Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Jeff Dean unveils Discovery Loop for automated AI research

AI summary

InfoQ 中文 AI&大模型 reports that Jeff Dean has founded Discovery Loop after leaving Google, where he worked for 27 years. At a Stanford HAI event, Dean said the startup is developing AI systems that design experiments, validate results and learn from prior research, initially focusing on machine learning research, machine learning engineering and hardware design. He also said the company has built an Agent Harness for large-scale agent collaboration and is running about 20 automated improvement loops, though these claims have not been independently verified.

In-depth insight

Discovery Loop offers a concrete organizational approach to AI-assisted research, but its infrastructure scale, productivity gains and ability to improve research workflows remain unverified vendor claims.

离开 Google 27 年,Jeff Dean 创业押注 AI 自我进化

在 Google 工作 27 年后,Jeff Dean 选择了重新创业。

这位参与过 MapReduce、Bigtable、TensorFlow、TPU 和 Gemini 等重要项目的技术老将,如今把目光投向了一个更具挑战性的方向:让 AI 不仅能够解决问题,还能自己设计实验、验证结果,并从过去的研究中持续学习。

在斯坦福大学 HAI 的一场公开活动上,Jeff Dean 首次系统介绍了新公司 Discovery Loop 的技术思路和研发进展。

Jeff Dean 透露,团队吸引了来自 Gemini、DeepSeek、Kimi 等模型项目的研究人员,也有参与 Go 编程语言项目的成员。公司成立前三周,他们进行了约 400 场面试。按照 Jeff Dean 在演讲中的说法,团队中大约 10% 的成员曾获得 ACM 计算奖,另有约 10% 拥有国际信息学奥林匹克竞赛奖牌。他们试图将长期从事大规模 AI 研究的资深工程师,与年轻研究人才组织在一起。

#Discovery Loop#Discovery Loop Agent Harness#Jeff Dean#Google#Gemini#DeepSeek#Kimi#Stanford HAI#World Labs#Fei-Fei Li

Original · 26 min readRead the original
Group chat

No comments yet. Start the conversation.