Sam McCandlish

# Sam McCandlish

Sam McCandlish is a theoretical physicist, co-founder and Chief Architect of Anthropic, the AI safety company that develops the Claude family of…

- [![Nextomoro](https://nextomoro.com/content/images/size/w100/2025/02/nextomoro-faviv1-d9c3f3ecaf595f91.png)](https://nextomoro.com/author/nextomoro/ "Nextomoro")

[Nextomoro](https://nextomoro.com/author/nextomoro/)

May 3, 20268 min read

![Sam McCandlish](https://nextomoro.com/content/images/size/w30/2026/05/sam-mccandlish-2.png)

# Bio

Sam McCandlish is an American theoretical physicist and AI researcher, co-founder of [Anthropic](https://nextomoro.com/anthropic/), the public-benefit corporation that develops the [Claude](https://www.anthropic.com/claude?ref=nextomoro.com) family of large-language models. He was previously a researcher at [OpenAI](https://nextomoro.com/openai/), where he was the second-named author of the January 2020 paper " [Scaling Laws for Neural Language Models](https://arxiv.org/abs/2001.08361?ref=nextomoro.com)." As of May 2026, he serves as Anthropic's Chief Architect with a focus on pre-training and large-scale model training, a role he assumed in October 2025 after Rahul Patil joined as Chief Technology Officer.

## At a glance

- **Education:** B.S. and M.S. in mathematics and physics, [Brandeis University](https://www.brandeis.edu/?ref=nextomoro.com) (2012); PhD in physics, [Stanford University](https://physics.stanford.edu/?ref=nextomoro.com) (2017), advised by [Eva Silverstein](https://profiles.stanford.edu/eva-silverstein?ref=nextomoro.com); thesis "Depth Perception in Holography." Postdoctoral position at [Boston University](https://www.bu.edu/physics/?ref=nextomoro.com) Physics (approximately 2017 to 2018).
- **Current role:** Co-founder and Chief Architect of [Anthropic](https://nextomoro.com/anthropic/), since October 2025; previously Chief Technology Officer.
- **Key contributions:** lead author of " [An Empirical Model of Large-Batch Training](https://arxiv.org/abs/1812.06162?ref=nextomoro.com)" (December 2018); second-named author of " [Scaling Laws for Neural Language Models](https://arxiv.org/abs/2001.08361?ref=nextomoro.com)" (January 2020); senior co-author of " [Language Models are Few-Shot Learners](https://arxiv.org/abs/2005.14165?ref=nextomoro.com)" (GPT-3, 2020) and " [Constitutional AI: Harmlessness from AI Feedback](https://arxiv.org/abs/2212.08073?ref=nextomoro.com)" (December 2022); pre-AI publication record in quantum gravity and holography.
- **Awards:** [NSF Graduate Research Fellowship](https://www.nsf.gov/funding/opportunities/grfp-nsf-graduate-research-fellowship-program?ref=nextomoro.com) (2012).
- **X:** [@samsamoa](https://x.com/samsamoa?ref=nextomoro.com); **GitHub:** [@samsamoa](https://github.com/samsamoa?ref=nextomoro.com); **LinkedIn:** [Sam McCandlish](https://www.linkedin.com/in/sam-mccandlish-5878301/?ref=nextomoro.com).

## Origins

Public biographical material on McCandlish is comparatively thin compared with

... (truncated, 23 KB total)