Dwarkesh Patel has sharply revised his timeline, arguing that AI matching the best human AI researchers may be only a couple of years away, possibly months. His central argument is compute scale: OpenAI's end-of-2027 fleet could support roughly 2.8 million GPT-3-training-run-sized experiments per year, and even with a third of the fleet dedicated to experiments, 10,000 agents could each run one every four days.
Termsslug: automated-ai-research
Automated AI Research
AI 研究自动化
A research paradigm where AI agents autonomously design, run, and evaluate experiments in place of human researchers.
Curated Intelligence
Hacker News AIT2·Models
62 pts
IT之家 · 智能时代T2·Industry
78 pts
Former Anthropic and OpenAI researcher Jacob Coxon testified at a NYC council hearing that top labs are targeting fully automated AI research by 2027-2028. Based on internal progress, he estimated the probability of humanity losing control to AI exceeds 50%, criticizing the industry's 'move fast and break things' approach as dangerously inadequate for frontier safety.