Dario Amodei
Scaling Laws and Anthropic
Concrete Problems in AI Safety, and the scaling hypothesis
Rating
91Impact score
Domains
ScalingAlignmentFrontier models
Scouting report
Co-authored Concrete Problems in AI Safety, which gave safety research a concrete agenda, and led the GPT-2 and GPT-3 work where the scaling laws were established. Left OpenAI to co-found Anthropic, whose interpretability and Constitutional AI research is published openly even though the models are not.
Curator’s note
The set’s open-weights sceptic, and its most rigorous public writer on scale.
Portrait
Painted from a photograph of Dario Amodei by UK Prime Minister, licensed CC BY 2.0.
Sources
- 01Dario Amodei en.wikipedia.org
- 02Concrete Problems in AI Safety (2016) arxiv.org
- 03Scaling Laws for Neural Language Models (2020) arxiv.org