OpenAI CEO Sam Altman admits AI’s benefits might not be widely distributed.
In a In a new essay posted on his personal blog,OpenAI CEO Sam Altman stated that the company was open to “compute budgets” among other “strange” ideas to “enable everybody on Earth to use AI” and to ensure the technology’s benefits are widely distributed.
The historical impact of technological advancement suggests that most metrics we care about, such as health outcomes, economic prosperity, and so on, are likely to improve over time. Altman wrote that while equality is improving on average, it does not seem to be a technologically-driven phenomenon. “This may require new ideas,” Altman wrote. It seems that the balance of power between labor and capital could easily be thrown off, and this might require early intervention. Already, AI is impacting on the labor marketand resulting in Job cuts departmental downsizing. Experts have warned that AI technology could lead to mass unemployment if it is not accompanied by government policies and programs for reskilling or upskilling. Altman has claimed that artificial general intelligence (AGI), which he defines in his book as “[an AI] a system that can solve increasingly complex problems at the human level, across many fields,” is close. Altman warns that this AGI may not be perfect in the sense that “it will require lots of human supervision and guidance”
[AGI systems] “will not have the greatest new ideas,” Altman wrote. “And it will be great at certain things but surprisingly poor at others,” Altman asserted. Altman, like OpenAI’s rival, Anthropic CEO Dario Amodei envisions thousands, or even millions, of hyper-capable AI system tackling tasks in “every field of knowledge work.” Altman noted that “you can spend any amount of money to get predictable and continuous gains” in AI performance. OpenAI has reportedly been in talks with investors to raise up $40 billion and has committed to spending up to $500 billion on a massive data network. Altman also argues that the cost of using “a certain level of AI” drops by about 10x each year. Altman says that while pushing the boundaries of AI technology will not get cheaper, users will have access to more capable systems as they progress.
The Chinese AI startup DeepSeek, and others, have developed inexpensive AI models that are capable. There are signs that training and development cost is also decreasing, but Altman and Amodei both argue that Massive investments are needed to achieve AGI-level AI and beyond.
Altman stated that OpenAI will make “major decisions and limitations in relation to AGI safety which will not be popular” if it is able to create AGI-level systems. It was once pledged that it would stop competing with any “value-aligned,” “safety-conscious” project that came close to building AGI, before it did, out of concern about safety.
OpenAI had originally intended to remain as a nonprofit. The company is converting its corporate structures into a more traditional profit-driven organization. OpenAI Apparently is aiming to reach $100 billion by 2029. This would be equal to Target’s and Nestle’s current annual sales. Altman said that OpenAI will “trend more toward individual empowerment” as it builds more powerful AI. This will prevent “AI being used to control their populations through mass surveillance and loss autonomy.” OpenAI has in the past open-sourced technology, but the company has a preference for proprietary, closed-source software development. Altman stated that “AI will permeate all aspects of society and the economy; we will expect to see everything smart.”
Altman said that many of us expect that we will need to give people greater control over technology than in the past, including more open-sourcing. We also accept that there are trade-offs between safety and individual empowerment. Visions of AI’s future.
Altman stated in a footnote that OpenAI did not intend to end its partnership with Microsoft, a close partner and investor, anytime soon. Microsoft and OpenAI had reportedly agreed on a contractual definition for AGI – AI systems that generate $100 billion in profit – that, if met, would allow OpenAI the opportunity to negotiate more favorable terms of investment. Altman stated that OpenAI, “fully expect[s] [tobepartneredwithMicrosoftonthelong-term”
Kyle Wiggers, a senior reporter with TechCrunch, has a special interest artificial intelligence. His writings have appeared in VentureBeat, Digital Trends and a variety of gadget blogs, including Android Police and Android Authority, Droid-Life and XDA-Developers. He lives in Brooklyn, with his partner who is a piano teacher, and plays the piano occasionally. Sometimes — but mostly unsuccessfully.
View Bio
