Sam Altman predicts AI will surpass human intelligence by 2030 (www.businessinsider.com)

OpenAI CEO Sam Altman told WELT he expects AI to reach and likely surpass human-level general intelligence by 2030, arguing the capability curve remains “extremely steep.” He said GPT-5 already outperforms him in many ways but still lacks human flexibility; nevertheless, he expects models in the next two years (surprising by 2026) to enable AI-driven scientific discoveries humans can’t make alone. Altman estimated AI could take on 30–40% of everyday economic tasks, shifting job content more than wholesale job counts, and reiterated OpenAI’s commitment to alignment—even while acknowledging past mistakes and defending the organization’s mixed nonprofit/commercial structure. Technically and product-wise, Altman previewed hardware ambitions: a “family of devices” (an Apple designer was hired) that rethinks the interface model—moving from app-switching to long-running, autonomous orchestration of complex, time-extended tasks that understand user intent and return when human input is needed. The remarks underscore two implications for the AI/ML community: an urgent need to accelerate alignment, safety and governance research as capability leaps continue, and fresh engineering challenges around embedding powerful models into new, persistent device paradigms and U.S.-based infrastructure.