Lovelace

“You can't just pour AI over your organization and expect everything to change,” says Lovelace CEO, Andrew Moore, in a new interview with TechTarget on AI hype, layoff anxiety, and what leaders actually need to know right now: “It irritates me when I see some of the foundation labs implying that AI is really easy, and that you can just bring in a bucket of AI, pour it over your organization, and everything changes. That's never the case, especially in anything safety-critical. You need as much care in designing how you're going to do it as you would in building a new headquarters for that company. There's a lot to think about, and a lot of creativity required in how you introduce automation.” https://lnkd.in/gGBYhE5V

Andrew Moore, our CEO, sat down with GTSC's Homeland Security Today's Kristina Tanasichuk to talk about one of the most pressing challenges facing government today: building AI systems that are transparent, traceable, and trustworthy enough for mission-critical decisions. Andrew breaks down how agencies can responsibly integrate AI into high-stakes operations without sacrificing accountability or accepting hidden risk. The goal is AI that decision-makers can actually trust, and how Lovelace is building the technology for precisely that. Watch: https://lnkd.in/ggZRF3Bb

We're excited to welcome four leaders to our new Strategic Advisory Board. Getting AI right in finance, defense, and other high-stakes environments takes people who've actually operated at that altitude: Diane Greene, who co-founded and led VMware, then ran Google Cloud. Few people have built and scaled enterprise infrastructure at that level. Lieutenant General H.R. McMaster, U.S. Army (ret.), served as U.S. National Security Advisor. He's seen firsthand what's at stake when decisions can't afford to be wrong. Philip Moyer, President and CEO of McGraw Hill and previously led Vimeo, brings a track record of scaling technology into organizations that serve millions. John Donovan, who led AT&T Communications and now sits on the boards of Lockheed Martin and Palo Alto Networks, giving him a rare view into both telecom and mission-critical infrastructure. We're also deeply grateful to Thomas Tull, co-chairman of TWG Global, whose early belief in Lovelace and guidance have helped us along our journey. There's a lot of hard work ahead. We couldn't ask for a more talented group to help us navigate it. Full announcement below. https://lnkd.in/esGXsSCY

Six weeks ago we showed a small cloud model + the Lovelace YottaGraph context engine could match Gemini Deep Research at a fraction of the cost. The question we got most: "can you do that with a local model?" So we swapped the last cloud dependency, the LLM, for an open-weight model running on a workstation GPU we had sitting in a closet. The result: a statistical tie with Gemini Deep Research, on the same investment banking benchmark. Zero API calls. Zero internet access. Zero per-report inference bill. We even stress-tested it off our home turf, on commodity supply chain outlooks, a domain our context engine wasn't built around, and it held up. The moat in grounded research was never the model, and now it isn’t the API either. It’s the context. https://lnkd.in/gMEJWZqv

Our latest benchmark: a locally hosted Gemma 4 model + the Lovelace YottaGraph context engine scored 9.83 against Gemini Deep Research's 9.87 (out of 10), statistically tied. The cost difference? It goes from roughly $7 per report down to a penny in electricity. No cloud API. No internet access. No inference bill. Just context, running inside your own walls. For enterprises handling sensitive or regulated data, this changes the calculus entirely. Frontier-grade research can now run end to end on open-weight models, inside your own walls, with nothing leaving the building. Context, not model size and not cloud access, is what determines research quality. And context is the one thing you can actually own. Full benchmark results on our blog, press release below. https://lnkd.in/gWZY-2JM

Big week for Lovelace. Just days after Andrew spoke on a panel at the Pennsylvania Defense and Innovation Summit, hosted by Senator Dave McCormick and attended by POTUS, he sat down live with CNN to talk about a problem that matters more than ever: how do you trust an AI's reasoning in mission critical, safety critical situations? Andrew's answer: "If an AI is doing chain of reasoning, every link in that chain needs to make sense. The critical piece is that we want to make absolutely sure that when an agent fills in a blank, the AI has the precise sources of information, ideally multiple independent sources, to fill in the blanks... This is all about making sure that information is available for a commander to know that they have the right information more quickly." That's the core of what Lovelace does: giving AI systems the traceability and grounding they need to be trusted in high-stakes decisions, and not just fast answers, but verifiable ones with every fact tracing back to its source. Reliable AI isn't a nice-to-have in defense, national security, finance, supply chain, or any serious industry. It's the whole game. https://lnkd.in/dS3CK2aA

PA's The United States Army War College was ground zero yesterday in the conversation around technology and national security - with The White House POTUS and United States Department of War SOW in attendance and hosted by Senator Dave McCormick. In the panel discussion for how AI is redefining national security - loved Dr. Andrew Moore's perspective for how *critical* academia and nimble startups are for defining and leveraging the next wave of AI technology (and of course you have to highlight #CMUAI!). It was an honor to have Gray Swan mentioned alongside Lovelace and Near Earth Autonomy as just a few examples of how our region's AI leadership is playing a key role. In addition to Dr. Moore, the panel included Antonio Gracias of Valor Equity Partners, Nir Bar Dea of Bridgewater, Troy Meink of the United States Air Force and was moderated by Mike Gallagher of Palantir Technologies.

Fantastic conversation with Ben Lorica, host of the podcast The Data Exchange, and Lovelace CEO Andrew Moore about why most "AI-ready" enterprises aren't AI-ready at all. Andrew's take: the model is not the hard part. The hard part is admitting your data isn't clean enough to trust an agent with it, and building the infrastructure to fix that, while adding billions of facts a week, without having an army of humans doing entity resolution by hand. Ben's show is in the top 0.5% most popular podcasts around, and he really knows his stuff when it comes to data, AI and ML. Check it out!

Great conversation between Ben Lorica and our CEO Andrew Moore on why agents need maps, not bigger context windows. The piece digs into how Lovelace's Elemental turns messy enterprise data (tables, PDFs, news, even satellite imagery) into structured, agent-navigable context graphs, and why entity resolution and auditability, not just retrieval, are the hard parts most teams underestimate. It's the infrastructure question every enterprise agentic AI effort eventually runs into. If you're not already following Ben's Gradient Flow newsletter, it's a great source for sharp thinking on data, ML, and AI. Worth subscribing!

Today in CIO Online - ‘When asked what prompted the decision to embark on the benchmark, Andrew Moore, CEO of Lovelace, and the former head of Google Cloud AI, replied...“We know that there is a crisis looming because of the expense of AI.” The Foundation Labs such as Google, OpenAI and Anthropic are, he said, “racing so fast towards more powerful AI that they, perhaps quite reasonably, are ignoring the expense in order to hit the finish line first. But if society is going to start deploying AI usefully, we feel it must be possible to do so without building thousands of new nuclear reactors and data centers.” Moore predicted that context, not compute, will define the next generation of AI systems... “Solve the context problem and you solve the cost problem.”’ https://lnkd.in/gcMy3jRf