# Jared Kaplan

> Jared Kaplan is an American theoretical physicist, machine-learning researcher, Anthropic co-founder and chief science officer. After studying holography and quantum gravity, he joined OpenAI and led the 2020 paper that mapped language-model performance against model size, data and compute. At Anthropic, he now helps direct technical research and oversees the difficult translation of those predictive curves into safety decisions for increasingly capable AI systems.

- **Role:** Co-Founder and Chief Science Officer at Anthropic
- **Organizations:** Anthropic, Johns Hopkins University, OpenAI, SLAC National Accelerator Laboratory, Stanford University
- **From:** Chicago, United States
- **Nationality:** American
- **Education:** BS in Physics and Mathematics, Stanford University, PhD in Physics; thesis: Aspects of Holography; advisor: Nima Arkani-Hamed, Harvard University
- **Known for:** Lead author of the 2020 Scaling Laws for Neural Language Models paper, which documented power-law relationships among language-model performance, model size, data and training compute., Co-author of Language Models are Few-Shot Learners, the paper that introduced GPT-3., Co-founder of Anthropic and head of its technical research efforts.

## Career timeline

- **2001** — Member of the United States Physics Team and graduate of the Illinois Mathematics and Science Academy.
- **2005-2009** — Completed a Harvard physics doctorate as a Hertz and NSF Graduate Research Fellow.
- **2009-2012** — Worked as a research associate at SLAC National Accelerator Laboratory.
- **2012** — Joined Johns Hopkins University as an assistant professor of physics and astronomy.
- **2014-2016** — Held an Alfred P. Sloan Research Fellowship.
- **2015** — Received an NSF CAREER Award for work on conformal field theory and quantum gravity.
- **2018** — Promoted to associate professor at Johns Hopkins.
- **2019** — Joined OpenAI as a researcher.
- **2020** — Led the Scaling Laws for Neural Language Models paper and co-authored the GPT-3 paper, Language Models are Few-Shot Learners.
- **2021** — Co-founded Anthropic with former OpenAI colleagues.
- **2023** — Became Anthropic's chief science officer.
- **2024** — Named Anthropic's Responsible Scaling Officer, charged with overseeing implementation of its Responsible Scaling Policy.
- **2025** — Named to the TIME100 AI list after applying AI Safety Level 3 protections to Claude 4.

## Achievements

- Lead author of the 2020 Scaling Laws for Neural Language Models paper, which documented power-law relationships among language-model performance, model size, data and training compute.
- Co-author of Language Models are Few-Shot Learners, the paper that introduced GPT-3.
- Co-founder of Anthropic and head of its technical research efforts.
- Helped pioneer Constitutional AI, according to the Hertz Foundation.
- Recipient of a 2005 Hertz Fellowship, a 2014 Sloan Research Fellowship and a 2015 NSF CAREER Award.
- Selected for the 2025 TIME100 AI list.

## Latest updates

- **2026-09** — Forbes included six Anthropic co-founders in the 2026 Forbes 400, estimating each fortune, including Kaplan's, at $15.5 billion.
- **2026-02** — Anthropic revised its Responsible Scaling Policy, removing an earlier categorical pledge to halt training when safeguards were not ready; Kaplan described the change as a pragmatic response to competitors continuing to advance.
- **2026-01** — Kaplan told The Atlantic he would prefer AGI to arrive in 2032 rather than 2028.
- **2025-08** — TIME named Kaplan to its TIME100 AI list, highlighting his decision to release Claude 4 under AI Safety Level 3 protections.
- **2025-06** — At TechCrunch Sessions: AI, Kaplan discussed AI agents, compute constraints and Anthropic's decision to restrict Windsurf's direct model access.

## Links

- Website: https://physics-astronomy.jhu.edu/directory/jared-kaplan/
- LinkedIn: https://www.linkedin.com/in/jared-kaplan-645843213

---

Profile page: https://yespress.io/jared-kaplan
Published by YesPress — https://yespress.io
Last updated: 2026-09-26