cycle."

And, listen—I think it's clear in hindsight that Smith was wrong to be optimistic, but it's also clear that he did genuinely care about Redfall, even if it came out a mess. So much so, in fact, that Smith reveals he pleaded with Microsoft to allow himself and other employees to come back into the office and finish the game's 1.4 patch.

"We finished the game and it was very poorly received. We then worked on it for a year. We patched thousands of bugs. We then got laid off, but we had in the pipe version 1.4. It was the one that fixed tons of stuff. So, we went to Microsoft and said, ‘I know we don’t work here anymore, but can we please release this?’

"And to their credit, they did not say no. They let six or seven of us sign forms that we could come back in the building. After being fired, we released 1.4, which I think if we had that on day one, it wouldn’t have set the world on fire, but it was a much, much better game."

Ben Horne, who was the production director at the time, chimes in: "It feels like an achievement unlocked for me in my career, to release something after I’ve been fired and the studio closed down."

Patch 1.4 wouldn't necessarily save Redfall from obscurity. PC Contributor Rick Lane went back to it, writing: "still isn't the game it needs to be, or the one Arkane Austin could have made. Nonetheless, I think the update shows that Arkane could have redeemed it in the long run." The tragedy, I suppose, is that we'll never know whether 1.5, 1.6, or 1.7 would've finally turned it into Smith and Horne's vision.

Take-Two Interactive, the publisher who oversees studios like Rockstar and Gearbox, had an AI guy. 'Had' being the operative word, seeing as he, and a decent chunk of its then-AI team, were laid off around a year after he took on the role. Dicken was previously a data scientist at Zynga, before becoming its director of applied AI in 2019, when a lot of generative AI and LLM technology was still in its infancy.

As someone who was the head of AI for one of the biggest videogame publishers on the planet, capable of making such tectonic swings that other developers are scampering to get out of its videogame's launch date, you'd think that Dicken would be one of the true believer types. It's my pleasure to, via an in-depth interview from The Side Quest, reveal that Dicken is actually overall pretty skeptical.

It's not as if he's given up on the technology entirely—but he certainly speaks with the same echo as many industry skeptics, including (perhaps tellingly) Take-Two CEO Strauss Zelnick himself.

First up, there's the public perception, which—when discussing Clair Obscur's lost Indie award wins after its own AI placeholder debacle—says that creatives in the industry "see the use of generative AI as an abdication of intentionality, as an abdication of authorial intent. And consequently they assume, in a lot of cases rightly (I'll caveat the hell out of what I'm saying here) but in a lot of cases rightly, that's a pervasive abdication of authorial intent."

As for the masses, "Western audiences, so North America, Europe, seem to be more strongly anti-generative. And anecdotally, what I've heard is that Asian audiences are more accepting and more tolerant of this stuff … The internet's not representative of the mass audience. So to what extent is the anti-GenAI sentiment mainstream versus vocal minority? I don't have an answer to that."

What has Dicken more skeptical from a place of authority, however, is the actual graft of how videogames are made. For example, a lot of videogames with AI-generated dialogue haven't taken off the ground because, well, more isn't always more:

"Games are curated experiences. When you look at a dialogue tree and there's three options, yeah, you could make that 50, 100, we could have made it 100 options without generative AI... you've got 100 options, but only two of them get you to where you actually are trying to go. Adding more options is not the answer."

Then there's the actual purported productivity gains where, even in areas like coding, where AI seems to've dug its heels in the most, Dicken says it's debatable: "I often see people throwing around, like, oh, my engineers are 100x engineers because of Claude Code … [but] individuals are not necessarily reliable narrators of what the impact is, so you need to be quantifying objectively, and very few people want to run those case studies."

On the flipside, "If you're a true believer, it's so obvious that it's having this massive impact, why would you run a case study? Why would you spend the time?"

It's that true believer-ism that Dicken seems the most vocally opposed to, though—batting aside the promises of generative AI as some miracle tool, because as he sees it, in a gold rush, people don't sell you the tools if they could sell you gold.

"If any of this shit actually did what people are billing it as, they would not be talking about it," Dicken says. "These would be protected like the crown jewels and Fort Knox combined … It's like the guy sat here telling you that there's a gold rush going and you need to get in on the gold rush, buy my pickaxe. The guy should be selling you the gold if he's actually got the gold."

That's not to say companies aren't able to screw you over if they have sold you the tools. For many, the price of tokens has risen substantially, a sort of looming 'oh, crap, we're in a bubble' situation that, among other things, caused the CEO of Palantir to have a bit of a meltdown on television.

Which isn't great if you're a company, because at least with employees, you decide how much you pay them—you don't get that luxury if you're leaning half of your workload on AI: "Where you bake it into a pipeline, and that pipeline has now rigidity, and you are now vendor-locked to something that you know is going to get its price jacked up."

The Side Quest compares it, aptly, to the Unity debacle wherein the ubiquitous engine suddenly jacked up its prices, leading to a year of outrage and what I can only describe as reputational implosion. "We're sleepwalking into that same situation," Dicken agrees.

And yet, the goldrush mentality persists—even as corporations repeatedly pants themselves: "These are organizations that are supposed to be serious organizations that you would imagine would have guardrails against this stuff … The FOMO and the newness and the velocity and all of these things are contributing to an environment where you can't get the guardrails down fast enough to actually prevent some of the crazy."

As for where he sees it all landing, Dicken's experience in working with AI since the technology's inception

"I think that there will be an amount of long-term changes. But in the way that the industry had to figure out how to adopt 3D tools like Maya, it's not going to be the seismic shift. That's my sense. Could be wrong … I think that they are not a trillion dollar business that's going to revolutionise everything. But as a piece of technology, I want to figure out what these things could do for games."

Still, even if I personally eye-roll at the prospect of having AI in videogames—outside of hypothetically neat toys—Dicken is at least cognisant of the fact you need to actually use the correct tool for the job, not just cram AI-generated slop into existing genres or types of games and expect it to work.

"There is a fundamental thesis to a game, and what you put into that game needs to align with that thesis." Examples where this goes wrong, Dicken adds, "is just shoving stuff in to tick the finance earnings call investor box, rather than trying to build the best damn thing you can build that's going to earn the most money, because it's building something that people want."

As I sit here, watching Saber Interactive's CEO have a normal one about how generative AI dialogue can totally improve videogames, I can't help but grit my teeth in the same, enamel-grinding fashion I have ever since companies like Ubisoft started promising a future with AI-powered "neo-NPCs."

If you'd believe the hype, these things are potentially good on paper, right? We live in a development environment where the ambitions of big-budget open world games are far outstripping their cost—wouldn't it simply be marvellous if you could populate a world with infinite dialogue chains and storylines?

In practice, the result is… pretty middling. While yes, some of the isolated tech demos we've seen have been impressive, the actual implementation of those we've seen in videogames, such as the life sim Seed, or Inzoi, which tried to pilot "smart Zois", fell flat.

Even when they're a source of entertainment, that fun is mostly found in the process of making them believe they're pregnant—and listen, I love toys. Putting aside my general distaste for generative AI, I don't think that every single videogame that comes out has to be a work of art. Toys are fine, you can smash action figures together and have a lovely time.

But that's not been the pitch, has it? Convai, an AI NPC start-up, wants to avoid a world without games where background "characters are given one line and repeat that one line over and over again, maybe two lines at best."

Ubisoft wants "fully personalised and immersive experiences", where its neo-NPCs make games "more alive and richer," whatever that means. Meaning Machine, a company that's pioneering reactive battle banter, is pushing for narrative-based games which are entirely powered by the tech. The dream is, in a lot of places, to use AI to try and tell stories.

The problem is, this pipe dream fails at the conceptual level, because this just isn't how good storytelling works. These people don't get it. And if they do get it? They think you're dumb. They think you're dumb as bricks.

In defense of miscellany

The main lie to the idea is the implication that background details, specifically what NPCs in said background say, aren't of any great importance. This couldn't be further from the truth.

One of videogames' biggest strengths is the ability to divorce themselves from pacing. Not entirely, mind—a singleplayer game with a linear narrative is still going to care about pacing a great deal. But an open-world game can take the time to establish, create, and put forward background details that would otherwise bog a story in any other medium down.

There are more obvious examples. Say, a town crier in a fantasy RPG telling you about some far-flung event you'll likely be dealing with—but there are more subtle instances of this sort of thing working brilliantly. My favourite example by far are Cyberpunk 2077's adverts, which are tailor-made to invoke the exact right amount of corporate overstimulation and dread in you as you muddle around Night City.

In a game with a stand-out story, all of these hundreds of little opportunities will be leveraged powerfully to draw you in—they can do more than just hand you information about the plot, they can create a sense of scale or place. Combat barks can characterise an enemy faction, journal entries can flesh out a culture, and so on.

The perception that these qualities in a game's overarching script are somehow "less important" is just completely backwards—they're the backbone on which the rest of the story lies. And even when they're not exactly award-winning, this background chatter still creates a game's sense of character.

Imagine a world in which lines like "Do you get up to the cloud district very often? Oh what am I saying, of course you don't " or "I used to be an adventurer like you, but then I took an arrow to the knee" had been scraped out and replaced with the ability to, like, make them think you're Abraham Lincoln reincarnated, or something.

Heck, the fact that you're probably summoning to mind videogames with hilarious or immersion-breaking background natter proves my point, right? You're already able to tell when it's gone wrong.

Starfield's setting feels bland and dull when compared to something like New Vegas not just because of its main yarn, but because in the latter, its locations and background babble are hand-crafted and designed to surprise you, and in the former, procedurally-generated planets lie next to grey and perfunctory cities with all the rough edges sanded off.

Even just being placed next to a bunch of crap that wasn't hand-made causes Starfield to feel shallower as a result. Now imagine if all of those cities had AI-generated NPCs with disconnected, jilted, random stories to tell you—stories you couldn't even have a shared laugh at with your mate when they were bad. Does that sound better?

In this sense, the idea that AI-generated NPCs are some miracle fix that will give smaller studios the ability to tell stories in bigger worlds just completely falls apart, because those worlds are only made to feel big and textured because of thousands of little scripts that are laboriously written out by hand.

At the risk of stating the obvious, you can't just skip over that hard work and expect the same results! There's far more creativity in all the smaller choices a team of writers makes than the broad, conceptual idea that started them off. Background chatter, combat barks, NPC dialogue, side-quests, it all matters.

They think you can't tell the difference

The reason these bigwigs think they can get this crap past you has a lot to do with the same, broader, executive misunderstandings of what makes good art: People. It's a sentiment that was echoed by id Software's baffled Chris Hays, who has been watching Microsoft push statement after statement that the studio was still the same id Software, even after half of its staff were laid off.

Or take EA's head honchos, who Dragon Age creator David Gaider said believed "you made an RPG and the nerds in the cave would always show up for an RPG, because it was an RPG." It's no coincidence that Gaider, ostensibly responsible for one of gaming's best fantasy settings, was pre-emptively exhausted at the idea of AI-generated story back in 2023.

Nor is it a coincidence that Cyberpunk's quest director was unimpressed by the technology, or Larian's head writer, who called its output worse than their worst drafts. People who know what actually makes a story good aren't cringing at generative AI because they're luddites whose jobs are being threatened—if it was able to put something compelling together, they'd spot it.

They think it's garbage because it's garbage, and they don't want to have a hand in feeding you garbage. It's really that simple.

The biggest insult of suits peddling this crap is the idea that you just don't care. That you don't care if the music you're listening to is crap, you don't care if their games look hideous, and that you don't care if a game's story leaves you confused and hollow. They believe that you will merrily accept the substitute of a bunch of keys being jangled in front of your face over something that was carefully hand-crafted to make you feel something.

To those people, videogames are products. Hasbro thinks it can spit out another Baldur's Gate just by throwing the same amount of money at a different studio, Amazon's then-boss of games blithely said, in 2024, that videogames don't have acting. This is the prevailing attitude—a row of artless suits who are trying to sell you a bridge. Don't buy the damn bridge, I beg you. It won't make your games better.

One of the unexpected stories this week has been that of a rather public spat between former Saber Interactive writer Stella Sacco, who claimed she was replaced in favour of AI, and the publisher's CEO, who tore into her via a statement sent to This Week in Videogames.

"Stella who?" asked Matt Karch, "I had to ask Claude because I have never spoken to her or seen her. We have already said we are using AI for certain parts of the game and will update the Steam page once we have finalized the feature set."

Karch also added that "it was always our intention to reveal this until Stella violated her confidentiality clause and revealed this before we were ready." Sacco, who claimed she was replaced "with ChatGPT" in 2023, called the statement "remarkably pathetic for the CEO of any company to be making."

You can read both parties' full statements in the articles linked above—the most recent development is that Saber Interactive has been seemingly pressed into adding a Steam AI disclaimer to the game's store page, which is present at the time of writing, but was notably not present when the game was put on the store on July 30.

"Once it's finalised we will update the Steam page," Karch told IGN. "In Rideshare we added an unrelated experimental mode to the game to play around with these concepts. It's not the core experience. And even in this case it required us adding two writers and multiple coders to work with AI tools to try and get a reasonable result. Maybe we are bad at business but the net cost of incorporating these features was significantly higher than leaving them out."

Karch then goes on to say that "my PR team is always terrified by even broaching the subject but I would be happy to engage in a reasonable conversation about how AI can and should help shape gamer experiences."

The disclaimers in question make reference to not just these AI-generated dialogue, but also "a mix of radio stations featuring human-made music and AI-generated music. Stations containing AI-generated music are clearly labeled as such in the game."

Karch chalks up the entire debacle to a "publicity stunt… This person was removed from the project three years ago. We have actually increased our writing staff since 2023. If I wanted publicity for this little game though, then I guess I have Stella to thank for it."

The entire situation, to put it extremely lightly, is a bit of a mess—regardless of whether or not Saber's landed itself in troubled waters (Sacco seems to think so), if I were a videogame publisher, I might ask myself why I felt the need to wait to be "ready" to reveal something core about how my game was made, and why I might be making my PR team "terrified".

Could it be that AI is broadly unpopular with the gamers Saber Interactive is claiming to serve? Surely not.

There's a problem with new technologies, in that cutting-edge technology develops quickly—in the span of years. Meanwhile, us human beings with our wet flesh bodies and our complicated bureaucracies and legal systems are glacially slow. You can see this playing out with generative AI.

