In his new essay on his personal blog, Openai CEO Sam Altman said that the company, among other “strange sound” ideas, ensures that “everyone on earth can use a lot of AI” He said he is accepting a “calculation budget” to ensure that the budget is in place. The advantages of technology are widely distributed.
“The historical impact of technological advancements suggests that most of the metrics we are interested in (health, economic prosperity, etc.) will improve on average and long term, but we also have an equality. The increase doesn't seem to be technically determined, and we can do this right. We need new ideas,” Altman wrote. “It seems like the balance of power between capital and labor in particular can easily be messed up, and this may require early intervention.”
Like Altman's “computational budget” concept, solutions to this problem may be easier to propose than implementing. Already, AI has impacted the labor market, resulting in job cuts and sector downsizing. Experts warn that mass unemployment is a possible consequence of the rise of AI technology without appropriate government policies and reskills and high-end programs.
Altman claims that although not the first time, artificial general information (AGI) defines it as “.[an AI] Many fields are nearby systems that can tackle increasingly complex problems at a human level. No matter what form, this AGI will not be perfect, Altman warns, in the sense that “it may require many human supervision and direction.”
“[AGI systems] Altman writes:
However, Altman argued that the real value from AGI comes from running these systems at scale. Like Dario Amodei, CEO of Openai's rival Anthropic, Altman envisions thousands or millions of AI systems working on tasks “in every area of knowledge.”
We might assume that it will be an expensive vision for it to come to fruition. In fact, Altman observed that in AI performance, “you can spend any amount and get continuous and predictable profits.” That's probably why Openai is reportedly in talks to raise up to $40 billion in the funding round, and is committed to spending up to $500 billion with partners on a vast data network.
However, Altman also claims that the cost of using “a certain level of AI” drops by about 10 times every 12 months. In other words, pushing the boundaries of AI technology doesn't make it cheaper, but users can access increasingly capable systems along the way.
Chinese AI startup DeepSeek's competent and inexpensive AI models seem to support that concept. Although there is evidence to suggest that training and development costs are also lower, both Altman and Amodei argue that large investments are required to achieve AGI-level AI.
Regarding Openai's plans to release an AGI-Level system (assuming it can actually be created), Altman said the company could make “some major decisions and limitations related to the safety of unpopular AGIs.” said it is expensive. Openai once committed to halting competition and began supporting “valuable” and “safety-sensitive” projects that approach building AGIs.
Of course, that was when Openai intended to remain a nonprofit organization. The company is in the process of transforming its corporate structure into a more traditional, profit-driven organizational structure. Openai reportedly aims to reach $100 billion in revenue by 2029.
This is true, and Altman said Openai's goal to build stronger AI is to “more trends towards individual empowerment,” and at the same time, “a large amount of AI used by authoritarian governments.” “It's about controlling the population through the loss of surveillance and autonomy.” Altman recently said he believes Openai is on the wrong side of history when it comes to open sourcing technology. Although Openai has had open source technology in the past, the company has generally supported its own closed development approach.
“AI will penetrate every part of the economy and society. I hope everything will be smart,” Altman said. “Many of us expect that we will historically need to control more technologies, including open sourcing, and there is a balance of individual empowerment that requires trade-offs. I accept that.”
Altman's blog post comes ahead of this week's AI Action Summit in Paris. This has already prompted other technical authors to outline their own vision for the future of AI.
In a footnote, Altman added that Openai is not actually going to end its close partners and investors' relationships with Microsoft any time soon, using the term AGI. Microsoft and Openai reportedly had a contractual definition of an AI system that could generate $100 billion in profits. But Altman said Openai “full hopes”[s] To partner with Microsoft over the long term. ”
