Jared Kaplan
Co-founder and Chief Science Officer of Anthropic.
At a glance
Roles
Papers to start with
Scaling Laws for Neural Language Models, Language Models are Few-Shot Learners
See all 2
Also on this page: Key facts · Links · Sources · Related pages
- Jared Kaplan is a theoretical physicist and a co-founder of Anthropic, where he is chief science officer.
- He holds a physics PhD from Harvard and has taught physics at Johns Hopkins since 2012.
- He joined OpenAI as a researcher in 2019.
- He led the 2020 paper on scaling laws for neural language models and co-wrote the GPT-3 paper.
- Anthropic named him its Responsible Scaling Officer in October 2024.
Why it matters
- For anyone who wants to know why AI models keep growing: his scaling-laws work showed how performance improves with size, data and compute.
Key facts
- Current role
- Co-Founder and Chief Science Officer, Anthropic1
- Education
- BS in physics and mathematics, Stanford; PhD in physics, Harvard (2009), advised by Nima Arkani-Hamed2
- Before AI
- Postdoctoral fellow at SLAC and Stanford; physics professor at Johns Hopkins since 20122
- OpenAI
- Researcher from 2019; worked on Codex2
- Responsible Scaling Officer
- Named at Anthropic in October 20242
- Honors
- Hertz Fellow (2005), Sloan Research Fellowship, NSF CAREER Award2
- Net worth
- $3.7 billion, Forbes estimate (2026)2
Roles
Founded
Papers to start with
The line under a date fills as it gets older: empty when new, full at 2 years.
-
Scaling Laws for Neural Language Models (opens arxiv.org)2
How language-model performance improves with size, data and compute.
-
Language Models are Few-Shot Learners (opens arxiv.org)2
The GPT-3 paper; he is a co-author.
Links
Sources
Last checked against its sources today
Is something on this page wrong or out of date? Tell us
The line under a date fills as time passes: empty when new, then one equal share each for a day, a week, a month and a year. For episodes and other items it is full at 2 years; for "last checked", at a year. A striped line means the date is known only to the month or year.