Artificial intelligence (AI) is advancing quickly and we are close to creating superintelligence, says Sam Altman in a new blog post.
Altman thinks the hardest advances in AI technology, like those behind systems such as GPT-4, are done. These breakthroughs will carry us far. AI’s biggest contributions will be speeding up scientific progress and increasing productivity. This could make the future much better than today. For example, he says, ChatGPT is already more powerful than any human in some ways. Millions use it daily for important tasks. A small improvement in AI can create huge benefits, but a small mistake, used by millions, could cause harm.
The near future and beyond
In 2025, AI agents (programs that perform tasks on their own) started doing real thinking work, like writing computer code. By 2026, Altman says, AI might discover new ideas. By 2027, robots could handle physical tasks. More people will create software and art with AI’s help, though experts using these tools will likely stay ahead of beginners. By 2030, one person’s ability to achieve more than in 2020 will be remarkable. Everyday life may not change completely, but the world will feel very different. Intelligence and energy, the main limits on progress, will become abundant, enabling almost anything with good planning.
AI is already boosting scientists’ productivity, and advanced AI can research better AI. This could lead to new discoveries, like better algorithms. Robots building more robots or data centers could speed things up further. As costs drop, AI’s price might match electricity’s. Progress will accelerate, but people will adapt. Jobs may disappear, but new wealth could support bold solutions. Challenges like AI safety and fair access must be addressed to ensure everyone benefits.
OpenAI is "a superintelligence research company," says Altman. "We have a lot of work in front of us, but most of the path in front of us is now lit, and the dark areas are receding fast."