Dario Amodei is co-founder and chief executive officer at Anthropic, an artificial intelligence (AI) safety and research company working to build reliable, interpretable and steerable AI systems.Previously Dario was vice president of research at OpenAI, during which he set the overall research direction at the organization, led the efforts to build GPT-2 and GPT-3, and led several teams focused on long-term safety research, including how to make AI systems more interpretable and how to embed human preferences and values in future powerful AI systems. Before working at OpenAI, Dario was a senior research scientist at Google, serving as a deep learning researcher on the Google Brain team, working to extend the capabilities of neural networks.
Co-founder and CEO Dario Amodei leads Anthropic, a safety and research startup focused on artificial intelligence, which develops dependable, comprehensible, and manipulable AI systems.
Dario was previously vice president of research at OpenAI, where he oversaw the organization’s overall research strategy, oversaw the development of GPT-2 and GPT-3, and oversaw multiple teams dedicated to long-term safety research, including how to improve the interpretability of AI systems and incorporate human preferences and values into more potent AI systems in the future. Dario was a senior research scientist at Google before joining OpenAI. He worked on the Google Brain team as a deep learning researcher, trying to improve the capabilities of neural networks.
Dario finished his PhD in physics at Princeton University, where he developed unique intracellular and extracellular recording devices and worked on statistical mechanics models of neuronal networks. His work at Stanford University School of Medicine as a postdoctoral scholar focused on using mass spectrometry to hunt for cancer biomarkers and to build network models of the cellular proteome.
2023
Texas-based Anthrop LLC filed a lawsuit against Anthropic in February 2023 alleging that Anthropic used its registered trademark “Anthropic A.I.” The business was dubbed the “Center of A.I. Doomerism” by Kevin Roose of The New York Times. According to him, a few workers “compared themselves to modern-day Robert Oppenheimers”.
Journalists frequently associate Anthropic with the effective altruism movement since several of the team members and founders were involved in or attracted to the community. Sam Bankman-Fried, the founder of the cryptocurrency exchange FTX, was among the investors in the Series B round before it folded in 2022.
Amazon announced a collaboration on September 25, 2023, with an initial investment of $1.25 billion and the possibility of Amazon spending up to $4 billion to become a minority shareholder. Anthropic agreed to utilize Amazon Web Services (AWS) as its main cloud provider and to make its AI models available to AWS users as part of the agreement. Google put $500 million into Anthropic the next month and pledged to contribute an additional $1.5 billion over time.
Latest news about Dario Amodei
OpenAI’s board has proposed a merger with rival Anthropic, a key AI competitor, following the firing of former CEO Sam Altman. However, Anthropic’s co-founder and CEO, Dario Amodei, declined both the merger proposal and the CEO position due to his commitment to Anthropic, which is seen as FTX’s most valuable asset. The proposal highlights intense competition and strategic maneuvers within the AI industry, with both companies vying for technological advancements and market influence in a rapidly evolving field.
Introducing Claude 2! Our latest model has improved performance in coding, math and reasoning. It can produce longer responses, and is available in a new public-facing beta website at https://t.co/uLbS2JNczH in the US and UK. pic.twitter.com/jSkvbXnqLd
Neural networks often pack many unrelated concepts into a single neuron – a puzzling phenomenon known as 'polysemanticity' which makes interpretability much more challenging. In our latest work, we build toy models where the origins of polysemanticity can be fully understood.
In "Language Models (Mostly) Know What They Know", we show that language models can evaluate whether what they say is true, and predict ahead of time whether they'll be able to answer questions correctly. https://t.co/ZmZJqycb39pic.twitter.com/mmZDS930Df
