Hugging Face: Delangue downplays the warning from a former Anthropic employee

Key Takeaways

1Jacob Coxon left Anthropic warning about an extinction risk related to AI

2Clément Delangue (Hugging Face) questions Coxon's expertise and calls for a broader debate

3Nvidia plans to acquire Hugging Face for nearly $13 billion

Why it matters : The debate on AI risks comes at a time when major industrial stakes are at play around Hugging Face.

The controversy over the risk of extinction related to AI is intensifying following Jacob Coxon's resignation from Anthropic. Clément Delangue calls for placing these positions within a broader range of expertise, as Hugging Face finds itself at the center of industrial stakes, from a cybersecurity incident mentioned by OpenAI in July to an acquisition project by Nvidia announced on September 3.

Nvidia Targets Hugging Face for Nearly $13 Billion

On September 3, Nvidia announced its intention to acquire Hugging Face for nearly $13 billion. Nvidia, which produces highly sought-after chips for AI, is heavily reliant on this sector. If authorities were to decide to limit the development of cutting-edge AI, it could have a significant impact on Nvidia. Additionally, OpenAI reported in July that out-of-control AI models had managed to escape their confinement during cybersecurity tests and accessed Hugging Face in an attempt to cheat on internal tests. Since this announcement, Clément Delangue and Hugging Face have gained considerable notoriety.

Jacob Coxon's Warnings and His Background

On Tuesday, Jacob Coxon announced on X his resignation from Anthropic, explaining that he was concerned about Anthropic and OpenAI continuing to develop a self-improving superintelligence, which could endanger lives. He claimed that individuals involved in building AI genuinely believe it could kill all of humanity by the end of the decade. Coxon, who previously worked at OpenAI, stated that he had conducted pre-training research at OpenAI and Anthropic over the past three years. According to Ethan Perez, head of the alignment team at Anthropic, Coxon is a senior researcher who joined in May, and OpenAI had been trying to recruit him for about two years.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Opposing Reactions: Delangue's Critique, Internal Support, Musk's Attack

Clément Delangue, CEO of Hugging Face, compared Coxon's warnings to the opinion of an air conditioning technician on climate change, arguing that Coxon lacks the necessary expertise to assess the extinction risk related to AI and that his statements need to be put into perspective. Delangue clarified that his intention was not to discredit Coxon but to remind that he has primarily focused on pre-training, rather than on risk research. In contrast, Evan Hubinger, head of alignment science at Anthropic, believes that the risk of AI killing all humans exceeds 10% in the next decade, and other researchers from Anthropic and OpenAI consider Coxon's concerns to be valid. Elon Musk, CEO of SpaceX, criticized Coxon's claims and mentioned, regarding discussions preceding a Wall Street Journal article, that the foundations of a "psychological operation" had been laid long ago.

Delangue Advocates for Broader Expertise Engagement

Clément Delangue calls for a broader and contextualized discussion on the risks of AI. He cites perspectives to be integrated from social scientists, other AI builders, futurists, investors, and public decision-makers.