Overview
- LeCun advised focusing on long‑shelf‑life subjects such as mathematics, physics, electrical engineering, signal processing, control theory and optimisation.
- He cautioned that taking only minimal math in typical CS programs can leave graduates unable to adapt to major technological changes.
- He clarified he is not telling students to avoid computer science as a major, urging them instead to load up on foundational courses rather than short‑lived tech electives.
- He maintained that scaling language models trained only on text will not deliver human‑level intelligence, arguing that such systems need far more than text to model reality.
- He added that core programming proficiency remains essential even as AI tools make coding more efficient.