Dispatch
Hinton concedes the open-weights fight at Ai4: 'I think that battle's been lost'
At the Ai4 conference in Las Vegas, Geoffrey Hinton, Fei-Fei Li and Andrew Ng split three ways on open models. Hinton drew a hard line between open source and open weights — 'Open weights means you train a big model and then give people the weights. That's very different' — and conceded 'I think that battle's been lost.' Ng argued against gatekeepers and for openness on competitive grounds, while Li rejected the binary outright: 'It's very dangerous to make this a dichotomy between complete openness... It's much more nuanced,' proposing layered openness modeled on nuclear physics regulation. All three backed some regulatory framework.
Source
↳ Follow the thread