Tag
Epoch AI 发布 InnovationEval 评测，测试前沿 AI 模型能否独立发现媲美人类研究者的新机器学习算法创新；结果显示近期模型即使投入数千美元 GPU 时间也几乎未取得进展。
A coalition including OpenAI, Anthropic, and over 1,300 AI employees published a report proposing 23 policy recommendations for the US government to manage risks from automated AI R&D.
Behnam Neyshabur announces Mirendil, a new company focused on self-accelerating AI R&D to democratize AI and accelerate science, backed by $200M seed from a16z, Kleiner Perkins, and NVIDIA, with a team from top AI labs.
Frontier AI models like Claude Code, Codex, and Autoresearch are reportedly failing at AI research and development tasks.