The "smartest" people are using AI the most. 43 of every 100 US workers with a graduate degree use generative AI. 10 of every 100 with a high school degree or less. 25 of every 100 across the whole workforce. We have seen this curve before. PCs in the 90s. Internet in the 2000s. Every productivity tool of the last fifty years. The educated tend to adopt first, capture the early wage premium, and the gap narrows over a generation. AI has the promise to flip that pattern. It amplifies expertise like every tool before it. It also delivers expertise the user never acquired. A high school graduate with a chatbot has access to legal reasoning, medical triage, financial analysis, and code generation that used to require years of training or a $400 hourly fee. Every prior wave of technology made it cheaper to do things you already knew how to do. AI makes it cheaper to know things in the first place. The trades worker who never had a lawyer gets tailored legal advice. The service worker who never had an accountant gets tailored tax advice. The single parent who never had a tutor for their kid gets one that moves at the right pace. The kid in the under-resourced school gets access to intelligence at the limits of their curiosity.

From "System of Record" to "System of Intelligence" In the next decade, you want to own the system of intelligence that pulls from the system of record, becomes the user’s one-stop shop for gaining context and taking action, and turns the SoR into something that’s primarily consumed at the API layer. The reasoning layer that sits above the database is where a new generation of companies is being built, and it’s where the majority of the next decade’s enterprise value of GTM software will end up. Full piece from a16z's Gio Ahern, Steph Zhang, and Alex Immerman:

The world owes itself another whole world economy. Global public debt has climbed from 68% of GDP in 2005 to 95% in 2025. History offers three paths out: austerity, inflation, or productivity. The UK brought debt-to-GDP down from ~250% after WWII to ~50% by 1980, mostly through nominal growth and inflation. Austerity is rare and painful. That leaves productivity, the option markets still underweight. If AI lifts productivity by even a point a year, the math on these debt stocks shifts from doom to manageable. The catch is timing. Productivity revolutions diffuse more slowly than markets expect. Debt service compounds immediately. Debt sustainability is really a productivity race. Whether AI-driven productivity arrives fast enough to outrun debt service may be the defining macro question of the decade.

The businesses getting the most out of AI right now aren't tech companies. They’re plumbers, agency owners, dentists, Etsy sellers. New data from @OpenAI (cc @RonnieChatterji) makes the pattern clear: tech startups account for about 5% of ~active~ U.S. users doing entrepreneurial work with ChatGPT. The other 95% are spread across services, retail, healthcare, and trades. AI adoption for entrepreneurship isn’t concentrated in tech. It’s happening inside everyday businesses, folding into routine work that used to be slower or outsourced or entirely overlooked.

The entire history of warfare can be read as a single, accelerating motion: extending the standoff between weapon and target. a16z's Daniel Penny, Alex Oliver, and Zach Chen on what this means for warfare today, and the urgency of building autonomous weapons:

You could freeze every frontier lab tomorrow. And… we would still have a decade of work ahead just figuring out how to use what already shipped. The data make that pretty clear. Even the leaders are only at ~23%. That means more than 3 out of 4 businesses still are not using AI in any meaningful way. Despite already having extremely capable models, true adoption is still early. The bottleneck on intelligence is now imagination, experimentation, and implementation across distribution, workflow redesign, training, trust, change management, etc. So... yeah... maybe we are "early innings".

Tokenized “real-world” assets (RWAs) have surged 10x in two years, now topping $30B — with nearly half held in U.S. Treasury debt. The growth reflects rising institutional demand to put traditional financial instruments onchain, from government bonds and commodities to equities and private credit. While U.S. Treasuries dominate today, the asset class is broadening, with more categories gaining meaningful share in recent quarters. Note: RWAs are traditional financial instruments such as government bonds, commodities, and equities that are represented onchain as tokens.

I started building the insights function for a company that tracked job changes and employment five days after the launch of ChatGPT in November 2022. In the last ~4 years, jobs have changed and companies (tech companies in particular) have changed how they are hiring. Some of this change is catalyzed by AI. Some of the change is a shift with interest rates and immune responses to market uncertainty. Jobs and hiring will keep changing. AI will accelerate this change. @DavidGeorge83's piece on the AI Job Apocalypse fallacy is a lens into what this change will look like.

We just announced the @a16z Design Engineer Fellowship 🧑‍🎨👨‍💻 What a remarkable time to be a designer, right? One day the timeline thinks your vocation is finished. "Design is the first casualty!" The next, we're so back - design isn't being eliminated, it’s absorbing engineering! It's the new system primitive. @pmarca called it the "Mexican Standoff" ... "AI empowers coders, designers, and product managers to believe they can do each other's jobs, eliminating the perceived need for specialized roles" ... and it's real. I won't pretend to know how the chips will fall, but what we do know is — when the frontier of a field is moving as fast as this one, it's incredibly hard to keep up. There's no curriculum, no playbooks where we're going. The rate limiter on acceleration is how quickly practitioners at the bleeding edge can share information, exchange notes on what worked/didn't in a high trust environment. How are you managing your teams? How are you driving & measuring AI adoption? What are you building in house, and what tools are you integrating? How does your team interface with other parts of your org, where do draw the line, set expectations? Right now, the people thinking these thoughts — making these decisions are scattered across hundreds of frontier companies, talking in small private groups or doomscrolling for tips, at risk of reinventing the wheel in isolation. We’re launching this fellowship to help: connect and bring you together in one place We’re looking for Design Engineering Fellows who: — are closing the gap between design and engineering - whether they came up through Figma and learned to ship; or moved up the stack and developed real creative + product sense. — ship real products, not just prototypes - and understand the gap between the two — hold both truths: AI changed everything and broke things along the way - the security vulns, the unmaintainable codebases, the distance between demo and production, and more — bring taste, systems thinking, and technical fluency to everything they build You might call yourself a Design Engineer, a Designer-who-Codes, a Frontend Engineer with taste, or something else entirely.. what matters is you care where this field is heading and want to put yourself among others who feel the same. If you’re interested, read more & apply below