AI is generating code at unprecedented speed, but nobody's checking if it actually works. Superintelligence is meaningless if it's not verified. That's why Carina Hong (Carina Hong) is building an AI mathematician at Axiom that doesn't guess, it proves. It scored 120/120

LLMs cracked language. But what about the other half of your brain? We’ve only solved half of AGI - the creative one that understands language. The data and math side? Still unsolved. BUT according to Jeremy Fraenkel, we are very close to changing that… cc Jeremy Fraenkel

1,100+ teams applied. 12 made it. We’re one of them. Base Batches 003. Bank like the 1%. Become Liminal.

The reason most people don't see the AI wave? They don't know how to ask questions. And even if you do, you need to manage your AI agents the same way you'd manage your employees. The progress isn't slowing down. It's speeding up. And most people have no idea what's coming.

I really enjoyed this discussion with @jason Carina Hong and Kanjun 🐙. Thanks This Week in AI for having me.

"You can't download your entire conversation history from OpenAI or Claude" They know more and more about you, and it becomes increasingly difficult to leave. They're going to understand you better than anyone. You already feel it. Kanjun 🐙 @jason

Stop treating AI agents like productivity tools. Treat them like a child you’re raising. In a world of automated "slop," the only value that remains is human judgment and taste. Demi Guo (Demi Guo) is the co-founder of Pika, where she is pivoting from AI video generation

The world's most valuable data is still waiting to be unlocked. Lin Qiao (Lin Qiao) built the inference layer for Llama at Meta and now runs Fireworks AI, processing tens of trillions of tokens per day. She joins @jason to discuss how open-source models are converging fast and