“AI researcher Yann LeCun is considered an "AI bear" because he argues that current methods like autoregressive pre-training on internet data will not lead to AGI.”