-
open source
2026-07-20
↺ 立场转变
After a cyberattack, turned to a Chinese open model: open models let them act without permission (a shift from 2024 concerns).
«Open models let us do that work without asking anyone's permission.»
-
open source
2026-07-14
Keeping AI closed doesn't make it safer but more dangerous, creating power asymmetry.
«You make it more dangerous because you create asymmetry of power»
-
safety
2026-07-14
The main AI risk is concentration of power; safety comes from a level playing field and transparency.
«The biggest risk in AI is concentration of power... make the world safer by leveling the playing field and transparency.»
-
products
2025-11-18
Predicts a future of many customized, specialized models rather than one universal model.
«a multiplicity of models that are more customized, specialized»
-
compute
2025-11-18
↺ 立场转变
Sharp framing: the bubble is in LLMs, not AI, and it may burst next year.
«I think we're in an LLM bubble, and I think the LLM bubble might be bursting next year.»
-
products
2025-06-16
Reiterates the small-models thesis: bigger isn't always better; specialized models are faster, cheaper, more accurate.
«bigger isn't always better... smaller specialized models are usually faster, cheaper and more accurate»
-
products
2025-05-29
Unveiling humanoids, stresses robotics must be open and affordable so it isn't dominated by a few with black boxes.
«these robots are open source, so anyone can assemble, rebuild, understand them, and affordable»
-
china us
2024-12-03 ~
↺ 立场转变
Voices concern about cultural bias in Chinese open models (Tiananmen example).
«A chatbot asked about Tiananmen won't respond the same way as one developed in France or the U.S.»
-
agi timeline
2023-06-29
Anti-hype view: AI is a new paradigm for building tech, but not Skynet or a sentient being.
«It's not Skynet or a super sentient being.»
-
regulation
2023-06-22
Testifying to the US House: open source prevents black boxes and makes companies accountable.
«Open-science and open-source prevent black box systems, make companies more accountable»