For a few years now, we've heard a lot of overexcitement from the videogame industry that AI is going to let smaller companies make bigger games, that soon you'll be able to generate your own concept art, your own music, your own videogame dialogue—and, of course, the deeply tedious phrase "democratising art".

In actuality, generative AI is more of a plagiarism machine, often spitting out carbon copies and, even when it's not doing that, still requires countless copyrighted works (which sometimes have to literally be burned) to actually function. That legal heat has spooked videogame companies, so says lawyer Haley MacLean in an interview with our friends over at GamesRadar+.

MacLean, both a corporate IP lawyer and a videogame specialist at Voyer Law, tells the site that no-AI clauses are on the rise, particularly in the last year: "It's turned around, especially in the past year. I would say about two to three-ish years ago, you'd see a little bit of it … Even the smaller, more simple publishing studios are like, 'Let's include this just to cover our butts.'

"So now it's just boilerplate, to the point that if I did encounter one that didn't have a gen AI clause, if the publisher was creating assets too, I would say we need to push for that to make sure the publisher is not using gen AI in the marketing, porting, QA of your game."

It's not all too surprising, even from a ruthless capitalist perspective. There's a tension between videogame platforms, who are incentivised to rid themselves of the tides of AI slop lapping at their shores, and developers themselves. The former is encouraged to add large, obvious disclaimers so their users can filter out the garbage, while the latter is utterly screwed when they do.

Back in June, for instance, a data analyst found out that AI disclosures could reduce the number of reviews a game got by 53%, making the reviews more negative overall, which MacLean says causes a pervasive attitude of "I don't want to use gen AI because my community would hate it."

But the legal quagmire is also becoming increasingly insurmountable as legislation catches up—one particular quirk? If you use generative AI to make something, it might be impossible to prove that you copied the thousands of artists who were fed into your algorithm, sure. But it also means that, legally speaking, you can't really argue YOU made it, either.

"Sometimes I feel like I'm going crazy when I see these massive corporations … It kind of feels like a lawyer is hovering slightly over them, because they'll say things like, 'Look at this awesome tool. I can just type in 'barrel' and a barrel appears. But I have to make sure to change it a little bit!' They sort of imply that the users should be transforming every generated asset a bit, but they know that their users aren't going to do that."

MacLean uses the example of the Tomb Raider: Legacy of Atlantis remake, which uses generative AI for its assets—the game's not out yet, so it's unclear if it'll meet this bar, but if it doesn't change those assets enough? "[if there was] an outdoor level and all the trees are AI-generated, I legally would argue somebody could go and use every single tree in the Lara Croft game, that they know is AI-generated, and put in their game, and I don't think those devs can turn around and say, 'Those are our trees.'

"No, they're not, because you didn't generate them, and legally speaking, you have no ownership title over them if you didn't put in the sweat equity to have an artist go in and, inspired by that original first pass of the tree, then transform it into an original tree asset … But I think these major companies who are always laying off, trying to cut the bottom line, increase the value of the company, they're going to cut corners when they can, and I think they're going to get sloppy."

It's sort of a 'snake eats its own tail' situation. In order to fulfil the promise of generative AI—that you could lay off a bunch of artists and just have your hallucination machine dream something up—you need to lower your headcount. But your remaining artists then need to meaningfully transform what AI generates, in order to legally lay claim to it.

And if there's not only fewer of them, but they also kinda resent you for laying off all their colleagues, then that increases the chance of legal trouble happening—or lost sales from just pissing your consumer base off. At which point, you might not even be saving the money that a handful of yearly salaries (which are pennies to the absolute billions these companies have) saves you by way of their absence.

"I think," MacLean says, "We're going to see lawsuits of those types in the coming years when they want their protection but they want to do everything faster and cheaper … What case law in the US and Canada has seen, time and time again, is that AI cannot own because it was not human-made.

"We're in that frustrating middle with new tech where it's very fun, it's very flashy, everyone's excited, but I think the legal issues are going to just be popping up more and more."

Original Story: After noticing the announcement of upcoming driving sim Rideshare Stimulator, writer Stella Sacco commented on social media that she'd once been the game's lead writer but had been fired in favor of ChatGPT. The game's publisher, Saber Interactive, denied the claim, stating that while it did use generative AI in the development of Rideshare Stimulator (which has not yet been disclosed on its Steam page), the story was written "entirely by real people."

Sacco acknowledged that the studio may have changed direction since she was removed from the project in 2023, but stood by her recollection. That could have been the end of it, but now a confusing, combative follow-up from Saber CEO Matt Karch is raising eyebrows.

In a statement sent to This Week in Video Games, Karch claimed that Sacco violated a confidentiality agreement by stating that she was fired in favor of AI, but also denied the assertion, crediting Rideshare Stimulator's "entire story script" to an unnamed writer.

Karch also stated that he didn't know who Sacco was until he consulted an AI chatbot, but nevertheless claimed that she was replaced by "someone more talented" and that he would retrospectively be "happy to replace her with AI" because AI is "programmed to be honest."

Here's Karch's full statement:

"Stella who? I had to ask Claude because I have never spoken to her or seen her. We have already said we are using AI for certain parts of the game and will update the Steam page once we have finalized the feature set.

"Let me clarify that the entire story script, were it to win an award, would go to someone who can walk up on stage and receive it. It was always our intention to reveal this until Stella violated her confidentiality clause and revealed this before we were ready.

"We look at our use of AI in this particular game, which will be a lower priced title, as a plus for consumers.

"As far as Stella is concerned, she was replaced by someone more talented and terminated for a variety of reasons a few months later. From what I can gather she enjoys controversy even more than I do. I would frankly in retrospect have been happy to replace her with AI. At least we would be dealing with someone programmed to be honest."

Sacco responded on Bluesky, standing by her original claim.

"I've seen the statement by Matt Karch, the CEO of Saber Entertainment, and all of the truly horrendous things he said about me. I think it should be obvious to anyone who reads that statement that he's lashing out because he did not want people to know there was AI use in the game," said the writer.

"Everything I said was true. Here's more truth: that statement is remarkably pathetic for the CEO of any company to be making, and combined with the AI use it will do far more damage to Saber's reputation than it ever could mine. This screed will be many people's first exposure to the Saber brand."

First reported by Tom's Hardware, it's come to light that, not only are the four "Backward Compatibility on PC" original Xbox games released last week running on a double layer of Xbox console emulation, but savvy independent developers have already used those emulators to run other Xbox games on PC.

Developer Nathan on X, "The Everything App," has a succinct topline explanation: "The new OG Xbox Backwards Compatibility layer released by Microsoft for the PC is actually an Xbox 360 emulator running XeFu (the OG Xbox emu for the 360)," he wrote, pointing to telltale indicators visible in a dump of XeFu (the Xeon Fusion Emulator) pulled from the Fuzion Frenzy PC re-release.

For an even more in-depth explanation, developer Milenko has an article on the subject, also on X. The big takeaway? It's a "Beautiful piece of engineering. Also: Emulators, all the way down."

Rewind all the way back to 2005: The very first instance of Xbox backwards compatibility on the 360 was XeFu emulation, as outlined in this 2018 video from Modern Vintage Gamer. When you popped your old Halo 2 or KotOR disk into a 360 back in the day, they weren't running natively on the console. The 360 emulated the original Xbox to run those games.

This tech is the basis of Microsoft's later backwards compatibility efforts on Xbox Series consoles as well. So instead of building a new Xbox emulator from scratch to run these games on PC, the company ported its in-house Xbox 360 emulator to PC, then ran XeFu through that. Pretty neat.

Meanwhile, the devs of the XWine1 project, an Xbox One translation layer for PC, shared that it's possible to use the in-development XWine1 with the dumped official emulator to get other OG Xbox and 360 games running on PC, though with "varying degrees of success."

As for how this will affect game preservation and end-user emulation quality in the long term, that remains to be seen. The xemu OG Xbox and Xenia 360 emulators have been in development and available for use for some time, but better public understanding of Microsoft's internal emulation methods can only be a good thing.

1,474,560 bytes may sound like a lot for those unfamiliar with the realities of modern data storage, but it amounts to a single 3.5-inch floppy disk. The hey-day of such a relic of physical media has long since passed, but floppy disks obviously still hold a certain allure for some judging by a certain Korean game dev contest.

Anyone can participate in the contest from Korean indie game site 2P Game Arcade, and the rules are very simple: devs just need to build a game that totals no more than 1,474,560 bytes (or 1.44 MB) after decompression (via Techspot).

Though there are no restrictions on which engine devs may use, the work itself must be an original creation made after the contest announcement date on May 26. The teeny-tiny game must also be playable as a standalone executable (browser-based games are not eligible). Devs have until midnight on September 4 to turn in their work.

There are prizes to be won, though a far cry from the $80,000 the EVE Online devs recently awarded to its own mod contest winners. The machine translation of the 2P Game Arcade contest page describes these prizes as only 'a small souvenir,' with first place taking home 720,000 won (about $490).

The judging criteria is just as straightforward as the rules: "1. Whether the game is complete. 2. Whether the capacity has been exceeded," and perhaps most importantly, "3. Whether the game is fun."

Considering how chunky modern game files are (not to mention what a premium system storage comes at these days), the whole premise of the contest is pretty refreshing. That said, 1.44 MB is barely enough room to swing around a pixel-based sprite of a cat. In the immortal words of Ru Paul, I can't wait to see how this turns out.

Unity has demonstrated a Unity Engine game running inside of an Unreal Engine game, marking a step in the direction of Epic Games CEO Tim Sweeney's vision for an interconnected gaming ecosystem.

The demonstration took place this week at the Unite Seoul keynote, where Unity also announced the next version of its game engine. Unity's goal with the project, in partnership with Epic, is to allow Unity developers to publish their games in Fortnite.

"You will see, for the first time, a Unity game rendering natively inside Unreal Engine," said Unity senior director of technical product management James Stone as he launched the demo. "Now this is not live in Fortnite yet, but it shows what's possible for a shared foundation.

"What you're seeing here is a game simulated in Unity, rendered with Unreal. Now, that's not video streamed, there's no smoke and mirrors. This is all real time. But it's not just rendering. Watch the avatar. That's an Unreal character running through a Unity world. That's physics, lighting, input: all of it synchronized between engines."

"Connect ALL the engines!" Sweeney exclaimed on social media in response to the demo.

The Unity partnership represents one of many avenues through which Sweeney and Epic are pursuing a future that connects content, economies, and social features across platforms and games—"the metaverse," for short. Sweeney discussed the effort in a recent conversation with PC Gamer, saying that the industry "should come to the realization that the value of connecting is higher for everybody than remaining separate."

Unity's new Fortnite integration will be available in early access in 2027. Unity 7, the next version of the engine, is also coming next year. Unity 7 promises new rendering, monetization, and collaboration features, as well as a protocol for connecting generative AI coding agents.

It's been six years since Epic first launched Unreal Engine 5, and at its State of Unreal event in Chicago today (head to 1h 40 mins in the above video) and in a separate blog post, we got our first look at what the next major release will offer to games, film, TV, and more industries. And it turns out that the biggest direction for the changes in the new engine came from Fortnite and UEFN.

If you're unfamiliar with the latter, it's basically a version of Unreal Engine 5 that's somewhat simplified and pared back to allow anyone to make levels or entire games for Fortnite. It's hugely popular and honestly very simple to use. I introduced my partner to it a few months ago, and despite having no experience in game development whatsoever, she created a fully functional Fortnite map and game mode within a day.

The sheer usability of UEFN is core to the changes in Unreal Engine 6, and when I talked to Epic at last year's State of Unreal event in Orlando, it explained that the separation between UE and UEFN would eventually go, with the two combined into a single package and offering the best of both worlds.

In the words of Tim Sweeney, it's "UE 5 plus UEFN equals UE 6, plus some more cool stuff on the way." The idea behind the merge is to allow developers to create something and then ship across every possible platform/store at the same time, including Fortnite itself. The unification also involves bringing APIs and code together across all of the various additions that Epic has for UE, such as MetaHumans.

That said, Fortnite isn't really going to be the showcase for Unreal Engine 6; that honour goes to Rocket League, and the very first glimpse of it all was dropped last month at the Paris Major event of the RL Championship Series.

Perhaps the most significant change in Unreal Engine 6 is the move to open standards for tools, code, APIs, etc. This isn't something that can be wholesale implemented overnight, and I suspect that not every element will be open in this manner, but the end goal is to give developers (games or otherwise) an easier path to getting content and code out to Epic's and external ecosystems.

Epic didn't say anything about specific features in the keynote, so no idea if there will be a major change to Lumen, for example, but it did say that UE6 is targeted for release at some point in 2027 ("2027-ish" was the exact phrase, though the blog post says "Early Access release at the end of 2027").

It also said a few things about Verse, the scripting language used in UEFN, and how the gameplay programming model in UE6 will be shifted to that language (though C++ will still be there underneath it all). Directly related to that will be something called Scene Graph, which will replace the current gameplay framework used in UE5. That will be built entirely on Verse, and with it, Epic plans to "build a full distributed software transactional memory system" for huge, interactive live worlds.

That won't mean anything to gamers, but for developers, it means that their "game code can be written as if it were running on a single machine without needing to coordinate custom networking code all over the place."

And naturally, AI is going to become a bigger feature of Unreal Engine, with the first step being taken in the new UE5.8 release and its MCP server plugin. That system will let you set up any LLM you want to use and give it various tasks to perform, from simple code refactoring all the way up to generating a full 3D scene that you can then tweak as required.

There's more to all of this, of course, but there's not quite enough information yet to glean any insight as to whether the step from UE5 to UE6 will be as dramatic as the jump from v4 to v5 was. My feeling is that it won't be, from a PC gamer's perspective at least, but game developers will probably relish the thought of having an Unreal Engine that's easier to use and quicker at producing the content required.

It's probably fair to say that Unreal Engine is probably the most comprehensive tool around for creating games, animations, and video effects. And, with the launch of UE 5.8, it's becoming even more extensive and even a little bit more AI-friendly. Whether you're just an Unreal Engine hobbyist like me, or a full-time game developer, you can download version 5.8 right now via the Epic Games Launcher.

Even if you don't plan on using any of the new features (of which there are a lot), it's always worth trying out the latest release just for bug and performance fixes. But what's actually new? The headline acts in UE5.8 are the introduction of Mesh Terrain and Procedural Vegetation Editor (PVE). Both are experimental features at the moment, but the former should be of great interest to anyone creating big, open-world terrains, as the tool basically generates full 3D meshes for you.

