Overview
- Altman said he would be very surprised if by 2030 there were no models capable of doing things humans cannot, a threshold he associates with superintelligence.
- He forecast that AI could in the near future handle roughly 30–40% of tasks across the economy, emphasizing task-level automation over wholesale job elimination.
- He asserted that GPT-5 already outperforms many people in some respects and suggested AI-only scientific discoveries could be plausible within a few years.
- He cautioned that advances may trigger mass layoffs, citing customer support roles as especially vulnerable and describing uncertainty around how programming work will evolve.
- He called for stronger alignment to human values, rejected claims that future systems would treat people like ants, and said OpenAI is exploring a small family of consumer devices with a former Apple designer.