Dario Amodei is the co-founder and CEO of Anthropic, the AI safety and research company behind Claude. Before Anthropic, he was Vice President of Research at OpenAI and a senior research scientist at Google Brain. His work connects large language models, training with human feedback, and AI safety.
Dario Amodei at a glance
- Role: co-founder and CEO of Anthropic
- Education: PhD in biophysics from Princeton University as a Hertz Fellow, then postdoctoral work at Stanford University School of Medicine
- Earlier roles: Google Brain, then VP of Research at OpenAI
- Known for: GPT-2 and GPT-3 research, learning from human preferences, Constitutional AI, and the essay “Machines of Loving Grace”
Source: Dario Amodei’s public biography.
How did Dario Amodei get into AI?
Amodei started in science, not software. His training in biophysics shaped how he sees AI: as a tool that could speed up research in biology and medicine. He moved into deep learning at Google Brain, then joined OpenAI, where he helped build GPT-2 and GPT-3 and co-led research direction.
| Stage | What happened |
|---|---|
| Science | PhD in biophysics at Princeton, postdoc at Stanford |
| Google Brain | Senior research scientist in deep learning |
| OpenAI | VP of Research, GPT-2 and GPT-3 |
| 2021 | Co-founds Anthropic and becomes CEO |
| September 2023 | Anthropic publishes its Responsible Scaling Policy |
What are Dario Amodei’s main contributions to AI?
- Learning from human feedback: he contributed to OpenAI’s 2017 work on training AI from human comparisons, a foundation of how modern assistants are refined.
- Scaling language models: GPT-2 and GPT-3 showed that more data and compute could unlock broad language skills.
- Constitutional AI: he is a co-author of the paper on training assistants with a written set of principles and AI feedback.
- Interpretability: under his leadership, Anthropic made understanding a model’s inner workings a core research area.
- Responsible scaling: Anthropic’s policy ties safety measures to the capability level of each model.
What does Dario Amodei believe about AI?
He holds two ideas at once. In “Machines of Loving Grace,” he describes how powerful AI could improve health, mental health, economic development, governance, and work. In his 2023 Senate testimony, he described short-term risks like bias and misinformation, medium-term risks like misuse in biology and cybersecurity, and long-term risks from systems that become hard to control.
That balance between optimism and caution explains much of Anthropic’s identity: build useful AI, and study its risks seriously while doing it.
Frequently asked questions about Dario Amodei
What company does Dario Amodei lead?
He is the co-founder and CEO of Anthropic, the company that makes Claude.
What did Dario Amodei do at OpenAI?
He was Vice President of Research and helped lead the work behind GPT-2 and GPT-3.
What is “Machines of Loving Grace”?
It is his long essay on how powerful AI could benefit humanity, especially in biology, health, and economic development.
The takeaway
Dario Amodei’s journey runs from biophysics to some of the most important AI research of the last decade. His mark is visible in how assistants are trained, in Constitutional AI and interpretability research, and in the push to treat AI safety as a science. To understand him, start with his essay, Anthropic’s research, and his own words.