PVE is somewhat similar, except that instead of creating landscapes, it produces vegetation (trees, bushes, reeds, grass, etc) from scratch, with the procedural system working in line with meshes already present in the world. For example, let's say you had a crumbled old archway in a forest, PVE will 'grow' trees around it, accounting for the natural source of light and competing vegetation.

Complementing these are MegaLights, which first appeared in experimental form in UE5.5 (but is now "production-ready"), an experimental fog screen space scattering feature, and perhaps most interesting of all, Lumen Lite. This is a mode for Lumen global illumination that Epic claims to be twice as fast as Lumen High Quality, while still preserving "much of the visual impact".

The release notes for UE5.8 specifically mention that "games that rely on global illumination for artistic purposes can run on Nintendo Switch 2 at 60 fps," so it's blatantly obvious what platform it was developed for. However, since it's also supported on PC, UE-powered games of the future could well offer this as a graphics option for low-end hardware users.

Tucked away, almost at the end of the new version details announcement—which blessedly includes improvements to shader compiling—is one more experimental feature, an MCP plugin. With this, you can implement any LLM of your choice to "connect to and understand both the engine and your project." In other words, if you want to use AI to create assets or code, carry out tests or refactoring tasks, then you should be able to hand that over to the LLM easily enough.

Epic left the note about this feature after everything else, and I suspect that's because AI is hardly flavour of the month in the world of PCs and gaming right now. After all, Epic already has a section of PC gaming fandom that takes a dim view of Unreal Engine games, whether you believe that's the fault of Epic or a given game's developers themselves, and AI-use is another PC gaming bug bear.

Anyway, because of the consternation about AI in gaming—from being the primary cause of the horrendous price increases for DRAM and SSDs, as well as the endless controversies over its use in games—the inclusion of this plugin is likely to draw ire from some quarters.

The thing is, many game studios will have to rely increasingly more on AI for certain workloads if they hope to stay afloat. With the likes of Microsoft and numerous other companies about to swing a sword of Damocles across all their gaming divisions, studio heads will be looking at every avenue that will result in them having a future.

I'm not suggesting that game devs must use AI, nor am I saying that the inclusion of the LLM plugin (specifically an MCP server) with UE5.8 is a good or bad thing; it's simply a sign of our times. One can argue that Unreal Engine's feature set is sometimes a little too far-forward in reach (console and mainstream PC hardware still isn't quite good enough to cope with an all-in Lumen and Nanite game at high fps), but this minor plugin is very much a 'here-and-now' thing.

Although many developers turn to Unreal Engine or Unity to bring their gaming visions to life, open-source Godot is becoming increasingly popular, thanks to its lightweight structure and intuitive scripting language. Now Microsoft has taken a 'first step' to expanding its user base even further, by releasing a reference sample software package for creating Godot games for Xbox on PC.

Yes, I know I should be writing XBOX on PC, given that Microsoft has decided to settle on the format, but I'll be writing Xbox until I'm in the grave. Anyway, nomenclature aside, it should be noted that Microsoft hasn't released an 'Xbox-version of Godot' or anything like that.

Instead, it's essentially a demonstration of how to implement all of the necessary Xbox on PC requirements (such as Xbox services, PlayFab, and GameInput) in a Godot-based game.

"This sample shows you how to bring sign-in, gamepad controller support capabilities and more into your game with Godot. It includes insights, re-usable integration code, and is currently compatible with the April 2026 GDK out of the box," writes Microsoft in its dev blog post.

Although Godot has been traditionally associated with 2D indie games, such as Brotato, Slay the Spire 2, and numerous cozy games, it does handle 3D graphics extremely well, and there's a decent spread of things being developed that sport visuals that wouldn't look out of place in anything made with Unity or Unreal Engine (full-blown ray-tracing aside).

I've dabbled with UE for many years, going all the way back to the late 1990s with its first iteration, having first dipped my toe into it all with UnrealEd and custom maps for Unreal Tournament. These days, Unreal Engine is a massively complex and capable package, used for games, TV shows, and more, but while it's relatively straightforward to get started with, mastering it is on a whole different level.

By having a much smaller remit, Godot is arguably easier to work with, although it's perhaps still not quite as novice-friendly as Unity. But since it's also entirely free and open-source, any revenue your game makes is 100% yours. Well, yours and the publisher's. Oh, and the platform that you sell it on, too.

Microsoft says that its Godot sample is a "source-only sample" and "not a product", as well as purely for PC at this stage, and that there is "no set update cadence for support or maintenance." That doesn't sound overly encouraging, I grant you, but the software giant does promise that "This is the first step in bringing Godot for Xbox on PC. We plan to evolve it over time based on what the community tells us is most valuable."

I suspect we won't see a full Xbox on PC version of Godot any time soon, or perhaps even ever, but if Microsoft really is serious about saving its Xbox brand, more Godot samples like this one would be a step in the right direction.

For well over a decade, Unity has been one of, if not the primary engine for indie game development, and has powered some massively successful games like Hollow Knight, Among Us, and the original Subnautica. But over the last few years, a growing number of developers have switched to Godot, an open-source engine that is similar to Unity in its scope and functionality.

For those developers who have switched, the reasons have often been ethical or financial, with various shenanigans from Unity putting developers off using the company's tech. But placing that aside, how do the two engines compare as tools? One game developer decided to find out, by making exactly the same game in both engines.

That designer is Thomas Grové, who runs the Japanese co-production developer Studio Interrupt. Grové was making a survival horror game with his son, and decided to use the project to test out both engines.

"For a long time, I've really wanted to see a side-by-side, apples-to-apples comparison between the two engines, not like the same game made by different developers or whatever," Grové says in a YouTube video discussing the experiment (via Automaton). "I thought this was my chance to finally do that and decide whether I want to move forward with Godot or keep using Unity."

Grové starts out by listing the features of his game, which is clearly at a very early stage of development. Said features include a complete character controller, a camera transition system, a scene-transition system, a tri-planar dither shader and an interactable object system. He built and ran this game with all these features in both engines, before comparing how each engine fared during the project.

The results were intriguing. From a functionality perspective, Grové explained that the engines performed similarly, with each doing some minor things slightly better or worse than the other. However, he concluded that Godot was significantly more efficient in areas like compiling, launching, loading etc.

He notes that Godot is over five times faster at loading a project, 20 times faster at exporting a project, and a whopping 31 times faster at compiling a script, which could make a big difference considering you would perform some of these tasks hundreds of times when designing a game. On top of that, Godot is also a significantly smaller program in terms of file size, taking up just 164 megabytes compared to Unity's 20 gigabytes.

Ultimately, Grové says that "If we look at all the data, Godot beat Unity on every metric except for the final output FPS", with both exported projects running at a max framerate way about Grove's minimum of 60FPS anyway. As such, he concludes that "I'm probably going to continue using Godot for this project."

While this seems like a slam dunk for Godot, some of Grove's viewers point out that the experiment may not be entirely reflective of each engine's performance. "The scene was just too simple to stress any systems, as evidenced by the very high framerate," writes YouTube user WitchfellGame, while the startlingly named jakegenocide says "If you aren't using a full, complete, game project, you aren't going to know which one is right for your project once it scales up."

Building two entire games from scratch just to test the performance of game engines isn't especially viable, however. And while maybe you could stress test further, the consistency and scale of difference in Grové's results do suggest that, broadly, Godot is the more efficient engine of the two.

These days, I'm firmly on the System Shock side of the System Shock/BioShock division that exists almost exclusively in my head, but I'll give Rapture this: it was goddamn stunning when I first took that bathysphere jaunt back in 2007. The barnacled art deco was a heck of a look, and it still is today. I can vouch for that: I replayed BioShock 1 and 2 just a couple years ago on my trusty Steam Deck.

In a recent chat with IGN, Irrational Games vozhd and 'the narrative Legos guy' Ken Levine said there's a reason for that: he's never been one to push hyper-real graphics tech to its limit. "I don't think we've ever been a company that was like, 'Oh my god, we need the latest and greatest technology,'" said Levine. "Outside of SWAT 4, we never really tried to do ultra-realism in our games."

Part of that is because, well, cramming the latest whizzbang rendering stuff in your game is pricey, but there's an artistic benefit too. "It doesn't age as well as more stylistic things. BioShock still looks good, I think because it wasn't trying to get every nut and bolt [looking] super realistic." It was trying to get every nut and bolt looking wet and gross, which is an objectively superior approach so far as I'm concerned.

It sounds like, in Levine's estimation, the endless chase of higher-fidelity graphics is a bit of a fool's errand, at least for his own studio. "Look at, say, the Switch 2, and the new Steam Machine coming out. Those are not massive technological upgrades. That wasn't their strategy. I think people are realising we're hitting a bit of diminishing returns with that."

What is the right approach? Great question. "I think if you have right art director, and the right approach, you don't need to be on the cutting edge of technology all the time, and even the stuff we're doing in Judas—all this narrative stuff we're doing—it's not CPU-intensive, it's work-intensive on our side."

I can't argue with him too much. Ultimately, you probably do need someone out there riding the bleeding edge, as that's how technology in general gets moved forward and developers like Levine and co get access to new tools they actually do want to use. But you certainly don't need everyone trying to fit path tracing and DLSS 8 billion into their visual novel about a cool horse (or something).

Heck, the most visually arresting game I've played recently was Metaphor: ReFantazio, not because of the tech (which seems barely different from Persona 5's), but because it feels like everyone was forbidden from saying no to the art team for the duration of that game's development. If that doesn't illustrate Levine's point, I don't know what does.

The Playdate is a curious little device. It's a handheld console with a crank built into the side, which developers can make use of for in-game controls. What devs can't do anymore, though, is get their games accepted onto the Playdate Catalog store if they use generative AI "for art, audio, music, text, or dialog."

So says an announcement on the Catalog developer blog. Panic Inc, the developer of the device, defines generative AI as the use of:

• Large language models (LLMs) such as ChatGPT, DeepSeek, Google Gemini
• AI-based image generation models such as Stable Diffusion, DALL-E, Midjourney
• AI-base audio generation models such as MuseNet, Suno, Udio, and others

The announcement clarifies that games using AI in their development will no longer be accepted, but custom-written functions like enemy AI are fine. It also says that games already listed on the Catalog store that have previously used AI will remain, but they will be "flagged as such", with explanations provided as to exactly how generative AI was implemented.

"There are lots of people in the world who are extremely eager and excited to create beautiful music, art, and words for your Playdate creation," the statement opines.

"If you need any help at all finding people to help with any parts of your title, you can check out the collaborations channel in the Playdate Squad Discord or the Playdate developer forum."

The use of generative AI in games is a hot topic and sure to bring up some strong opinions. While a survey at the start of the year claimed that 50% of game developers think its inclusion is bad for the industry, some investors seem to think otherwise.

It must be a tempting option for the developers of small games for a relatively niche device, though, where budget and time constraints are pushed to their limits. Still, it's a no-go for gen AI-enhanced game development on the console-with-a-crank. The wheels, they keep turning. Something like that, anyway.

Gabe Newell remains the CEO of Valve, but it's no secret that he stepped back from actively running the company some time ago: he says himself he's been "retired for a long time." Newell now divides his days between various other pursuits, including Starfish Neuroscience, a company he co-founded that focuses on neural interfaces, and Inkfish, a marine research operation. As well as buying $500 million superyachts.

I've always been a little fuzzy about exactly when Newell stepped back from the game design side of things, not least because Valve itself often credits staff equally, though he was clearly intimately involved in the design of Half-Life and its sequel, as well as spearheading the company's acquisition of fan-made properties like Counter-Strike and Team Fortress. But an interview with Portal 2 project lead Josh Weier by Kiwi Talkz, recently resurfaced by PCG sister site GR+, gives a rough idea of the timeline for Newell moving away from being hands-on with development: and the main reason why.

"If you haven't met him, Gabe is a very imposing guy," says Weier. "He's literally just a tall dude, and also his mannerisms can be imposing. So when I'm like 20 on [Half-Life 2] and he's taking me out for lunch, giving me game ideas, I'm like [mimes eating and looking at the ground] 'Yes Mr Newell!'"

Weier laughs: "Head down freaking out! And I think over time Gabe realised 'oh OK, I should let the game teams do what I pay them for' and he would stop in and give us feedback. I think his things were, he was worried about… we redesigned GLaDOS for Portal 2, because she was going to have a bigger role and felt like the old model wasn't going to quite give us what we needed. He was very adamant about us giving GLaDOS a face and we were 'ah don't really think so, we're gonna go a different direction.'

"So we kind of went back-and-forth with him a bit, and then after a while he was like 'alright, you guys got it, I'm just gonna leave you to it.' And he would check in with me personally to say 'how you doing, are you stressed, how's life going?'"

The real problem from Newell's perspective though was that he'd become almost too forbidding a presence: even the kind of top-tier developer that works at Valve couldn't quite bring themselves to push back hard enough.

"At that point I think he was realising… I think he always wanted to be part of the team, but being Gabe and being in his position, that never really worked," says Weier. "Because people would be like, 'whatever you say!' And he was more 'no no no, I want to be part of the team and come up with ideas.'

"And that was really hard for people, so I think there was a period where he stepped back and was like, 'Alright I guess I'm just not going to be able to interact with everyone that way and I'll just work at a higher level', and that was alright.'"

Portal 2 began development after the release of the original (2007) and released in 2011. While Valve had always been a successful company, this was also roughly the time period where Steam was exploding in popularity and beginning to truly dominate the PC market, so obviously the company's priorities were shifting from being solely a game development studio.

The danger, which Newell clearly spotted and avoided, was him becoming too dominant a creative presence for Valve, and ending up in a position where he was cramping the ideas of the company's game developers rather than helping to refine and realise them. On the one hand that's sad for obvious reasons, but on the other Newell obviously found many other outlets for his generational talents.

A lot of my favourite games ever made fall under the expansive rubric of "eurojank." Stalker, the first Witcher, Mount & Blade—the list goes on. But maybe it's a little unfair to confine the jank phenomenon exclusively to Europe. After all, what is 2004 RPG Vampire: The Masquerade – Bloodlines, which is still getting fan patches 22 years later, if not a prime cut of jank?

