Benjamin Mann
Co-founder, Anthropic
關於
Benjamin Mann is a software engineer and co-founder of Anthropic, the AI safety company behind Claude. Before co-founding Anthropic, he worked at Google, OpenAI, and the Machine Intelligence Research Institute, after shifting his focus toward AI safety. His personal bio says he spent six years as a software engineer at Google before moving into AI safety work, and later helped found Anthropic. At Anthropic, he has worked on product engineering and the company's mission of building safer transformative AI systems.
主要貢獻
- Co-founded Anthropic, one of the leading frontier AI labs focused on AI safety and alignment
- Worked at OpenAI before Anthropic and has been described as one of the architects of GPT-3
- Helped bring AI safety concerns into product and engineering strategy at a major AI company
- Previously worked at Google and MIRI before focusing on frontier AI systems
- Publicly discusses AGI timelines, AI safety, labor-market disruption, and responsible deployment
影片與訪談
Anthropic co-founder: AGI predictions, leaving OpenAI, what keeps him up at night | Ben Mann
Interview covering AGI predictions, why Mann left OpenAI, Anthropic's mission, and risks that concern him.
View Details
No Priors Ep. 118 | With Anthropic Co-Founder Ben Mann
No Priors interview with Ben Mann on Anthropic, Claude 4, frontier AI progress, and AI safety.
View Details思想連結
Dario Amodei
曾經合作執行長暨共同創辦人・Anthropic
Mann 是〈Language Models are Few-Shot Learners〉的第二作者，而 Amodei 署名在最後——GPT-3 這篇讓 scaling 再也無法被否認的論文，作者名單的兩端剛好是他們兩人。2021 年，兩人又在同一波離職潮中出走，共同創辦 Anthropic。這組連結標記出本站一再回望的裂縫：證明 scaling 有效的那群人，也正是斷定這件事需要換一個機構來做的人。
arxiv.org · en.wikipedia.org
Amanda Askell
曾經合作哲學家暨 AI 對齊研究員
相隔一年的兩份作者名單上都有他們：2020 年的 GPT-3，Mann 參與打造那個讓所有人意外的模型；2021 年 Anthropic 的對齊論文，Askell 則主導追問這樣的模型該是什麼樣子。工程師與道德哲學家並列在同一份名單上，並不是研究團隊常見的樣子，卻很接近 Anthropic 創立時對「這件事究竟需要什麼」的主張。
arxiv.org · arxiv.org
