At the WIRED World Expo in Miami on November 5, Hugging Face CEO Clément Delangue poured cold water on the famous "doomsday equation." His assessment of the probability of human extinction, or P(Doom), as he calls it, was "close to zero." At a time when a crowd of founders is busy playing up existential threats, the head of the open-source camp chose to speak plainly: rather than apocalyptic imaginings, what deserves closer attention right now is what is happening elsewhere.
The real risk is behind locked doors Delangue pointed the finger at another danger: a handful of large companies secretly honing frontier models in closed environments. In his view, the main danger of AI does not come from the open ecosystem, but precisely from these "behind-closed-doors" participants. Once training and applications are enclosed behind the high walls of a few laboratories, the outside world can neither audit nor check them, and the risk is instead quietly magnified.
Dismantling monopoly with openness His solution is to open the doors: let more developers openly build, review, and improve models, creating an ecosystem of mutual checks and balances. In his view, open-weight models are not just a technical preference, but a tool to prevent a few companies from monopolizing talent, compute, and AI capabilities all at once.
The platform itself is still growing The company offered evidence of continued expansion: it now has more than 20 million AI developers, up from about 18 million disclosed when Nvidia announced its $12.93 billion acquisition in September; Hugging Face has also sold 50,000 Microducks, bipedal robots that can be programmed with the help of AI.