So says Andrii Verpakhovskyi, designer on the original Stalker games, in a recent interview in Edge magazine. "Some of my favourite games back in the day were Vampire: The Masquerade – Bloodlines and Arcanum, both built by Troika Games, which was the core team [from the] Fallout games before them.

"Those games were janky as hell, but they had this same soul that was in games described as Eurojank, which is why I think it’s unfair to kind of geofence the genre." The man has a point!

Eurojank, as a pseudo-genre category, refers to a bunch of systems-heavy and ambitious games that tend to originate in Central and Eastern Europe. They often do things that no other game you've ever played even tries, but—owing to a lack of resources—they're also often crammed to bursting with weird and hilarious bugs.

Back when Verpakhovskyi and co were working on those original Stalkers, they didn't really think of themselves as making a particularly regional sort of game. "We had absolutely no notion that what we were doing we were doing differently from people building games outside of the post-USSR space. We never even drew any kind of lines between Japanese games of Nintendo fame or Sega Genesis, and the games that came out of the US, Canada, the UK and western Europe."

They were just devs making games and experimenting along the way. "We were all newcomers to the industry and people were not even specifically trained in engineering or arts." Which perhaps explains some of Stalker's myriad quirks better than the place it came from.

And while you and I might use jank as a loving term, to describe those games with lofty systemic ambitions that end up butting against the limits of a developer's budget or technological capacity, it can sometimes be an insult used by those who lack vision. "I feel like Steam especially now has smothered creativity with how rigidly titles need to fit into a specific orthodoxy," Greg Pryjmachuk, founder of Minskworks and dev behind Jalopy, told Edge. "Anything that doesn’t colour within the established genre lines is quickly punished."

Not here, mind you. At launch, 2024's Stalker 2 kept aloft the janky banner of its forebears. I loved it, and wrote as much in our Stalker 2 review.

Higher education is one of the bigger choices you'll ever make in life, and finding your path is not as simple as choosing an RPG class. When it comes to a career in game design, every journey starts at the same place: picking the right university. Let this be your first stop.

Each year, the Princeton Review ranks the 50 best undergraduate and 25 best graduate programs in the United States and beyond, using a thorough set of criteria that includes graduate employment rates, early-career earnings, faculty industry experience, curriculum relevance, and demonstrated student success in the games industry.

The list spans the best of the best programs in the country, so no matter where you land, you'll be one stop closer to your dream career. To get you into the mindset of a game designer, we've also included homework projects that explore common challenges faced by developers today, as well as a discussion on the role of AI programmers and map designers in game development.

Want to know more? Check out The Princeton Review's website for further info: www.princetonreview.com/game-design

UNDERGRADUATE RANKINGS

1. New York University

2025 Grads Hired: 53%
2025 Grads Mean Salary: $84,800
Faculty: Eric Zimmerman (Gamelab), Dr. Mitu Khandaker (Spirit AI)
Graduates: Jenny Jiao-Hsia (IGF), Carol Mertz (Exploding Kittens)

2. University of Southern California

2025 Grads Hired: 63%
2025 Grads Mean Salary: $107,400
Faculty: TreaAndrea Russworm (Author, "Gaming Representation"), Mark Bolas (Founder, Fakespace Labs)
Graduates: Jenova Chen (Thatgamecompany), Brent Strong (Disney)

3. University of Central Florida

2025 Grads Hired: 36%
2025 Grads Mean Salary: $48,500
Faculty: Dr. Peter Smith, Dr. John Murray
Graduates: Richard Ugarte (Epic), Alexia Mandeville (Niantic)

4. Rochester Institute of Technology

2025 Grads Hired: 88%
2025 Grads Mean Salary: $77,000
Faculty: David I. Schwartz, Flip Philips
Graduates: Elan Lee (Exploding Kittens), Steven Van Slyke (co-inventor of OLED)

5. Abertay University

2025 Grads Hired: 44%
2025 Grads Mean Salary: $41,000
Faculty: Fraser Simpson (YoYo Games), Ruth Falconer (Women in Games)
Graduates: Zoe Sams (Riot), David Hynd (Rockstar)

6. DigiPen Institute of Technology

2025 Grads Hired: 62%
2025 Grads Mean Salary: $96,096
Faculty: Claude Comair (Nintendo), Ellen Guon Beeman (EA, Microsoft)
Graduates: Kim Swift (Portal), Alex Williams (Studio Wildcard)

7. Clark University

2025 Grads Hired: 68%
2025 Grads Mean Salary: $69,000
Faculty: Ezra Cove (Lord of the Rings Online, Turbine), Amanda Theinert
Graduates: Gary Goldberg (Fable Vision Studios), Stanley Pierre-Louis (ESA)

8. University of Utah

2025 Grads Hired: 36%
2025 Grads Mean Salary: $71,113
Faculty: Michael Young PhD, Jose Zagal PhD (Game Studies)
Graduates: Doug Bowser (former COO of Nintendo), Nolan Bushnell (founder of Atari)

9. Savannah College of Art and Design

2025 Grads Hired: 34%
2025 Grads Mean Salary: $62,500
Faculty: Nye Warburton, Cyril Guichard (Electrotank)
Graduates: Chad Dezem (Insomniac), Harrison Pink (Blizzard)

10. Michigan State University

2025 Grads Hired: 53%
2025 Grads Mean Salary: $62,000
Faculty: Brian Winn (Will Winn Games), Ricardo Guimaraes (Blizzard, Ubisoft)
Graduates: Brian Murray (EA), Steven Messinger (Rockstar)

11. Champlain College

2025 Grads Hired: 78%
2025 Grads Mean Salary: $50,226
Faculty: Jessie Leigh Gagnon (Riot Games, Phoenix Labs), Jason Della Rocca (Execution Labs)
Graduates: Heather Conover (Riot), Alex Martens (Rockstar)

12. Drexel University

2025 Grads Hired: 43%
2025 Grads Mean Salary: $42,857
Faculty: Dr. Paul Diefenbach (OpenWorlds Inc.), Dr. Frank Lee
Graduates: Girish Balakrishnan (Netflix), Greg Lebanov (Chicory, Wandersong)

13. Worcester Polytechnic Institute

2025 Grads Hired: 50%
2025 Grads Mean Salary: $82,000
Faculty: Ben Schneider (Standing Stone, Turbine), Ralph Sutter
Graduates: Michael Gesner (Riot), Jon Radoff (Beamable)

14. Northeastern University

2025 Grads Hired: 20%
2025 Grads Mean Salary: $80,000
Faculty: Christopher Barney, Celia Pearce (IndieCade)
Graduates: Devin Yang (Owlchemy Labs), Ryan Maloney (Blizzard)

15. Purdue University

2025 Grads Hired: 68%
2025 Grads Mean Salary: $67,066
Faculty: Robert Howard, Jeffery Kesselman
Graduates: Jeff Hanna (Volition), Keith Self-Ballard (PlayStation)

16. Falmouth University

2025 Grads Hired: 56%
2025 Grads Mean Salary: $28,311
Faculty: Tanya Krzwinska, William Huber
Graduates: Rex Crowle (Media Molecule), Sophie Shepherd (Respawn)

17. Shawnee State University

2025 Grads Hired: 25%
2025 Grads Mean Salary: $73,500
Faculty: Bastien Lecouffe-Deharme, Greg Lyons
Graduates: Dan Clark (Epic), Derek Lewis (Insomniac)

18. University of Florida

2025 Grads Hired: 33%
2025 Grads Mean Salary: $99,000
Faculty: Dr. James Oliverio, Dr. Hyo Jeong Kang
Graduates: Jesse Rapczak (Studio Wildcard), Joshua Jayaheri (Raven Software)

19. Full Sail University

2025 Grads Hired: 10%
2025 Grads Mean Salary: $62,109
Faculty: Patrick Kelly, Miguel De La Cruz
Graduates: Alejandro Garcia-Tunon, Elbert Perez (Unity)

20. Vancouver Film School

2025 Grads Hired: 11%
2025 Grads Mean Salary: $65,000
Faculty: Mike Wuetherick (Unity, Cloudhead Games), Joe Bonar (Unity)
Graduates: Armando Troisi (MY.GAMES, Ubisoft), Jack Kelly (New World Interactive)

21. Quinnipiac University

2025 Grads Hired: 60%
2025 Grads Mean Salary: NR
Faculty: Roger Clark (Rockstar), Elena Bertozzi
Graduates: Donovan Sepulveda (Riot), Zaria Brogdon (Rockstar)

22. New England Institute of Technology

2025 Grads Hired: 24%
2025 Grads Mean Salary: $50,000
Faculty: William Culbertson (Hasbro), Jordan Dubreuil
Graduates: Daniel Laba (Blizzard), Daniel Cuccia (Sony)

23. Bradley University

2025 Grads Hired: 15%
2025 Grads Mean Salary: $45,000
Faculty: David Abzug (Deep Silver), Ray Hazlip (Deep Silver)
Graduates: Diana Hughes, Joseph Lax (Sony Santa Monica)

24. Miami University

2025 Grads Hired: 56%
2025 Grads Mean Salary: $56,690
Faculty: Geoffrey Long PhD (Microsoft, Disney), Ben Nicholson
Graduates: Ellie Helbling (Naughty Dog), Brian Beebe (Riot)

25. DePaul University

2025 Grads Hired: 77%
2025 Grads Mean Salary: NR
Faculty: Ed Keenan (Midway), Allen Turner (Disney)
Graduates: Kevin Geisler (Young Horses), Mark Nauta (Firaxis)

26. Howest University of Applied Sciences

2025 Grads Hired: 54%
2025 Grads Mean Salary: NR
Faculty: Tristan Clarysse (Larian, Bioware), Kevin Hoefman (Larian)
Graduates: NR

27. Rensselaer Polytechnic Institute

2025 Grads Hired: 8%
2025 Grads Mean Salary: $74,917
Faculty: Rush Swope (Guerilla Games), Maurice Suckling (Epic, 2K)
Graduates: Curtis Priem (Nvidia co-founder), Zach Barth (Zachtronics)

28. University of California, Santa Barbara

2025 Grads Hired: 33%
2025 Grads Mean Salary: $155,000
Faculty: Pradeep Sen PhD, Tobias Hollerer
Graduates: Cody Corona (Electronic Arts), Eric Roth (Screenwriter)

29. Abilene Christian University

2025 Grads Hired: 63%
2025 Grads Mean Salary: $52,200
Faculty: Dr. Brian Burton, Rich Tanner
Graduates: Morgan Davis (Gearbox), Chase Yamaguchi (Sony)

30. Kent State University

2025 Grads Hired: 34%
2025 Grads Mean Salary: $65,000
Faculty: Christopher Totten, Justin McCrea
Graduates: Joey Gage (EA), Nina Marotta

31. Marist College

2025 Grads Hired: NR
2025 Grads Mean Salary: NR
Faculty: Dr. Karen Schrier (Nickelodeon), Amy Finkel
Graduates: Darren Sugg (Epic), Rex Dickson (Blizzard)

32. University of California, Irvine

2025 Grads Hired: 25%
2025 Grads Mean Salary: $83,000
Faculty: Katie Salen, Constance Steinkhueler
Graduates: Rob Pardo (Blizzard), Ben Brode (Blizzard)

33. LaSalle College Vancouver

2025 Grads Hired: 44%
2025 Grads Mean Salary: $70,000
Faculty: Dr. Amir Jahanlou, Lisa Smedman
Graduates: Jarett Metcalfe (EA), Dani Merrithew (Blizzard)

34. Academy of Art University

2025 Grads Hired: NR
2025 Grads Mean Salary: NR
Faculty: David Goodwine, Brian Hess
Graduates: Ciaron Ireland (Striking Distance), Tala Furniss (Playtika)

35. University of Miami

2025 Grads Hired: 32%
2025 Grads Mean Salary: $54,000
Faculty: Dr. Lindsay Grace, Dr. Anna Queiroz
Graduates: Pingtin Ji (Ubisoft), Casey O'Brien (Respawn)

36. Northern Arizona University

2025 Grads Hired: 50%
2025 Grads Mean Salary: NR
Faculty: Jared Duval (Serious Games), James Palmer
Graduates: NR

37. Dakota State University

2025 Grads Hired: NR
2025 Grads Mean Salary: $41,964
Faculty: Peter Britton, Erik Pederson
Graduates: Matt Engesser (Rockstar, Epic), Jimmy Chattin (Microsoft)

38. University of Michigan-Dearborn

2025 Grads Hired: 20%
2025 Grads Mean Salary: $80,000
Faculty: Brice Maxim, Foyzol Hassan
Graduates: Austin Krauss, Shane Costello

39. New Jersey Institute of Technology

2025 Grads Hired: NR
2025 Grads Mean Salary: $78,265
Faculty: Richard Thomspon, Hye Yeon Nam
Graduates: Jason Chin (Nvidia), James Tichenor (Meta)

40. University of Wisconsin-Stout

2025 Grads Hired: 10%
2025 Grads Mean Salary: $40,000
Faculty: Dr. Andrew Williams, Kim Loken
Graduates: Jose Estrada (Riot), Eleanore Falck (Epic)

41. Stetson University

2025 Grads Hired: 100%
2025 Grads Mean Salary: $68,000
Faculty: Dengke Chen, Matt Roberts
Graduates: Daryl Holt (EA Sports), Ben Burbank (Valve)

42. Oklahoma Christian University

2025 Grads Hired: 25%
2025 Grads Mean Salary: $35,000
Faculty: Jacob Touchstone, Dr. Robert Nix
Graduates: Alex Hallenbeck (Bungie), Laurel Zwirtz (Ember Labs)

43. Bloomfield College of Montclair State University

2025 Grads Hired: 7%
2025 Grads Mean Salary: NR
Faculty: Samuel Leigh (Viacom), Oliver Hong (IGDA)
Graduates: Josiah Hunt (Cloud Chamber), Michael Murphy (Blizzard)

44. Indiana University Bloomington

2025 Grads Hired: 20%
2025 Grads Mean Salary: $70,000
Faculty: William Emigh (Studio Cypher), Michael Uslan
Graduates: Max Lancaster (Psyonix), Kees Leyundjik (Nintendo)

45. University of Massachusetts Dartmouth

2025 Grads Hired: NR
2025 Grads Mean Salary: NR
Faculty: Firas Khatib, Clinton Rogers
Graduates: Matt Mrazik (WB Games)

46. ArtCenter College of Design

2025 Grads Hired: 25%
2025 Grads Mean Salary: NR
Faculty: Zachary Adams, Tim FitzRandolph
Graduates: Syd Mead, Michael Bay (Director)

