Andreessen Horowitz

Without AI-related infrastructure investment, there is very little investment growth, at all:

In comparison to the previous tech-leap forward driven by PCs, when it comes to AI, the cost of intelligence is dropping much more quickly. It took PCs almost two decades to achieve the affordability gains that AI has achieved in ~3 years. That’s a remarkable trajectory, and as with computers, it seems like falling costs are helping push AI demand up-and-to-the-right (with a lot more runway to go).

A16Z American Dynamism presents: Ulysses. Mariana Minerals. Radiant. Coming soon.

AI appears to be accelerating entry-level hiring in a big way: Per an analysis of Revelio Labs’ job data and Ramp’s spending data, “High-Intensity AI Adoption” corresponds to a ~6pp increase in entry-level headcount, two years after adoption. “Low-Intensity” AI-Adoption, by contrast, corresponds to a ~.50 pp decline. Now, there are few ways to interpret this, from “AI is creating entry-level hiring,” to “AI adopters are the growth companies and therefore of course they’re the ones doing the hiring,” to “Ramp’s data may not be a representative sample of the economy writ-large, and therefore it’s hard to conclude anything at all.” That’s all well and good, but one interpretation that almost certainly does not follow is that “AI is killing entry-level jobs.”

11x's revenue agents are generating hundreds of millions of dollars in pipeline for their customers. We sat down with CEO Prabhav Jain and went through their full internal AI operating stack. They are only selling to customers who will see incredible value from 11x. So he built a Claude skill every sales rep uses. It pulls in emails, call notes, Slack messages exchanged with prospects. Reps only close deals with confidence >85%. Claude integrations pull every call from Granola, Slack, and email exchanged across the whole cycle. Then it generates a Notion doc and a kickoff deck. Agents are good at context, humans are good at service. Additionally, 11x runs a company-wide fleet of internal codebase-read agents they call lemmings. This is built with LangChain's LangSmith Fleet, which lets you describe in plain english a fleet of agents that work across your internal code and tools, and at mention them from Slack. But 11x also realized that as coding agents wrote more PRs, the bottleneck would move to verifying PRs. So early on, they decided to invest heavily in building their own code harness. Prabhav also built a simple fleet of personal agents with intel for each day, all running on Claude Cowork scheduled jobs. PS – 11x is hiring in sales, success, growth, engineering, and design.

We're thrilled to lead Runta's seed round. AI models have required rethinking the GPU stack. And now agents are doing the same thing for CPUs. Ironically, compute's evolution over the last decade - driven by developers pushing for higher abstraction, opinionated state management, and devX - has made it less suited for AI. Runta is rebuilding the execution layer agents actually need - a computer that's extremely efficient, runs locally or in the cloud, with the security and policy controls to run agents safely in the enterprise. Or anywhere. We've known Guanlan Dai for nearly a decade, since his days leading Cloudflare's edge proxy and Kong's core proxy. Martin Casado is joining the board. Congrats to Guanlan and the Runta team. By Martin Casado, Yoko Li, Guido Appenzeller

Marc Andreessen's advice to college students: "Gain AI superpowers. I think it's actually very straightforward." "You have the enormous stroke of luck that you have arrived at the moment in which there is this new capability for augmenting human ability on a thousand fronts at the same time, that's just dropped into our laps, and it's going to get much better from here." "You are gonna have the opportunity to have this be something that is absolutely key to your skill set and key to everything that you can accomplish as a professional or as a creative for the next 50 years." "I would just lean in incredibly hard on that. Walk into every job interview with, 'Here's my portfolio, resume, whatever. Here is how I use this technology. Here are the capabilities that I'm bringing to the table.'" With Erik Torenberg on MTS Live

Reactors have a surprising amount of detail. A few years ago, John Salvatier wrote an essay about building basement stairs and concluded "reality has a surprising amount of detail." I couldn't agree more. We are building microreactors at Radiant. The small details, the ones that often seem inconsequential, are the ones that de-risk your product for your customers. A bolt is just a bolt, right? When I was at McMaster-Carr, we sold 91,130 distinct products called "a bolt." Not one of them, on its own, is necessarily the right bolt to hold the lid on a nuclear pressure vessel. We have 56 bolts on the top of our reactor pressure vessel alone. If you're trying to do something that has never been done before, you can't assume the details will sort themselves out. The more difficult your mission, the more the details are critical to understand. This is why we built our five-phase test campaign at Idaho National Laboratory's DOME facility. Each phase provokes specific behavior and measures it against what we predicted. Where they disagree, a detail was hiding. Winning is the hundredth reactor, and the thousandth, every bolt the right bolt, every pipe, sensor, and line of code built so reliably that nuclear becomes boring instead of a spectacle. As Salvatier said, “If you wish to not get stuck, seek to perceive what you have not yet perceived.” Don't take anything for granted. Find the details you cannot yet see. I wrote more about this, linked in the comments.

Marc Andreessen on why history is one of his greatest source of inspiration: “I'm a big fan of history.” “There's a sort of a classic cliché in the Valley, which is like we don't respect history very much... we're all about the future, right? We're building the future. And so I think there's this natural tendency to assume that the past isn't very relevant.” “My general view is the people who came before us had a harder time doing what they did than we do because the world was in a more immature state.” “Everybody takes it for granted, and then there's a person or some set of people who think, ‘No, no, there's a better and, you know, different and better way to do things.’” YC Startup School 2016, with Qasar Younis