ABAB Pedia

Ben Mann

Ben Mann (Benjamin Mann) is a co-founder of the AI company AnthropicAnthropicAnthropic is an artificial intelligence research and product company founded in 2021 that develops the Claude models, assistant, and tools such as Claude Code, and operates as a public benefit corporation. Dario Amodei and Daniela Amodei are its co-founders, and its research covers model safety, interpretability, and steerability.Open the full entry. He previously worked at OpenAI on infrastructure, efficiency, and safety for GPT-3 and is one of the authors of the GPT-3 paper; as of 2026 he works on Anthropic’s Labs team, which incubates experimental products.

Contents9 sections
Key facts

Early Life: Education and Early Work

Mann studied computer science at Columbia University and worked at research organizations such as the Machine Intelligence Research Institute and at startups focused on AI and automation.⁠[1][1]

Early Life: Waze Carpool at Google

Before joining OpenAI, Mann was a senior software engineer at Google, where he helped build Google’s carpooling service, Waze Carpool.⁠[1]

The GPT-3 Paper

Mann was a member of the technical staff at OpenAI, working on infrastructure, efficiency, and safety for GPT-3. On May 28, 2020, OpenAI researchers posted the paper introducing GPT-3, “Language Models are Few-Shot Learners,” on arXiv; Mann was listed second among its 31 authors, after Tom B. Brown.⁠[1][2][2][2][2]

Anthropic’s Series A

Mann is one of the co-founders of Anthropic, an AI safety company based in San Francisco. On May 28, 2021, Anthropic announced a $124 million Series A; the company was led by siblings Dario Amodei and Daniela Amodei, and its team had previously conducted research into GPT-3 and other topics.⁠[1][3][3]

Lessons From Building Claude

In 2023, as Anthropic’s co-founder and product engineering lead, Mann gave a talk at The AI Conference titled “Lessons Learned while Building Anthropic’s LLM Product, Claude,” covering his top lessons from building Claude, including techniques for aligning an AI system with human values and bootstrapping scalable oversight of model outputs. He has said his goal is to develop AI systems that are helpful, harmless, and honest.⁠[1][1][1]

Expansion of the Labs Team

On January 13, 2026, Anthropic announced the expansion of Labs, a team that incubates experimental products at the frontier of Claude’s capabilities; former chief product officer and Instagram co-founder Mike Krieger moved to Labs to build alongside Mann. This approach of experimenting at the frontier and then scaling what works had already produced Claude Code, the Model Context Protocol, Skills, Claude in Chrome, and Cowork.⁠[4][4][4][4][4][5]

Inviting Early Testers

Related Organizations, Websites, and Public Accounts

Sources

  1. Ben Mann - The AI Conference (opens in a new window)
  2. Language Models are Few-Shot Learners (opens in a new window)
  3. Anthropic raises $124 million to build more reliable, general AI systems (opens in a new window)
  4. Introducing Labs (opens in a new window)
  5. Anthropic’s Labs team gets a shake-up as Instagram co-founder Mike Krieger joins experimental AI unit (opens in a new window)
  6. Investing in Conway, the Creator of Underdog (opens in a new window)