47. Louisiana State University

2025 Grads Hired: 4%
2025 Grads Mean Salary: NR
Faculty: Marc Aubanel (Electronic Arts), George Robins
Graduates: Janiece Campbell (Epic), Ashley Aucoin (Meta)

48. Oklahoma City University

2025 Grads Hired: 10%
2025 Grads Mean Salary: $50,000
Faculty: Justin Shaw, Shirley Zhang
Graduates: Andrew Remeniski, Jessica Jennings

49. Cleveland Institute of Art

2025 Grads Hired: NR
2025 Grads Mean Salary: NR
Faculty: Steven Gutierrez (Serious Games), Harrison Walsh (Niche Worlds Cult)
Graduates: Zack James, Tim Elek (Epic)

50. New York Film Academy

2025 Grads Hired: NR
2025 Grads Mean Salary: NR
Faculty: Glenn Storm, Andrew Allen
Graduates: Leandro Ribeiro (Afterverse), Guillermo Quesada (Amazon)

GRADUATE RANKINGS

1. New York University

2025 Grads Hired: 59%
2025 Grads Mean Salary: $84,800
Faculty: Eric Zimmerman (Gamelab), Dr. Mitu Khandaker (Spirit AI)
Graduates: Jenny Jiao-Hsia (IGF), Carol Mertz (Exploding Kittens)

2. University of Central Florida

2025 Grads Hired: 80%
2025 Grads Mean Salary: $94,646
Faculty: Benjamin Noel (EA), Nicholas Zuccarello (Sony)
Graduates: Carlos Barbosa (WB Games), Brian DeSanti (Blizzard)

3. University of Southern California

2025 Grads Hired: 81%
2025 Grads Mean Salary: $97,857
Faculty: TreaAndrea Russworm (Author, "Gaming Representation"), Mark Bolas (Fakespace Labs)
Graduates: Jenova Chen (Thatgamecompany), Brent Strong (Disney)

4. Rochester Institute of Technology

2025 Grads Hired: 79%
2025 Grads Mean Salary: $83,000
Faculty: David Schwartz, Flip Philips
Graduates: Elan Lee (Exploding Kittens), Steven Van Slyke (co-inventor of OLED)

5. University of Utah

2025 Grads Hired: 80%
2025 Grads Mean Salary: $87,325
Faculty: Michael Young PhD, Jose Zagal PhD
Graduates: Doug Bowser (COO of Nintendo America), Nolan Bushnell (Founder of Atari)

6. Southern Methodist University

2025 Grads Hired: 45%
2025 Grads Mean Salary: $104,744
Faculty: Dr. Elizabeth Storz Stringer (Atari), Joowon Kim MacDonald
Graduates: Grace Liu (Netflix Games), Daniel Rivin (Disney)

7. Abertay University

2025 Grads Hired: 49%
2025 Grads Mean Salary: $49,000
Faculty: Dr. Kenneth Fee, Joseph DeLappe
Graduates: Zoe Sams (Riot), David Hynd (Rockstar)

8. Clark University

2025 Grads Hired: 72%
2025 Grads Mean Salary: $70,000
Faculty: Terasa Ulm, Scott Niemi
Graduates: Stanley Pierre-Louis (ESA), Gary Goldberg (Fablevision)

9. Savannah College of Art and Design

2025 Grads Hired: 73%
2025 Grads Mean Salary: $62,500
Faculty: Nye Warburton, Cyril Guichard (Electrotank)
Graduates: Lucas Slominski (ZeniMax), Chris Schmidt (Bioware)

10. Drexel University

2025 Grads Hired: 50%
2025 Grads Mean Salary: $80,000
Faculty: Dr. Paul Diefenbach (OpenWorlds Inc.), Dr. Frank Lee
Graduates: Anna Nguyen (NetherRealm Studios), Greg Lebanov (Wondersong, Chicory)

11. Michigan State University

2025 Grads Hired: 60%
2025 Grads Mean Salary: $62,000
Faculty: Amanda Cote, Susan Bonner
Graduates: Brian Murray (EA), Geoff Johns (DC Entertainment)

12. DigiPen Institute of Technology

2025 Grads Hired: 37%
2025 Grads Mean Salary: $179,000
Faculty: Claude Comair (Nintendo), Brian Schmidt
Graduates: Luis Villegas (Bungie), Taralyn von der Linden (Pixar)

13. Northeastern University

2025 Grads Hired: 40%
2025 Grads Mean Salary: $80,000
Faculty: Celia Pearce (IndieCade), Bob De Schutter
Graduates: Angi Liu (Blizzard), Wendi Zhang (Activision Blizzard)

14. Worcester Polytechnic Institute

2025 Grads Hired: 70%
2025 Grads Mean Salary: $82,000
Faculty: Ben Schneider (Turbine, Standing Stone), Ralph Sutter
Graduates: Michael Gesner (Riot), Jon Radoff (Beamable)

15. Falmouth University

2025 Grads Hired: 71%
2025 Grads Mean Salary: $43,386
Faculty: Tanya Krzwinska, William Huber
Graduates: Rex Crowle (Media Molecule), Sophie Shepherd (Respawn)

16. University of Florida

2025 Grads Hired: 85%
2025 Grads Mean Salary: $99,000
Faculty: Dr. James Oliverio, Dr. Hyo Jeong Kang
Graduates: Joseph Murphy (EA), Pooja Agshikar (Microsoft)

17. Full Sail University

2025 Grads Hired: 33%
2025 Grads Mean Salary: $62,109
Faculty: Dr. Shawn Stafford, Chris Keeling
Graduates: Erin Eberhardt (Dreamhaven), Aaron Hamilton Cook (Blizzard)

18. DePaul University

2025 Grads Hired: 91%
2025 Grads Mean Salary: NR
Faculty: Ed Keenan (Midway), Allen Turner (Disney)
Graduates: Sara Asmussen (Blizzard), Kevin Geisler (Young Horses co-founder)

19. Bradley University

2025 Grads Hired: 40%
2025 Grads Mean Salary: $45,000
Faculty: David Abzug (Deep Silver), Ray Hazlip (Deep Silver)
Graduates: Alex Cline (Shapeshifter Games), Annabelle Pries (Blind Squirrel)

20. Purdue University

2025 Grads Hired: 7%
2025 Grads Mean Salary: $67,066
Faculty: Robert Howard, Jeffery Kesselman
Graduates: Jeff Hanna (Volition), Keith Self-Ballard (Bungie)

21. Carnegie Mellon University

2025 Grads Hired: 64%
2025 Grads Mean Salary: $74,000
Faculty: Jesse Schell (Schell Games, Disney Imagineer), Derek Ham
Graduates: Neil Druckmann (Naughty Dog co-president), Bei Yang (Meta, Disney)

22. Rensselaer Polytechnic Institute

2025 Grads Hired: 33%
2025 Grads Mean Salary: $74,917
Faculty: Maurice Suckling (Epic, 2K), Rush Swope (Guerilla Games)
Graduates: Dr. Bryan Kim, Dr. Matthew Gantt

23. University of Wisconsin-Stout

2025 Grads Hired: 100%
2025 Grads Mean Salary: $44,000
Faculty: Dave Beck, Andrew Williams
Graduates: Hue Vang (Dreamworks), Andrew Murphy (Apple)

24. American University

2025 Grads Hired: 50%
2025 Grads Mean Salary: $70,000
Faculty: Andrew Phelps, Krzysztof Pietroszek
Graduates: Ryan Seymour (Smithsonian), Kelli Dunlop (TakeThis.org)

25. New York Film Academy

2025 Grads Hired: 2%
2025 Grads Mean Salary: NR
Faculty: John Zuur Platten, Craig Caton-Largent (Digital Domain)
Graduates: Leandro Ribeiro (Afterverse), Guillermo Quesada (Amazon Studios)

Homework: Creative accessibility

In recent years, the industry has reinforced the importance of designing videogames with accessibility in mind. After all, the more people who can comfortably play your game, the better chance you have of reaching an audience. The most common forms of accessible game design are options that allow players to tweak colorblind settings, alter models that trigger certain phobias, or tone down flickering lights. But sometimes, a game calls for accessibility considerations specific to the project.

Clair Obscur: Expedition 33, the breakout hit RPG of 2025, looked beyond the usual settings when designing for accessibility. Developer Sandfall Interactive took note during playtests that its jerky combat camera was causing motion sickness for some players. To get ahead of that problem, the team included a camera stabilization option at launch that mitigated motion sickness without making any major artistic sacrifices.

Imagine you're developing a first-person shooter, a strategy sim, or a puzzle game that relies on color. How would you anticipate accessibility needs and design for them?

Programming believable AI characters

Artists can produce amazing concept art and designers can think up compelling gameplay loops, but videogames never make it out of the conceptual phase without talented programmers constructing the underlying logic. Their work is often invisible—a line of code that makes a character's jump or car's drift feel perfect—but one place a programmer's work gets to shine is in the non-player characters that populate a world. Programming believable behaviors into friendly NPCs or hostile creatures is crucial to the player believing the virtual world they're inhabiting.

A recent example of great AI programming is Arc Raiders, the hit 2025 extraction shooter that attracted millions of players through its unique PvPvE format. Players can optionally fight each other on its large maps, but they're encouraged to band together because of the presence of Arc—a fleet of mysterious robots that roam the planet's surface. The Arc, and specifically the behaviors that developer Embark Studios programmed into them, are the glue that binds the whole experience together. Arc drones will dynamically patrol the map, changing up their routes and thoroughly investigating noises. They're smart, but crucially, Embark programmers ensured they aren't so smart that players can't outmaneuver them. Careful stealth and awareness of their vision cones can mitigate their threat.

The Arc's dynamic perception is impressive, but the real magic trick is how the robots react to damage inflicted by players. When players shoot off one of a flying drone's propellers, it will dynamically attempt to compensate for the loss in power, drifting and dipping through the air before it either finds balance or crashes into the ground.

The same goes for ground-based Arc robots. When a four-legged Leaper loses a leg, it compensates by changing its walking pattern. That adherence to physics-based rules, as enabled by talented programmers, makes the Arc feel alive. They're curious, inquisitive, and worthy foes with a survival instinct, not rigid constructs with a few pre-determined actions.

Programmers are also at the heart of our favorite RPGs. Baldur's Gate 3 is known for its excellent writing, but talented programmers working behind the scenes constructed the complex web of quest outcomes that made its characters so reactive to the player's choices.

Sometimes, the best way to program believable AI characters is to double down on social interactions. In Playground Games' upcoming Fable reboot, every single citizen of the open-world Albion is a real, simulated person with a name, home, and daily routine. That much has been done before, but where Fable goes the extra mile is simulating every individual citizen's personalities and their perspectives on the world. The people of Albion will form an opinion of the player,, and that opinion can change based on one-on-one interactions or choices made throughout the story.

Fable's programmers call this system the "living population." Passing citizens will remark upon developments in the main story. A shop clerk might have a high opinion of the player if that player owns the shop they're working in. If the player develops a reputation in town as a tycoon, some citizens will find them off-putting by default.

If you find yourself fascinated by the little details and behaviors that make great videogame AI and want to be a part of it, then programming may be your starting point.

Homework: Finding fun

Often in game design, creativity goes hand in hand with flexibility. It's rare that a game comes together exactly as it was pitched. Instead, developers tend to find their game's unique hook along the journey. That can mean taking your original idea and refining it, but sometimes the project's entire premise can change when an even better idea comes along.

The 2025 Steam hit REPO, developed by Semiwork, was not originally conceived as a co-op horror game. At first, the team was making a singleplayer horror cleaning game in which players picked up trash around a mansion while avoiding a monster. By showing creative flexibility, the team realized it should pivot to multiplayer (because everything is more fun with friends). That change eventually led to moving away from the cleaning concept altogether, as the team found cleaning together wasn't as fun as it originally sounded. What was fun, they discovered, was cooperatively carrying fragile objects through precarious mansions. Thus, REPO was born.

Imagine you're developing a playable prototype for a 3D platformer. Your playtesters love how the jump feels, but aren't responding well to the game's setting or characters. How do you pivot?

The art of map design

One of the most specialized disciplines in game development is multiplayer map design. Multiplayer arenas can be sculpted from anything—a California golf course in Battlefield 6, a neon-soaked city in Fortnite, a long-forgotten spaceport in Arc Raiders—but unlike traditional level designers, map makers have the unique challenge of creating spaces that delight players upon first impression while also standing up after hundreds or thousands of hours of playtime.

Good maps have to be as distinct or visually interesting as any singleplayer space, but crucially, they also have to be fair. Mapping challenges every facet of spatial design: great mappers have an understanding of architecture, nature, composition, lighting, and can visualize how players are likely to move through spaces from any direction. The balance is considering all of these principles while also taking competition into account: Are the sightlines clear? Is the space too busy or confusing? Will it be fun to navigate the hundredth time players see it?

One way designers achieve fairness in maps is through symmetry: even if two sides of a map have different theming, they'll often share a similar or even identical shape to ensure one side isn't advantageous over the other. This is why maps in competitive shooters tend to take the general shape of a rectangle or bowl—no matter which side your team is on, it takes roughly the same time to reach the center.

We see this sort of "camouflaged symmetry" used on the largest map in Battlefield 6. Mirak Valley is a wide, flat expanse set in Tajikistan. On one end of the map is a small village with lots of indoor cover. On the opposite end is a bombed-out village with fewer standing buildings, but an elaborate network of trenches. The two sides achieve the same thing—locations with cover that act as capture points—with the same general land mass, but the map team at DICE utilized cover in two different ways. The side with fewer buildings favors vehicles, while the village gives infantry players a better fighting chance. It's not 100% symmetrical, but it's close, and it's more interesting to opt for variety.

Another way to create variety in maps is when the terrain can be changed by players. The destructible maps in the likes of Rainbow Six Siege X and The Finals force mappers to really get their creative juices flowing: How do you maintain competitive integrity when the map changes every session? The developers at Ubisoft tackled this problem in Siege by designing a visual language for "soft" walls that can be destroyed by players, and "hard" walls that cannot. In this way, designers can tightly control what sightlines can be altered by players and maintain a building's intended shape.

