Hot Take: The best unknown builders/founders are not quiet. They are buried...
They post messy, unfiltered brainstorms.
They sometimes write badly, with no signs of AI polished posts.
They don’t have a content system
.... 🧵
But when they explain the problem/fix/challenge, you can feel they touched it with their hands...
X is weird because the interesting people often look small first.
Drop one unknown builder more people should follow
Codex is helping an astrophysicist simulate black holes better...
Chi-kwan Chan says current algorithms compute hit a wall. Near a black hole, plasma gets so hot and diffuse that fluid shortcuts stop working...🧵
3/3 That's the real shift: AI as a research multiplier, not a replacement. If science can test every claim, it may be one of AI's best use cases. Where else does this work?