While I believe in recursive self improvement, I think we are grossly underestimating how hard it might be. Results from evals like PostTrainBench suggest you can train AI to make incremental progress but novel ideas are harder to generate.
maybe the missing ingredient for true generality is not another collection of capabilities, but a sufficiently general mechanism for acquiring capabilities
image/video gen models are meant to democratize creativity, like anyone can now generate whatever they want but what actually seems to be happening is it gets read as cheap, try-hard, or low-effort rather than creative
@lalitcap@zoink 2 current LLMs are pattern learners & we dont have evidence they're conscious. humans also do pattern matching & we're conscious, but whether consciousness is just pattern matching we dont know yet
this is a good sign. as ai keeps advancing, having strong open-weight models alongside closed ones keeps the field more accessible and more accountable
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.
AI will transform every industry, power every company, and be built by every country.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
The world needs both frontier closed models and frontier open models.
https://t.co/AUKzoQ5Ikb