How Fast Is Frontier AI Labor Growing?
Dwarkesh Patel says the effective AI workforce at frontier labs is growing about 10 times a year, and one lab could hold more labor than the number of people on Earth by the end of the decade.
Compute at the frontier, in flop terms, is growing four or five times a year, while the compute needed for a given capability falls about three times a year. That is about a 10 times larger effective population each year. It matters less today because the systems cannot yet do full jobs. If the trend holds, OpenAI goes from about 10 million AI laborers this year to 100 million the next year and a billion the year after. Even if compute scaling slows, it does not take many more years before each company has more labor equivalents than there are people on Earth. That is plausible by the end of this decade, with most of the work output concentrated in the two labs that keep taking more of the world's compute.
Dylan Patel says the same 10 times yearly growth holds even without recursive self-improvement, once the models are competent remote workers, software engineers, or researchers. With recursive self-improvement the growth could be 100 times or 1,000 times a year, or a mix of smarter models and a larger population. He does not see a path where this stays spread out.
Dylan Patel – Anthropic & OpenAI will have most of the world’s compute by 2028 ·
From the transcript
One thing I find crazy about these scenarios is just how much of the world's future labor supply ends up in
very few companies. And also how fast that labor supply grows year over year. So if like, computer the frontier, you know, in flop terms is growing four or five X a year. And further, the computer required to achieve the likability is like decreasing 3x a year. So the computer at the frontier, basically the effective AI population size at the frontier labs is increasing 10x year over year.
And so that doesn't really matter that much right now because AIs are not good enough to do full jobs or be like, as autonomous as people in their capacity to do work or pull off schemes or whatever. But if the current trend continues, you have a world where OpenAI goes from having, say, 10 million basically AI laborers this year to 100 million the next year to a billion the year after that.