The Finals, meanwhile, sets no such limitations. Literally everything but the ground itself can crumble—a freedom that creates unparalleled dynamism, but also forces the map team to find other ways to distinguish arenas. The team at Embark Studios goes all-in on setting for this reason: its globe-trotting maps recreate real world locations across various time periods, including a futuristic Las Vegas, Monaco, Mexico, Korea, and even 16th century Japan.

The rise of battle royale games over the last decade has created a whole new category of map design: the 100-player mega map. Creating one large map instead of a dozen smaller ones is a major undertaking that requires close collaboration between individual mappers. The average Fortnite map, for instance, has over a dozen distinct points of interest in addition to kilometers of hills, roads, rivers, and mountains between them. It's a lot to consider, but all the extra space gives designers a blank canvas to sculpt beautiful views, incorporate multiple biomes, or hide secrets.

If these are the sort of challenges that get your mind racing, you may have the heart of a multiplayer map designer.

In a small ray of light for our fallen world, an indie dev who poured their heart and soul into their game seems to have actually been rewarded for it. As spotted by Dexerto (via GamesRadar), solo dev Cakez learnt his game Tangy TD had made $250,000 in a week while he was streaming, and he seemed truly unable to believe it.

Cakez has been documenting the process of making Tangy TD on his YouTube channel and streams for the past four years. A clip of him opening his day-one sales report and seeing it had made $30,000 went viral all by itself, and when the time came to check out his week-one report, he invited the community to see it with him.

And, well, an already-emotional Cakez couldn't really believe what he saw: $245,123 gross lifetime Steam revenue ($197,847 net revenue). That's from 28,078 copies sold. Our hero couldn't believe it, and was immediately overcome by tears as his wife whooped in celebration beside him. It was, I must admit, an incredibly wholesome sight.

In a stream, Cakes stayed humble about the whole thing: "Thank you very much. I feel like I really don't deserve this. But I appreciate it... It's so amazing to see how many people have come out to support me, essentially, and what I do. It's just crazy, I really don't know what to say. I don't know why people are so nice."

Which is all rather heartwarming, if you ask me. As for Tangy TD? It's doing pretty well on Steam: 89% of its reviews at the time of writing are positive, with one player writing "Saw the reaction vid. Bought the game. It's pretty good. Love a game dev that doesn't expect anything. GG's my dude."

It is about one hour to presentation time and Alexander Helliwell has decided to experiment. His mod—an attempt to convert Prologue: Go Wayback's default temperate woodland biome to a scorching desert—was looking pretty slick last time I passed by: all dazzling sunsets and dry heat and, an added bonus, running quite well. Turns out that yanking most of the game's world out and replacing it with sand carries a performance uplift.

It's still performing quite well, but a glance at his screen reveals a new, seemingly major change. Now the sands are pockmarked by great stretches of an unidentifiable black substance, the aftermath of experimentation on the mod team's part. "Is that oil?" someone asks. Helliwell turns in his seat, eyes wide: "I don't know."

Let's jam

In the last few days of 2025, PlayerUnknown Productions summoned its developers for one of its regular in-person conflabs. Though its offices are in Amsterdam, very few of its staff actually live anywhere near the city, and to compensate for the deficit in physical communication, company policy is to have as many staff as possible make the schlep to Mokum every couple of months.

Devs roll in from across Europe on trains and planes from Finland, Hungary, the Czech Republic and beyond. At least one bleary-eyed staffer has spent 12 hours on a train from Marseilles. But they are not—this time—just here to endure a barrage of stand-ups. They're here for a modjam, the same kind of minor modding festival that fans of games like Minecraft throw all the time, with the aim of spinning up game mods in a limited period.

It's an experiment: what happens if you loose a dev team on their own game and let them do whatever they want to it, using a rough assemblage of UE5 asset-store props, the game's pre-existing systems and objects, and whatever they can gin up themselves over the course of two days' modding?

Quite a bit, it turns out. Over 48 hours, Prologue's devs spin up new mechanics, items, systems, and full-blown total conversion mods. Some of them even conjure up ad-hoc marketing departments, producing ads and mock-up Steam pages featuring their creations. At the end of the jam, they'll present their work to the rest of the studio: what they learnt, where they think the ideas could go from here. An idea that goes down particularly well might find itself incorporated, one way or another, into a future update to the game itself.

But it's also an experiment in modding as a concept. "Modding was one of the bigger requests that we've had from the community," says studio founder Brendan Greene, the PlayerUnknown in PlayerUnknown Productions. "They want to be able to fuck around and do their own thing.

"This is what the modjam is for. It's like: let's do some experiments and see how possible this is and what we can do."

What we can do

PlayerUnknown Production's developers split into 10 projects, consisting of anything from a single dev to a full team of five. I rove between them in a way that must be quite annoying, accompanied by studio head of publishing Maurice Tan, whose task is to enlighten me as to precisely what I am looking at on people's monitors, while simultaneously ensuring that I don't commandeer an empty desk and unearth forbidden game design secrets or start developing PUBG 2.

My main posting is with senior 3D character artist Hakan Kamar, who's working with four other PlayerUnknown Productions devs to create a project that converted Prologue into a full-on horror game. If you were out of earshot of copyright lawyers, you might call it a Slenderman mode: "Something follows you, you don't know what, and you need to progress through the forest finding items, and then eventually escape.

"The thing we want to test is: 'Can we make Prologue a horror game?'"

The answer, so far as I can tell, is 'yes, though not without some difficulty.' In place of Prologue's usual loop—spawn in, try to survive and make your way to a distant weather station in a generally barren world—Kamar and co's horror mod has you collect scattered pages while avoiding the creature-legally-distinct-from-Slenderman. In place of the weather station, looming over the map is a great grey monolith, lit up by floodlights in the dark. It's the Half-Life-2-esque laboratory from which your pursuer escaped.

It is also far, far too large. Kamar is futzing with it as I watch, murmuring that "100 times is too big" as the sheer scale of the building sends the game's camera careening through the level geometry and down into the heart of the Earth.

Elsewhere, other devs on the horror team beaver away on their own bits of its creation, heads down in full production mode. In evening meetings the full squad—consisting of art director Jorry Rosman, senior character concept artist Sarah Sneeboer, principal audio designer Miguel Salas, programmer Yoan Rock, and Kamar—ponders choices like whether the creature should teleport around the island or be forced to walk everywhere, which skin materials to use so it doesn't practically glow in the dark, and whether its heavy-limbed "Sopranos walk" is a positive or negative (positive, obviously).

Meanwhile, Salas works on the audio for the mod's creature-detection system, a Silent-Hill-style radio that gets increasingly distorted the nearer the monster is—the strains of Tiny Tim's Livin' In The Sunlight slowing to static as your doom approaches.

It's a remarkably serious operation for what might, at first, seem like an end-of-year jolly. Rosman even gets cracking on a full-on marketing campaign for the mod, replete with Steam pages and product screenshots. The team treats the jam like, well, game development, with all the gravity that implies, save for their first choice of name: "What Happens In The Darkest Days In The Dark?"

Right to the source

The studio bustles during the day I spend in its office. Teams work on their projects, yes, but they also flit back and forth between desks, taking in developments in each other's work, offering criticism, advice, and expressions of sheer bafflement, and occasionally coming away with new inspiration for their own projects.

It's almost too perfect: a physical microcosm of the kind of organic, freewheeling internet of yore that has been almost entirely obliterated by corporate consolidation. No wonder Greene is keen on the Modjam idea—this is his entire pitch for Prologue and the games that will follow it. "The internet is dying," he says. "What was once one of humanity's greatest hopes of technology, meant to bring us closer, has decayed into a minefield of bots, of misinformation and digital splintered digital bubbles… But we don't come online for bots. We come online for each other.

"Even here today, we have all the team in for that human connection, rather than doing stuff online all the time. You need that. Now, I think that human connection is needed more than ever."

Greene's vision for Prologue isn't short on ambition—he sees it as step one of a metaverse, an open-source "global commons" that, unlike similar pitches by Mark Zuckerberg, "cannot belong to a single company." He imagines—not Prologue itself, but the tech and ideas behind it—one day providing the infrastructure for an internet more like the one we had 20 or 30 years ago, just in 3D: a space for the spontaneous formation of human relationships (and, you know, sometimes people calling you a prick over a Star Wars argument). "People don't want 'content.' They want each other."

Am I fully bought in? Not quite, but then if you'd shown me PUBG in 2016 I wouldn't have bet that'd become the defining template for seemingly all videogames for years to come either. Regardless of what I think, Greene's grand plans for Prologue and its successor games means that events like the Modjam are kind of essential. People have to be able to build on this foundation for it to become anything like the sunlit, open-source metaverse that Greene and the team imagines down the line.

Black hole sun

The good news for that vision is: it all seems to be going quite well. As Kamar and the horror mod team gin up Steam page assets and cover the collectible pages with royalty-free baboon drawings, other devs labour on their own mods. In one corner of the office, Helliwell and environment artist Boy Best convert the game from a frigid European forest to an Arizonan desert—"You're no longer trying to stay warm, you're now trying to stay cold and cool and stay hydrated."

Despite only two days' work, a casual onlooker (me) would be hard-pressed to notice the game was not always desert-based. There are even new sandstorm mechanics to obscure your sight and scramble your sense of direction. Helliwell has, perhaps inevitably, labelled some of these mechanics "Darude – Sandstorm" in the recesses of the game code. Occasionally his speakers quietly emit the song itself, in what must surely be a boon for team morale.

Meanwhile, a team consisting of technical game designer Guillaume Lucas, senior FX artist Dominikus Reiter, programmer Adrien Fabre, and environment artist Thijs de Haas is working on making Prologue a little bit more Far Cry, adding vehicles like a hang-glider and, tantalisingly, a shopping cart to the game.

The latter might actually be the more technically demanding of the two—attempting to cajole Prologue's engine into allowing the player to fill a cart with loose objects and move them about can prove too much for it. During the project's presentation to the studio, the team runs into an inevitable crash when the trolley gets overfull, prompting whoops and cheers from their assembled fellow devs.

But my favourite of the lot is the black hole, created by technical artists Michael van den Berg and Djanco Dewus. It's, well, a black hole. A great void manifests in the sky and pursues the player about the map, sucking up potentially useful items, starting fires, and growing ever-larger as it does so.

The aim is to force the player to keep moving—a little dash of PUBG's circle in Prologue—but I mostly find it entertaining as an exercise in random destruction. You can't help but feel Prologue's devs are working out a little of their frustration with the game's sometimes-untameable engine when they create mods whose aim is to lay waste to the entire thing.

The past is prologue

Is this the future? You won't catch me making definitive pronouncements for people to mock from the comfort of their orbital habitats 50 years from now, but I'll say this: if the modjam proves anything, it's that you'll never truly kill devs' desire to tinker with tech for the sheer love of it, no matter how corporatised and monopolised our online spaces become.

Whether it's the future or not, it's not a bad foundation to work from.

Microsoft has just shared an update to DirectX, labelling it "the biggest wave of new tooling features in DirectX’s history." The company, alongside major "GPU hardware partners", has shared its "dream of bringing console-level GPU developer tools to Windows."

Microsoft reports, "AMD, Intel, Nvidia, and Qualcomm have worked closely with us throughout feature development, each making significant contributions to make this release possible."

The new changes come in a few different sections. The first is the introduction of DirectX dump files. As elaborated on in the Microsoft Game Dev blog, "GPU-related bugs can emerge at any point in a game’s lifecycle, and developers have sought more thorough crash-dump infrastructure to understand their causes."

These dump files can help with both retail users and developer machines, giving a better understanding of what has gone wrong and why. Dump files will reportedly give logs on the hardware and driver state, plus the state of DirectX, and what was happening in the game before the dump file was created. Giving more information in condensed forms like this can help with all kinds of troubleshooting.

Another big update to DirectX is the introduction of "Live Shader Debugging". Effectively, this would allow for real-time shader debugging, which is a celebrated feature on Xbox already. It is currently planned for release in 2027. Microsoft claims it is designed for "needle in the haystack" style problems and also reports it's "the deepest GPU tooling collaboration with hardware vendors in Windows history."

This debugging on PC was only shown off as a preview at this year's GDC, but Microsoft says the public can expect more details in the coming months.

Microsoft has also noted its new shader explorer, which it claims is "a new way to inspect, understand and debug compiled shaders, with deeper live analysis".

This new explorer works with Pix, Windows' debugging tool, and Microsoft has shown off a version working on both AMD and Intel hardware. Interestingly, the shader explorer is not paired to the driver, so developers can analyse shaders for GPUs they don't own. This could potentially help with optimising for the myriad hardware forms that make up PC gaming, especially in smaller dev teams with less access to different hardware.

Microsoft points out specific integrations that AMD and Intel have implemented for these new bits of tech, but it points out that "all four have invested in PIX plugin support to surface their hardware-specific information through PIX’s standardized UI."

Given Microsoft's push towards PC gaming with its hybrid next-generation console, Project Helix, there's a good argument for why Microsoft might want to enable console-like dev tools for Windows devices. Project Helix is losing part of its traditional console appeal with its PC gaming capabilities, so giving developers tools to fine-tune and more easily report problems could help with long-term optimisation.

Back in my day, I would pass the time waiting for my Sims family to load back into existence by staring out of the window and appreciating all the natural beauty that surrounded my folks' place. Not so anymore—compression tech has gotten so good that the only fronds I have time to contemplate are the polygonal progeny of SpeedTree.

One may be forgiven for thinking load times can't get much speedier, but Microsoft has released further details about a swish new GPU decompression algorithm in DirectStorage 1.4, assuring that they can. The company first released these details at GDC 2026, during the DirectX State of the Union 2026: DirectStorage and Beyond talk.

Also on Wednesday, Microsoft released a public preview of DirectStorage 1.4 and the Game Asset Conditioning Library, which now both support the Zstandard codec as an option for compressing game assets on Windows. The Zstandard codec is already widely used, but Microsoft claims that the additional support offered via the Game Asset Conditioning Library can improve asset compression ratios by up to 50%. To put it another way, noticeable texture pop-in might just be about to pop-out of existence.

Microsoft describes Zstandard (or Zstd) as "an open standard that improves compression ratios, enables faster load times, and provides smoother asset streaming for content-rich games."

