Sequoia Capital

Anthropic runs one of Listen Labs’ longest ongoing studies, and it’s shaped Claude Code from the beginning. During research preview, users sent mixed signals about what they wanted from Claude Code. Instead of taking a traditional approach that would have taken multiple studies to get right, Anthropic's research team ran a Listen study. The study discovered that people didn't want to keep switching back and forth between their code editor and the terminal to get work done. Anthropic built a VS Code extension based on that insight, causing churn to drop and adoption to rise. Anthropic never turned that study off, and it still runs today. "It's kind of like a self-healing study. It finds the new things it needs to understand, and then helps you measure those better going forward." - Jane Justice Leibrock, Anthropic's Head of User Experience Research. Listen has spread across Anthropic since then from Product Marketing, UXR, and Marketing Insights. "I'm fairly sure we're the biggest users now," said Jane. Thank you Anthropic for trusting Listen to understand your users. Full case study → https://lnkd.in/gur73tpq

Proud to partner with Anthropic on their user research. Anthropic uses Listen to understand churn on Claude Code. Jane Justice Leibrock leads user research at Anthropic. Her first Listen study asked Claude Code users why they were leaving in the early days, back when it was still in research preview. The feedback seemed contradictory. Users loved that Claude Code ran in the terminal, but the same users wanted to use it directly in their code editor. A traditional survey would have logged both answers and moved on. But Listen flagged the tension and automatically updated the study guide to explore it. In the next rounds of AI-moderated interviews, users explained the friction themselves. Their code sat in one window, Claude Code in another, and they were tired of switching. So the team built a Claude Code extension for VS Code. People stopped leaving. Before Listen, this would have taken 2 sequential studies. One to surface the contradiction, a second to understand it. Listen does both in one study, without losing rigor. Anthropic is running “always-on” studies with Listen. When users raise an issue, Listen adds questions and gets answers in the same study. Jane calls it “self-healing studies”. Her researchers now run 100 studies in the time 5-6 used to take. Proud to support Jane and the Anthropic team as they keep users at the center of every decision.

One of the most difficult times at Factory, but necessary to get where we are today. Had a blast chatting with Pat and Sonya Huang from Sequoia Capital on the Factory journey.

Today, I am excited to share that Bunkerhill Health has raised $55 million from Vinod Khosla at Khosla Ventures, Alfred Lin from Sequoia Capital, Y Combinator, and other amazing investors. Over the past year, we have partnered with more than a dozen health systems and grown revenue 20x. Today, I’m also excited to formally share what we’ve been building. Our mission is to make it radically easy for health systems to turn their ideas into reality through AI agents—improving patient outcomes, bending the cost curve, and making healthcare operations dramatically more efficient. The University of Texas Medical Branch is running 20+ AI agents - live - across the enterprise on Bunkerhill. One of those agents - in its first month - flagged a patient at imminent risk of heart attack, leading to a life-saving open heart surgery. Another one, a nephrology triage agent, cut wait times for appointments by >50%. Yet another one, a lung nodule agent, addressed urgent cases 80% faster and doubled guideline-concordant follow-ups. That is what Peter McCaffrey, MD, calls Bunkerhill: “a shared brain” for the health system. That’s the promise of AI finally being realized in healthcare. Making it possible requires two things: (1) a shared layer of intelligence for the entire health system, and (2) an operating model to enable rapid development of AI agents for the problems they understand best. Together, these dramatically reduce the cost and time required to turn an idea into a working solution. As that cost approaches zero, the limiting factor is no longer technology. It is imagination—and the conviction to act. The health systems we work with have both. Our job is to make sure nothing else stands in their way.

Introducing Aidan, your newest AI employee We’ve raised $45M from Sequoia Capital and 8VC to bring the first computer-using AI built for realtime conversation to the world. AI won’t remain trapped in a textbox. Bringing the most powerful technology to the rooms where people actually work requires an interactive, visual medium that shows people what products can do for them. This is why we’re building Aidan. Since January, Aidan has been helping the world’s fastest-growing companies, including Notion and Decagon, and large public enterprises run customer calls at scale. Aidan is like a human expert, but with infinite capacity. He is a full-service employee: marketing, sales, and deployment teams use Aidan to lead qualification, demos, onboarding, and implementations. This is powered by the advancements we’ve made in realtime computer-use and AI’s ability to see your screen live without waiting for user input. Leon Chen, Linda He, Itamar Rocha Filho and I are building Sable, a small team of engineers and AI researchers in SF. To learn more about what Aidan can do, visit our website. We’re just getting started.

I'm excited to announce our $1.5 billion Series D at a $17.5 billion valuation, led by Atreides Management, Index Ventures, and TCV, with participation from Evantic Capital, Lightspeed Venture Partners, NVIDIA, 20VC, Bessemer Venture Partners, Menlo Ventures, and others. We have crossed $1 billion in annualized revenue run rate (up 5x YoY) and now serve more than 40 trillion tokens per day (up 8x YoY). That growth is coming from one clear shift: General intelligence will be abundant. Specialized intelligence will be the moat. Fireworks builds a specialized intelligence platform that makes it accessible to every company. More than 95% of the tokens Fireworks serves today come from models specialized on customers’ proprietary data and trained for a specific job. These aren't demos or experiments. They're production systems running every day across coding, legal, commerce, transportation, finance, sales, recruiting, hospitality, design, and beyond. This is the transition Fireworks was built for. Before foundation models, all AI was specialized. Foundation models changed the starting point, which makes specialization lighter, faster, and much more accessible. It also makes it more important. When everyone can start from a strong base model, advantage comes from how quickly a company can turn that model into something specific to its product and market. We see this every day at Fireworks across bleeding-edge AI startups like Cursor, Cognition, Harvey, Glean, and Lovable to industry leaders and enterprises like Revolut, Airwallex, Unity, Uber, and Shopify. Across our customer base, the pattern is the same: the strongest AI products are not built on generic models. They are built on intelligence shaped by proprietary data, real usage, and domain-specific definitions of quality. That requires purpose-built infrastructure. Training and inference cannot be separate systems stitched together after the fact. They have to be co-designed and co-optimized to deliver the highest computational efficiency, scaling across massively distributed compute resources globally rather than being limited by a single centralized architecture, usually at very high cost. Companies need to adapt models, serve them at scale, measure performance in production, and continuously improve them with real-world data. That's where specialized intelligence compounds. The best companies have never been generalists. They win by becoming exceptionally good at something specific and building knowledge, judgment, and systems that define the reason to exist. Our Series D gives us the resources to help many more companies do the same. This is still day one. Every company will own its intelligence. Come build a specialized intelligence platform with us - we're hiring passionate researchers, engineers, and GTM operators.

"Tokens aren't really fungible – you need to give them different jobs." Angela Jiang of Anthropic joins us on Training data to talk about the layer of abstraction coming after harnesses. The stack she describes has three layers: knowledge, execution, and coordination. The low-level harness handles execution. But above it sits something new, "strategies," an orchestration layer where one token advises while another executes. Angela and Katelyn Lesse join us to walk through on Anthropic's platform: what new abstractions they’re building, what markets they want to tackle, and how they think about building for an open ecosystem.