San Francisco (United States) (AFP) – Founded five years ago by a breakaway group from OpenAI, Anthropic is now preparing a stock market listing that will be the biggest test yet of AI’s long-term financial viability. Here are five things to know about the company behind the Claude chatbot.
– OpenAI exiles –
Anthropic’s leaders learned to build AI at OpenAI, the ChatGPT maker, where they became known for their work on safety. Dario Amodei joined the then-nonprofit in 2016, where he ran its safety team and a team studying “scaling laws.” Scaling laws describe how AI models gain ability as they are trained on more data. His younger sister Daniela joined OpenAI in 2018 and became vice president of safety and policy. After disagreements with OpenAI leadership over how AI should be built, the Amodeis and several colleagues left in December 2020 and launched Anthropic four weeks later. The bitter rivalry between the two San Francisco-based companies has continued since.
– Siblings and housemates –
Chief executive Dario Amodei and Daniela, the company’s president, grew up in San Francisco’s Mission District. Their father was a leather craftsman and their mother managed public library renovations. According to the New York Times, two other Anthropic cofounders were the Amodeis’ housemates in San Francisco before the OpenAI years: physicist Jared Kaplan, now chief science officer, and researcher Chris Olah, whom Dario Amodei had once tutored in math. The remaining three cofounders also came from OpenAI. Sam McCandlish is Anthropic’s chief architect and was previously chief technology officer. Tom Brown oversees the computing infrastructure. He was among OpenAI’s first 20 employees. Jack Clark heads the company’s public benefit work and founded the Anthropic Institute, an internal think tank. He is a former technology journalist who joined OpenAI in 2016.
– Rationalists and effective altruists –
Dario Amodei’s thinking was shaped by two related movements. The first is rationalism, associated with AI thinker Eliezer Yudkowsky, which holds that AI could become powerful enough to destroy society. The second is effective altruism, which aims to make charitable giving as efficient as possible. “Effective altruists” also supplied Anthropic’s early money. Facebook cofounder Dustin Moskovitz and Skype cofounder Jaan Tallinn backed the first funding round. Sam Bankman-Fried invested $500 million through his company FTX in non-voting shares. After FTX collapsed and Bankman-Fried was sent to jail, that stake was sold for $884 million to repay the exchange’s customers. Amodei said in 2023 that he did not consider himself an effective altruist, though he was sympathetic to some of its ideas. Critics including Meta’s Mark Zuckerberg and Nvidia’s Jensen Huang have called him a “doomer.”
– Big Tech and big funds –
Anthropic’s IPO filing shows that two of its largest backers are Amazon and Google’s parent company Alphabet. Other participants include Founders Fund, the venture firm started by Peter Thiel, along with Microsoft and Nvidia. In May, Anthropic raised $65 billion at a $965 billion valuation, which put it ahead of OpenAI for the first time. Former Google chief executive Eric Schmidt is among the earliest outside investors. He put tens of millions of dollars into Anthropic’s first funding round in 2021.
– Clash with the Pentagon –
In January, the Pentagon, which the Trump administration asked to have unrestricted use of Claude, the company’s chatbot. Anthropic refused to allow two uses: domestic mass surveillance and lethal autonomous weapons. In February, a furious Trump ordered federal agencies to stop using Anthropic’s technology. Defense Secretary Pete Hegseth then directed that the company be labeled a supply-chain risk, a designation normally used for foreign adversaries. In September, the Defense Department posted on social media: “Americanism, not effective altruism.”
© 2024 AFP
