Dario Amodei

Overview

Dario Amodei is the co-founder and chief executive of Anthropic, the company behind the Claude models.^[1][4] Before founding Anthropic he worked at OpenAI, where he became Vice President of Research.^[4] Anthropic was founded in 2021, after his OpenAI years.

Education and early career

Amodei attended Caltech and then earned a bachelor’s degree in physics from Stanford. He received a PhD in biophysics from Princeton, where he held a Hertz Foundation Fellowship in 2007 and won the Hertz Thesis Prize in 2011. He was also a postdoctoral scholar at Stanford University School of Medicine.^[1]

Baidu is a Chinese internet company, and Google Brain is Google’s AI research team.^[1]

Key contributions

In June 2017 Amodei co-authored “Deep reinforcement learning from human preferences” with Paul Christiano, Jan Leike, Tom B. Brown, Miljan Martic and Shane Legg.^[2] The method trains an AI system using human judgments. People compare pairs of behaviours, a model learns to predict those preferences, and reinforcement learning then adjusts the system against that learned signal. This general approach later became one of the main ways to steer language models toward helpful behaviour. The idea is that people can often express preferences by comparing two outputs more easily than by writing a precise score, so the system learns from which output people prefer.

At OpenAI he worked on GPT-2, on which he is a co-author of the paper, on “Scaling Laws for Neural Language Models,” and on GPT-3.^[1] The GPT-3 paper states that “Dario Amodei designed and led the research.”^[3] GPT-3 is a large language model, and its paper examined what such a model can do when it is given only a few examples of a task in its input.^[3]

In October 2024 he published the essay “Machines of Loving Grace: How AI Could Transform the World for the Better.” It describes a possible “country of geniuses in a datacenter” and a “compressed 21st century” of scientific progress.^[5]

Positions

Amodei left OpenAI at the end of 2020. OpenAI’s December 2020 organisational update thanked him for his contributions over four and a half years.^[4]

In 2021 he co-founded Anthropic with Daniela Amodei and other former senior OpenAI members, and he has been its CEO since.^[1]

Impact and recognition

In 2025 Amodei was named to the TIME100 AI list and to TIME’s list of the 100 most influential people. He was also included in TIME’s Person of the Year coverage of the “Architects of AI.”^[6] The essay mentioned above is one of his main written statements about how AI could affect science and the wider economy.^[5] It is an essay rather than a technical paper, so its claims are best read as his stated views rather than as established results.