Tom Brown

Tom Brown holds a deep, highly specialized background in software engineering, machine learning architecture, and large-scale distributed computing. His critical formative experience was gained at OpenAI, where he wrestled with the immense engineering challenges of training incredibly massive neural networks across thousands of GPUs. This hands-on, practical experience in overcoming the technical bottlenecks of massive-scale compute provided the essential engineering foundation for Anthropic's rapid ascent.

Tom Brown is a co-founder of Anthropic and one of the most consequential software engineers in the history of artificial intelligence. Before establishing Anthropic, Brown was a senior engineer at OpenAI, where he served as the lead author of the historic GPT-3 paper ('Language Models are Few-Shot Learners'). His work on GPT-3 proved that massive, generalized language models could perform an incredibly wide variety of tasks without specific fine-tuning, fundamentally changing the direction of AI research globally. Driven by a desire to focus more heavily on AI safety and alignment as these models rapidly approached human-level capabilities, Brown left OpenAI in 2021 to co-found Anthropic. His unparalleled experience engineering the world's first massive-scale language models was instrumental in building the initial architecture for Anthropic's Claude.