newsfilter
← Interviews
20VC with Harry Stebbings1h 6m

Yann LeCun: Meta’s New AI Model LLaMA; Why Elon is Wrong about AI; Open-source AI Models | E1014

Yann LeCun, Elon, David Marcus, Harry Stebbings

Interview, Fireside Chat

Yann LeCun traces the historical resurgence of deep learning from his 1980s postdoctoral work with Jeff Hinton and Yoshua Bengio to his prediction that current autoregressive large language models will soon become obsolete due to their lack of planning capabilities and persistent memory. He advocates for a shift toward open-source ecosystems that enforce hardwired safety constraints and hierarchical objectives, arguing that such architectures will replace text prediction with systems capable of genuine world modeling and common sense. LeCun concludes by dismissing doomer narratives and mass unemployment fears, asserting that the AI revolution will ultimately generate a new renaissance of creative roles while requiring societal adjustments to wealth distribution rather than technological restriction.