Jared Kaplan

Kaplan is a theoretical physicist who was a professor at Johns Hopkins University before moving into AI. His collaboration with OpenAI produced the influential 2020 paper on neural scaling laws, which showed that model performance improves predictably with compute, data, and parameters — a result that shaped how the entire industry allocates training resources.

In 2021 he co-founded Anthropic, where as chief science officer he guides the research behind the Claude model family, including work on scaling, alignment, and interpretability. His scientific framing of scaling as a predictable phenomenon underpins much of how frontier labs plan their model roadmaps.

Kaplan is one of the research leaders who give Anthropic its scientific credibility, bridging fundamental research and the practical development of models that lead in coding and agentic tasks.