Dario Amodei

Career

1. 
   Anthropic

   • CEO and Co-FounderFeb 2021 – present
2. 
   OpenAI

   • Vice President of ResearchDec 2019 – Dec 2020
   • Research DirectorSep 2018 – Dec 2019
   • Team Lead for AI SafetyJul 2016 – Sep 2018
3. 
   Google

   • Senior Research ScientistOct 2015 – Jul 2016
4. 
   Baidu

   • Research ScientistNov 2014 – Oct 2015
5. S
   Skyline Targeted Proteomics Environment

   • Software Developer (part-time)May 2013 – Dec 2014
6. 
   Stanford University

   • Postdoctoral ScholarDec 2011 – Dec 2014

Education

• PhD, (Bio)physics, Princeton University, 2011
• BS, Physics, Stanford University, 2006
• Progress towards BS, Physics, Caltech, 2001–2003

Lesser-known facts

• Did a PhD in biophysics at Princeton (2011) after a physics degree at Stanford.
• Worked as a geophysicist at Schlumberger in 2004, years before AI.
• Credited as a co-inventor of reinforcement learning from human feedback.
• Made the USA Physics Olympiad team in 2000.
• Started college at Caltech as one of Tom Tombrello's Physics 11 students, then transferred to Stanford.
• Worked part-time as a software developer on Skyline, a targeted proteomics tool, and consulted for Applied Minds.
• Worked at Baidu from November 2014 to October 2015, then at Google Brain, before joining OpenAI in 2016.
• Set out his scaling view in a 2017 document called "The Big Blob of Compute Hypothesis" and says he still holds it.
• Runs a biweekly all-hands called the DVQ, short for Dario Vision Quest, speaking for an hour to the whole company.

Research papers before Anthropic

• Concrete Problems in AI Safety2016 · author 1 of 6
• Deep Reinforcement Learning from Human Preferences2017 · author 6 of 6
• Fine-Tuning Language Models from Human Preferences2019 · author 6 of 8
• Language Models are Unsupervised Multitask Learners2019 · author 5 of 6
• GPT-3: Language Models are Few-Shot Learners2020 · author 31 of 31
• Learning to Summarize from Human Feedback2020 · author 8 of 9
• Scaling Laws for Autoregressive Generative Modeling2020 · author 18 of 19
• Scaling Laws for Neural Language Models2020 · author 10 of 10

Known for

• GPT-2
• GPT-3
• Scaling laws for neural language models (2020)
• Deep Speech 2
• Constitutional AI
• Responsible scaling policy

Honors

Hertz Doctoral Thesis Prize, 2012 · Hertz Fellow · McKnight Technological Innovations in Neuroscience Award · Star Employee of the Quarter, Q2 2015

Interviews and talks

• Dwarkesh Podcast: "We are near the end of the exponential" (2026)dwarkesh.com