Jack Clark
Jack Clark is co-founder and Head of Policy at Anthropic, an AI safety and research company he co-founded in 2021. He previously served as Policy Director at OpenAI and, before that, worked as a technology journalist covering AI and distributed systems for publications including Bloomberg and The Register. He writes Import AI, a widely-read newsletter about AI research and its implications, and has been involved in AI policy initiatives including the AI Index at Stanford University and the U.S. National AI Advisory Committee.
“Mirendil-Base is a pretrained LLM 'built by one researcher in 3.5 months, with less than $10m in compute... This includes all the research and small-scale experiments it took to get there, as well as the final training run.'”
“In 2027, huge amounts of software will be built via AI systems end-to-end autonomously developing whatever they need to do for whatever tasks they're being put to.”
“I now think it's more likely we'll see full RSI become possible in 2027, whereas before I thought it was more likely in 2028.”
“As we figure out how to get agents to productively coordinate... we could see even greater returns to scaling.”
AI-extracted from podcast / newsletter / paper summaries. May contain errors.