DirectStorage already incorporates another compression codec, GDeflate. A collaboration between Microsoft and Nvidia, it's a similarly open-source bit of GPU compression tech that was first introduced back in DirectStorage 1.1. The first game to officially support it, by way of Nvidia's own RTX IO, was Portal: Prelude, an at the time 15-year-old mod.

Last year's Spider-Man 2 port also used the tech via DirectStorage, though arguably not well at launch. More recently, Resident Evil Requiem is one of the few games that use its GPU data decompression—though not all rigs will harness this tech, even though they're capable of it.

According to Microsoft's own benchmarks, GDeflate is still the way to go for developers wanting to let the GPU handle asset loading, but Zstd is better for CPU-intensive projects. Zstd does also offer a better compression ratio overall compared to GDeflate—but only by about 2%.

A number of different codecs were evaluated for inclusion, but "Zstd stands out by delivering competitive compression ratios and decompression performance, broad availability on hardware and software across operating systems, and widespread adoption in OS, cloud, and web scenarios." Basically, its inclusion alongside GDeflate offers developers another option when figuring out what will suit their workload best.

Even greater improvements to this compression tech will be ushered in down the line by "GPU partners [working] towards future hardware specific optimizations for Zstd." Those GPU partners chime in with some much-needed context, too.

"Aligning the industry on an open compression standard builds a foundation for future game titles to deliver immersive experiences with even larger worlds than realistically possible today," Daniel Staheli, CVP of Software Development at AMD, said, "We plan to make optimizations for AMD GPUs available in a public driver during the second half of 2026 and look forward to seeing how developers use these investments to enhance the player experience.”

Rival GPU manufacturer and now massive AI company Nvidia echoed a similar sentiment, with distinguished engineer Patrick Neill saying, "NVIDIA is excited to bring Zstd support to game developers, with decompression optimizations tailored for NVIDIA GeForce RTX GPUs arriving in the second half of this year."

Intel is also on board, hoping to "tune Zstandard decompression through DirectStorage across our GPU architectures". Qualcomm also said it will bring Zstd-tuned driver updates "that reflect our investments in Zstd decompression on our platforms" by the end of the year.

To aid in 'aligning the industry on an open compression standard,' Microsoft has also made its Zstd GPU decompression compute shader open-source. It's still in active development, but currently interested devs can get their hands on a version of the shader that is "optimized for content chunked to 256KB or smaller, consistent with modern game packaging patterns for streaming workloads."

While a lot of the language here implies a PC focus, Microsoft is definitely still interested in console hardware. Just for a start, the new CEO of Microsoft Gaming, Asha Sharma, said 'hello' by promising to "recommit to our core Xbox fans and players." Then there was the announcement of next-gen console Project Helix, a hybrid box that will apparently play both PC and Xbox games—though another GDC talk revealed this week that it's still years away.

Even so, it's clear Microsoft still wants to keep its thumb firmly stuck in the gaming pie. CEO Satya Nadella recently said the company will "always" invest in gaming, and that Xbox "at its best lifts the entire company." Smart compression improvements like DirectStorage 1.4 may help Microsoft pave its way to a brighter gaming future…bit by bit.

In November 2024, developer Maxime Foulquier posted footage of Bloodborne Top Down Arena, a fan tribute mixing FromSoftware's relentlessly goth setting with Vampire Survivors mechanics. The project looked kinda cool, if not a little close to its source material.

Foulquier went quiet on the project after posting that video, and as it turns out, for good reason: In March 2025 he received a sternly-worded letter from lawyers representing Sony. Rather than announce straight away that the project was done for, Foulquier decided to wait.

"I didn’t say anything in March 2025 because I decided to repurpose all the assets I had created for that project into a new game set in my own universe, inspired by Bloodborne," Foulquier wrote on X, where he has also posted an image of the cease and desist letter. "My plan was to reveal the letter once the game was ready to communicate.

"Unfortunately, plans have changed. I’d rather speak about it now, as my disappointment is immense. The disappearance of Bluepoint is a tremendous loss for the industry and for the dream of a Bloodborne remake."

Foulquier posted the letter just after Sony confirmed it will close famed remake studio Bluepoint. It has long been hoped—but never hinted at by Sony—that Bluepoint might work on either a Bloodborne remake or remaster. A pipe dream for sure, but Bluepoint did remake Demon's Souls and Shadow of the Colossus, two other Sony-owned games with severely loyal followings.

With the closure, one of the Bloodborne community's most powerful coping mechanisms has been dashed. It was for the hope of a Bluepoint-fronted Bloodborne remake that Foulquier received the cease and desist letter with equanimity.

"I was incredibly excited about this project and worked on it a lot in my spare time," Foulquier wrote. "Then I received the letter you can see here. I told myself, 'Okay… I expected this might happen. Maybe it’s because a real Bloodborne remake is actually in development, with Bluepoint leading it for the PS6 launch?'"

It's not the first time Sony has killed a fanmade Bloodborne project: Bloodborne Kart was, as the name implies, a kart racer set in Yharnam, before Sony's lawyers got involved. It's now called Nightmare Kart.

"Why are you making such absurd decisions, Sony?" Foulquier wrote. "Shutting down such a talented studio while you’re not even making a Bloodborne remake and on top of that, sending cease and desist letters to small fan projects? Are you serious?"

While we wait, perhaps futilely, for Sony to decide it wants to make a lot of money from a cherished classic, Bloodborne PC emulation has made a lot of progress of late.

Ubisoft announced sweeping changes to its business today, confirming more layoffs, major internal restructuring, and the cancelation of seven projects including Prince of Persia: The Sand of Time. At the same time it renewed its pledge to double down on generative AI.

In a document released today Ubisoft describes its "radically new value-creation" business model, which it claims will help it become "a more gamer-centric organization" thanks to a "rightsized and more agile organization".

It goes on: "The new operating model will further empower the execution of the Group's strategy, centered on Open World Adventures and GaaS-native experiences, supported by targeted investments, deeper specialization, and cutting-edge technology, including accelerated investments behind player-facing Generative AI."

Ubisoft's embrace of generative AI is nothing new. Last year the company revealed Teammates, a "playable generative AI research project" featuring AI-powered NPCs capable of dynamically interacting with the player. Teammates was an evolution of an earlier experiment dubbed Neo NPCs, and it's easy to see how it could be implemented in future Ubisoft games, especially its sprawling open world affairs.

But Ubisoft's "accelerated investments" in gen AI come at a time when studios are facing pressure to move away from the tech. When Larian revealed last month it is experimenting with AI, the blowback was so severe that the studio had to host a damage control Reddit AMA all about it. Arc Raiders' near perfect launch was clouded by its implementation of gen AI.

Attitudes towards gen AI will undoubtedly keep shifting with time (though it's hard to imagine it'll go the way of NFTs, which Ubisoft also embraced), but if Ubisoft announced Far Cry AI right now it's fair to say the backlash would be enormous.

Still, Ubisoft CEO Yves Guillemot seems extremely enthused by the tech, claiming last year that it's "as big a revolution for our industry as the shift to 3D."

Even with the hardware-hungry AI industry causing component shortages and sending memory prices soaring, there's perhaps one thing I wouldn't have thought to blame on the datacentres: massive font licensing fees for game developers.

Let me hit the backspace for a spell. At the end of November, commercial font service Fontworks ended its most affordable license plan, LETS (Leading Edge Type Solution), which originally allowed game devs to use their Japanese fonts for only 60,000 yen (or about $380) a year.

Parent company Monotype Imaging, which bought Fontworks back in 2023, then introduced a new pricing structure costing easily fifty times as much as the old plan (via Automaton). US-owned Monotype's pricing structure initially offered no local pricing for Japan, but still expected devs to cough up $20,500 per year, in addition to agreeing to a ridiculous 25,000-user cap on games using the licensed fonts.

Thankfully, the parent company has backpedalled a bit since, allowing developers to extend their licenses under the old affordable plan for the time being. But why such an astronomical price hike in the first place? Allegedly, heavy investment in AI that did not pay off for the company.

According to Monotype employee Mayur Pahwa, the company had "leaned heavily on the buzz surrounding artificial intelligence." He goes on to write, "Company leaders spoke enthusiastically about AI-driven typography solutions, hinting at a future where design and personalization would be automated at scale. Significant money and resources were funnelled into these projects, with entire engineering pods dedicated to experiments that promised to 'redefine the future of fonts.'"

Ultimately, a number of these AI projects were "shelved quietly, with little to show for the investment." What Pahwa calls Monotype's "AI Misadventure" allegedly resulted in waves of ongoing layoffs, and may now explain that astronomical font licensing fee; Monotype may be scrabbling to recoup its losses. Pahwa quotes one anonymous colleague, commenting on the situation, "We burned millions and got nothing. Now they’re cutting people to pay for those mistakes."

As for the Japanese game developers affected by the font fee hike, there's little good news to be found. Monotype doesn't just own Fontworks, and a number of other Japanese font makers feature among its portfolio; since Monotype was purchased by private equity firm HGGC back in 2019, the company has hoovered up Fontsmith, URW Type Foundry, as well as Hoefler & Co, according to Automaton. As such, there are few other reasonably priced services offering Japanese typefaces for devs to turn to.

Obviously, this is far from the only example of the twin villains of private equity or AI sinking their fangs into game development. To just point one more finger at AI, surging RAM and SSD prices mean Larian will be doing lots of optimisation in Divinity's early access, way ahead of the studio's original development schedule. As for private equity, I have only two letters for you: EA—though thankfully, Lincoln has many more letters and even a few hundred words on why EA's $55 billion acquisition has everyone terrified.

Bottom line, as Monotype's layoffs demonstrate, the concern is just how many people will be swept away by big tech's volatile whims.

AI-generated voice 'acting'—despite the fact I've never heard an example that hasn't immediately made me less interested in a videogame—has still become more ubiquitous in gaming. Arc Raiders, by all accounts a dang successful videogame, uses it for text-to-speech lines.

Speaking to PCGamesN, the voice actor for Baldur's Gate 3's Astarion, Neil Newbon (who knows a little bit about voice acting because he's won friggin awards for it), was very down on AI voiceover, echoing statements both he and his fellow BG3 cast members have made before on the subject:

"Comparatively speaking, the amount of money it costs to do those voice lines compared to the rest of the game's development is peanuts. When a game's successful, I don't really get why [developers] don't go 'well, at the time we couldn't afford to do it—it was too much or too difficult—but now we've been super successful, why don't we go back and actually redo the lines with actors?' That's an option; I'm just saying."

It's hard not to empathise with Newbon, here. The rational, sensible grown adult in me whole-heartedly agrees with him—compared to other development costs, hiring someone to go into a booth for a couple hours is basically nothing, especially when your game's making a million bucks.

The cynical goblin in my heart I've got locked up in a basement is, unfortunately, completely unsurprised that a studio like Embark would go for the cheaper, tackier option simply because it's… well, cheaper. We're in an industry plagued by layoffs where seemingly history-blind and overambitious executives make big, stupid decisions, then turn around and grind studios through crunch and burnout while they buy vintage cars with their paychecks.

Point being, we are not necessarily dealing with sensible people here—if all you care about's a budget, then why make something better if no-one's complaining all that much?

"I don't really think it's legitimate," Newbon continues. "If you're going to not record the lines in the first place and just use AI to take somebody's voice and manipulate it however the hell you want, that's a problem. You're robbing that performer of that day's fee, and you're robbing that performer of the ability to look after themselves or their family—most actors, notoriously, are not rich. Most of us struggle [for] our entire career."

Andrew Wilson, the CEO of EA, took home 260 times more money than what the median EA employee made this year. I share this not to argue with Newbon, more to just—sigh and nod along with him. Anyway: "The justification is difficult. 'Ah, we couldn't really afford it;' well, maybe find a way, now that you've made your money, to go back over those lines and do it better."

He also, rightly so, states that it all sounds bloody uncanny. Not enough to turn players away, clearly, but enough to make some (and certainly this writer) feel like they're playing a Temu version of the real deal:

"It's dull as hell. I don't believe it; it takes me out of the immersion. People are saying it helps the immersion because it's reactive. It takes me out of the experience because I just hear something that doesn't sound like a human being in jeopardy, or in combat, or excitement, or whatever emotion you're supposed to be aiming for. It doesn't feel real."

Oil and water, cats and dogs, Epic and Unity. Some things just don't go together. Until now, that is. Epic Games has announced plans to team up with probably its greatest rival in game development, Unity, to bring games made using the Unity engine to Fortnite. Epic also says, "Unity’s enhanced commerce platform" will come to Unreal Engine.

What's more, the collaboration will see Unity bring Unreal Engine support to its platform. To quote the killer excerpt from Epic's announcement:

"Developers will have the ability to publish Unity games into Fortnite, one of the world’s largest gaming ecosystems with more than 500 million registered accounts worldwide, and participate in the Fortnite Creator Economy. Also announced today at Unite, Unity’s Annual Developer Conference, Unity will be bringing Unreal Engine support to their cross-platform commerce platform."

If you're wondering what led these former mortal enemies to get together in such a sweeping collab', over to Tim Sweeney, Founder and CEO of Epic Games. “Just like the early days of the web, we believe that companies need to work together in order to build the open metaverse in a way that’s interoperable and fair,” he said.

Matt Bromberg, President and CEO, Unity, echoed that, commenting, “Choice and open systems create growth for everyone in the gaming ecosystem.” Of course, for Epic, this is presumably all about its ongoing quest to transform Fortnite from a mere smash-hit battle royale into a broad, sprawling and even more profitable gaming ecosystem. As they say in the UK's satirical tome, Private Eye, trebles all round!

Explaining Epic's rationale, Sweeney also told The Verge about his vision for a decentralised Fortnite that's more akin to the open web. “Fortnite’s still this big thing operated by Epic completely,” he said, “but there will be a day when you can, from the program that’s currently called Fortnite, go to other sites that are completely controlled by other companies.

"We have nothing to do with them. We make no revenue from them. We don’t have commercial agreements with them. They’re just like websites on the web.”

In that context, Unity's role here is to be a test candidate for that more open future. For now, Epic will ensure that content inserted into Fortnite is reviewed and "ratings compliant." But, eventually, the plan is for a fully open platform.

"We’d only have reviews for stuff that we host, and anybody would be able to put anything anywhere, and people would be able to browse to it like you do in a web browser today,” Sweeney said.

