.@christinahkim says the frontier isn't benchmarks anymore. It's usage. Eval scores are saturated, but daily life isn’t. The real signal of progress is how many people use AI to get real things done. That’s how we’ll know we’re approaching AGI.

Post-training is an art form. @christinahkim says shaping model behavior means sculpting character. Every reward is a tradeoff. GPT-5 was a reset: balancing helpfulness with honesty, warmth without sycophancy. The assistant shouldn’t just answer. It should feel right to use.

"Move fast and make things." @pmarca says this is the real motto of the AI era. Nothing broke. Keep building.

the a16z pod is nothing without its people We are breaking down everything you need to know about GPT-5 with the @openai team. 📸 @sarahdingwang @eriktorenberg @christinahkim @isafulf

Permissioned data access appears to be under siege. If data access becomes a closed-loop system, what breaks — and what is lost? I’ve met hundreds of B2B AI startups. Here's how I think this reshapes the landscape and how founders can respond 👇

.@pmarca in 1996 with @charlierose: "The internet is everywhere. In 10 years, we probably won’t even be talking about the internet." Everyone connected. Any network. Infinite use cases. The internet became the substrate. Everything else followed.

.@pmarca says embodied AI is already here. From autonomous drones to self-driving cars, and it's just the beginning. " The general purpose robotics thing is going to happen, and it's going to happen in the next decade… at giant scale." The U.S. has a choice: lead the Fourth Industrial Revolution, or live in a world of Chinese robots.

Everything technology touches collapses in price. @pmarca says that’s why TVs, video games, basically anything with electronics in them, keep getting cheaper, while housing, healthcare, and education keep rising in price. "Technology hasn’t really touched them… Productivity growth hasn’t happened in them." The challenge now: bring technology where it’s needed most.

App generation isn’t a winner-take-all market. The best platforms specialize in prototypes, personal tools, or production apps. They’re complements, not competitors. Using one makes you more likely to use another. New thesis by @illscience and @venturetwins on where this is going.

App generation is one of the fastest growing AI spaces. Cos are scaling to $100M+ ARR at record pace, seemingly in a race to dominate the market. But we think it won't be "winner takes all." The market will segment, with different winners in different areas. Our thesis 👇