DP Dwarkesh Patel Dwarkesh Patel Hosts the Dwarkesh Podcast, a long-form interview show on AI progress, timelines, and the people building frontier models, along with economists and historians.

“The effective AI population size at the frontier labs is increasing 10x year over year. Even if compute scaling slows down, it doesn't take many more years before each company individually has more labor equivalents than there are people on Earth. We talk a lot about centralization of power from nationalization, but we don't think enough about the fact that most of the world's minds, in terms of work output, are about to be concentrated within two labs. If these AIs are misaligned, then most of the world is misaligned, basically, because most of the world's minds are there.”

Dwarkesh Podcast · AI Research & Frontier Labs · August 2026

“The effective AI population size at the frontier labs is increasing 10x year over year. Even if compute scaling slows down, it doesn't take many more years before each company individually has more labor equivalents than there are people on Earth. We talk a lot about centralization of power from nationalization, but we don't think enough about the fact that most of the world's minds, in terms of work output, are about to be concentrated within two labs. If these AIs are misaligned, then most of the world is misaligned, basically, because most of the world's minds are there.” — Dwarkesh Patel, Dwarkesh Podcast

His own extrapolation of the compute trends he and Patel had been costing out, redirected from economics to concentration of power. The move is to treat compute as population rather than as capex, which turns a spending forecast into a claim about where the world's cognitive labour will sit.

Dwarkesh Podcast · 2026-08-25 Listen to the episode → More from Dwarkesh Patel → More AI Research & Frontier Labs quotes →

More from Dwarkesh Podcast

“I wrote a blog post a few months ago about how RL, at least policy gradient RL, is even more inefficient than you might think. The inefficiency one thinks about naively is that you have to roll out a whole trajectory to get any learning signal at all — instead of completing the next word in a sentence, the model has to do two days' worth of work to figure out if it even did the project correctly. If you're doing this through RL, the model would try 'the sky is halcyon' — nope, wrong. 'The sky is told' — nope, wrong. This is a totally untrained model, so you'd have to do this on the order of 100,000 times just to stumble on 'blue' before you get any learning signal at all.” — Dwarkesh Patel, Dwarkesh Podcast “When I learn how an LLM works and how simple RLVR is as an algorithm, I'm sort of stunned by the kinds of things it can do — that it can learn to build very complicated code repositories simply from getting a yes or no. And here, I feel like AlphaGo actually seems less impressive in retrospect the more you understand it — because you're putting in a lot of bias, you're telling it how to titrate exploration as things go on, you're building this very explicit tree search for it.” — Dwarkesh Patel, Dwarkesh Podcast