As for timing, the announcement says that "Developers will be able to manage pricing, promotions, and live operations with Unreal Engine, coming early next year," while additional details about the partnership and product timing will also be shared early in 2026.

Best gaming monitors 2025

All our current recommendations

Pick a category Best overallCore ComponentsChairs & DesksSystems & LaptopsScreensPeripheralsCases, Cooling & Storage

Choose a list Best gaming monitors 2025Best OLED gaming monitor 2025

---

BEST OVERALL

1. MSI MPG 321URX QD-OLED

Check Price

BEST 4K

2. LG Ultragear 27GR93U

Check Price

BEST BUDGET 4K

3. Gigabyte M28U

Check Price

BEST 1440P

4. Xiaomi G Pro 27i

Check Price

BEST BUDGET 1440P

5. KTC H27T22C-3

Check Price

BEST BUDGET 1080P

6. AOC C27G4ZXE

Check Price

BEST ULTRAWIDE

7. Asus ROG Swift OLED PG34WCDM

Check Price

BEST 32:9

8. Odyssey OLED G9

Check Price

BEST BUDGET ULTRAWIDE

9. ASRock Phantom PG34WQ15R2B

Check Price

BEST WOLED

10. LG Ultragear 32GS95UE

Check Price

BEST 1440P OLED

11. MSI MPG271QRX

Check Price

BEST BUDGET OLED ULTRAWIDE

12. Alienware AW2725Q

Check Price

BEST DUAL-MODE

13. Alienware AW2725QF

Check Price

---

Read the full guide: Best gaming monitors 2025

POWERED BY

After more than 30 years of running Activision, from 1991 to 2023, you might think that Bobby Kotick knows a thing or two about what it takes to run a successful videogame company.

He's a controversial figure, but Kotick also bought a failing company for $500,000 and turned it into a publisher that Microsoft spent $75.4 billion acquiring.

After listening to him suck up to Elon Musk on the Rushmore podcast (cheers TWIV), however, it sounds like he might have lost his strong instincts.

"Elon truly would be the best owner of any game company," he said. "No, he would. There's not a person who'd be more qualified to build and run a game company."

Musk's videogame credentials largely amount to liking videogames. He might actually be quite good at Quake and Diablo 4, too, though it's really hard to tell given his penchant for paying people to play for him.

In January, he streamed some Path of Exile 2, but seemed confused by basic game mechanics, which would have been fine if he hadn't been boasting about his skills, or if he wasn't playing on a geared-up, high-level character.

It turns out that he used an account boosting service, which he also admitted he did in Diablo 4. The 54-year-old billionaire didn't really see a problem with it, though. "It's impossible to beat the players in Asia if you don't, as they do!"

On the podcast, Musk still claimed to be "pretty good at videogames".

Musk also discussed one of his favourite videogames, Deus Ex, and once again used it to criticise the Covid lockdowns. Musk's failure to understand his favourite game has been going on for a while, and it's prompted Deus Ex writer Austin Grossman to call him out a few times.

"That kind of political weight and social satire is a real common thread between Deus Ex and Dishonored," Grossman told us. "It is the thing that Elon Musk likes, creepily. It is creepily in Elon Musk's worldview. So that may be its longest, worst legacy."

Not mentioned in the podcast was Musk's current gaming venture. Last year, he announced that xAI would develop games using AI, which he claimed would "make games great again". He said it would release an AI-generated game in 2026.

Former Sledgehammer boss Glen Schofield had a different take from Kotick, his former employer, calling Musk "full of crap" for claiming xAI could develop a game in a year.

Musk does seem to have slammed on the breaks, though. On the podcast he suggested we'd see fully AI-generated games appearing in the next three years.

Regardless, anyone who wants to cut out the humans who made all the games he loves so much, or who thinks that the truly atrocious AI tech demos we've all seen suggest that AI games will ever be anything other than hilarious mistakes, is not the kind of person who should own a game developer.

Kotick wasn't done with the flattery, though.

When Rushmore host and Endeavour CEO Ari Emanuel suggested that it was a mistake for Musk to buy Twitter, and that he should have invested in videogames instead, Kotick came to his pal's defence.

"For the betterment of society he needed to actually own X," he said. Yes, the man who let the bigots return to Twitter was doing it all for the betterment of society, and not just because he's a terminally online populist.

"No, he's way too capable," Kotick added. "The things he's working on, there are too few people in the world that actually could do that. So those are the priorities of how do you actually save humanity, improve society, he's got his priorities right."

One thing we can agree on, Bobby, is that it would be better if Musk spent more time tweeting and less time trying to ruin videogames. At least for me.

Remember Agent? After first mentioning it in 2007, Rockstar formally announced it in 2009—an espionage game in an open world, GTA meets James Bond, with all the cars-turning-into-submarines gadgetry that implies. After bouncing between different Rockstar studios for a while, Agent vanished off the radar, with its death finally being confirmed in 2021.

In a rare interview, Rockstar co-founder Dan Houser appeared on the Lex Fridman podcast recently to discuss all things GTA and Red Dead Redemption, though Agent also came up. "It had about five different iterations," Houser explained. "I don't think it works, I concluded. I keep thinking about it sometimes, I sometimes lie in bed thinking about it and I've concluded what makes them really good as film stories makes them not work as videogames—or I need to think through how to do it a different way as a videogame."

When you think about spy movies, from The Ipcress File to the latest Mission Impossible, while they may involve car chases and shootouts they're all about agents with a more singular mission than GTA's free-roaming loose cannons. "Those films, they're very very frenetic and they're beat to beat to beat," Houser said. "You've got to go here and save the world, you've got to go there and stop that person being killed and then save the world. And an open world game does have moments like that, when the story comes together, but for large portions it's a lot looser and you're just hanging out, and you're just doing what you want."

Which is why, he suggests, Agent never came together. The open world formula Rockstar had created, while well-suited to the life of a gangster or a cowboy, couldn't simulate being a spy in the same way. "That's why it works well being a criminal," Houser said. "Because you fundamentally don't have anyone telling you what to do. And we try and create external agency through these people forcing you into the story at times, but as a spy, that doesn't really work because you have to be against the clock. For me, I question if you can even make a good open world spy game."

The missions in the GTA London games where you pretend to be the Austin Powers-esque Endeavour Chambers may be the closest we ever get to a Rockstar open world spy game. Meanwhile, IO Interactive's 007 First Light seems like it might be more mission-based, as you'd expect from the Hitman studio. It's due out on March 27, 2026.

Have you ever been sitting there, watching an anime, and gone: Man, I wish I could get drawn into chaotic adventures just like Momo Ayase and Ken Takakura to help me work through my personal baggage, motivating me to become a better person? What am I saying, of course you've done that. We've all been there. Moving on.

Anyway, turns out there's a future in which something similar could be possible. Earlier this month, Yokohama City University's Minds1020Lab began trialling "anime therapy", as reported by itmedia (thanks, Automaton, for the translation).

Here's how it works: Participants will choose one of six anime avatars, each with their unique backstory. Then, in what looks to be a visual-novel style interface, a psychologist will take on the role of said anime character and conduct a counselling session. Some questionnaire-based psychological evaluation happens before and after these sessions, too.

Something interesting that's implied here, per Automaton, is the fact that a large chunk of these sessions will also be devoted to learning about these anime characters and their place in the fictional world that Minds1020Lab has set up, here.

And listen, it might be easy to see the words "anime-based therapy" and roll your eyes—and I definitely think we'll need to wait and see the findings of this study, which'll conclude in June 2026, before saying whether it's effective or not. But from where I'm sitting, this kinda rules.

I've been roleplaying in MMORPGs for a long time, and playing TTRPGs like D&D for 10 years, and while you shouldn't be treating your D&D table as a replacement for, like, actual therapy? I can't say I've never worked through my emotions while mucking about with fellow writers in World of Warcraft, or exploring my character's journey in a solid D&D campaign. Sometimes you figure things out in play.

In other words, I don't think the stories we tell each other are just frivolous nonsense, I think they do have a solid impact on our lives—fiction is powerful! And if anime therapy helps someone figure their stuff out, who cares whether it's delivered by a therapist in a mute-grey office or a therapist pretending to be a cool anime character in a visual novel interface: If it works, then it works.

Things are dire in the game industry at the moment. Which is, unfortunately, a sentence I've been writing for a lot of moments—two years of moments, actually. I am now writing it again, wincing in sympathy, as former Riot executive Greg Street has taken to LinkedIn to have an (understandable) moan about his studio's lot.

For context, Street worked on World of Warcraft (you might know him as Ghostcrawler) as well as Riot's still-fermenting League of Legends MMO, before heading out to form his own company in 2023. Fantastic Pixel Castle has been working on Project Ghost for a minute, before having the funding rug pulled out by NetEase earlier this month.

The studio's been on the hunt for money, but having a hard time, as per the post itself (thanks, GamesRadar+). "Please indulge me in feeling salty for a moment. I understand that almost nobody in the video game business is making deals right now, and those that do happen are very small.

"There are many reasons for this state of affairs, but one of the least excusable ones is some of the gigantic investments in games or studios in the previous few years that didn't pan out. A fraction of that could have created a lot of smaller games."

As our own resident MMO-knower, the cynic in me does understand that MMOs just aren't really in vogue anymore, and I'm tempted to lean towards saying well, what do you expect? But it's not as if Street is wrong, here. I mean, it's hard to believe that all those execs pumping out huge live service games are justified when Concord and Suicide Squad exist. Every so often you get a Marvel Rivals, but it's not a surefire thing.

"I'm not blaming every game that didn't hit big," Street continues, "Making games is very hard after all. But when you hear about projects where the team was burning money because the game vision kept changing, or the build was never stable, or playtests just weren't fun, or the leadership team was a revolving door, then you know that game was unlikely to ever deliver a good experience for players."

I mean hey, sometimes the playtests even were fun. Phil Spencer loved that MMO Zenimax were developing, before Microsoft decided to take it 'round the back of a shed with a 12-gauge. I don't know if Project Ghost would've hit big, but I'm already exhausted with the capriciousness of the gaming industry, and I report on it second-hand.

All of this, Street says, "can make it slightly frustrating when you do have, as we do: A team with a proven track record, that works well together with very high morale, without any leadership drama, that has taken a game halfway through development, whose vision has not strayed and whose tech is very stable, which hard data suggests players love and will play.

"And yet," Street says, "There is still no available funding." Rough.

In the world of PCs and PC gaming, debates are usually quite evenly distributed. AMD vs Intel. AMD vs Nvidia. WASD vs arrow keys. Cats vs dogs. You'll find arguments for either side most of the time. However, when it comes to Unreal Engine 5, comments from PC gamers are mostly on the side of "it just runs really badly." Epic Games has tried to counter such claims many times before, but with the debate still running, CEO Tim Sweeney has stepped in to say that the main reason why UE5 games don't run well is how they're developed in the first place.

He said this at the recent Unreal Fest event in Seoul, as reported by Korean site This is Game (via RedGamingTech). Press members asked about what Epic Games was doing to address optimisation issues for games developed on UE5. "The primary reason Unreal Engine 5-based games don't run smoothly on certain PCs or GPUs is the development process," replied Sweeney (alas, my grasp of Korean is non-existent, so hopefully the machine translation has got it spot on).

More specifically, he suggested that one particular problem is the choice of PC platform used in the formative stages of a game's creation. "Many developers begin by developing games for high-end hardware, then optimize and test on lower-spec devices in the final stages."

He's certainly not wrong. I've seen many examples of pre-alpha code being run on fairly high-end hardware, but that's only because those are the rigs that all the developers at the studio use. Once the bulk of the game is done, coders then work on getting it to run just as well on other configurations, but with less time being available to do this compared to the early stages of development, a comprehensive optimisation process isn't always possible.

While it is easier to scale up rather than down, one factor that can complicate the issue is when developers target consoles rather than PCs to judge performance. For example, the makers of Avowed (Obsidian Entertainment) picked the Xbox Series S for their benchmark hardware, and despite being a relatively weak platform, the PC version was decidedly wobbly on release. On face value, code for an x86-based console should be easy enough to translate to a PC, but they're disparate enough to make it quite a challenge.

Sweeney also acknowledged the fact that tweaking code for a range of platforms is simply a very time-consuming process. "Since optimization requires significant manual labor, we plan to provide automated optimization features for various devices, enabling developers to work faster and more easily.

"In addition to providing training on the overall testing and optimization process, it's crucial to raise awareness that optimization should occur early. If necessary, our engineers can directly intervene to raise awareness of optimization techniques and provide technical assistance."

The latter is arguably more useful than the former, because it takes time for Unreal Engine to be updated to include such tools and features, and even if they're released quickly, game developers are unlikely to use that version of UE for their game. In fact, because AAA development takes so long, many big releases come out on quite old versions of Epic's software package.

Take Stalker 2 as an example. That's an Unreal Engine 5.1 game, a version of the engine that's missing a huge amount of improvements that Epic has added over the years. The devs are currently working on shifting the whole thing across to version 5.5 (and perhaps may even make the jump to the much-lauded 5.6), but even if the update makes everything all super peachy, the damage has already been done to UE's reputation.

Avowed actually started on Unreal Engine 4, then switched to UE5, with the final release being version 5.3 in February of this year. You might wonder why game developers don't just update the tools they use the moment Epic releases a new version of Unreal Engine. After all, Unreal Engine 5.3 came out in September 2023, and 5.4 appeared in April of last year.

Well, if you've done a year or two of solid coding, you'll be familiar with the engine's quirks and foibles, and more importantly, you'll have stability. By that, I don't necessarily mean software crashes; it's about having as much consistency in the development process as possible, to avoid having to waste precious time adapting to new things. Sticking to the same engine, even if it's not the latest one available, isn't laziness—it simply involves fewer risks.

Ultimately, Sweeney is right in that the main reason why UE5 games don't often run well on PCs is the game development process itself, though I disagree that it's mostly down to choosing the wrong target platform. Of course, he's not going to blame his company's product, but at least Epic acknowledges the issue and it's trying to do something about it, just as Stalker 2's devs and others are as well.

I suspect that none of this will stop some folks from having the view that Unreal Engine 5 just runs really badly on PCs, because there will always be games based on that engine which get launched in a poorly optimised state. For those cases, either Epic and UE5 will incur the usual flak, or the developers themselves will be in the firing line.

Excuse me for a moment while I dance around my office, yelling 'same as it ever was', while a certain Talking Heads song plays in the background.