Get your free personalized podcast brief

We scan new podcasts and send you the top 5 insights daily.

Tom Milton estimates labs have single-digit verification staff and maybe 50 people globally work on compute verification. He contrasts this with Project Vela, where the US spent billions in today's money on seismology and satellites to detect nuclear tests. A talented researcher could quickly become one of the field's most useful people.

Related Insights

OpenAI's team found that as code generation speed approaches real-time, the new constraint is the human capacity to verify correctness. The challenge shifts from creating code to reviewing and testing the massive output to ensure it's bug-free and meets requirements.

The intense talent war in AI is hyper-concentrated. All major labs are competing for the same cohort of roughly 150-200 globally-known, elite researchers who are seen as capable of making fundamental breakthroughs, creating an extremely competitive and visible talent market.

The core challenge in the AI race isn't monetization but model creation. The global pool of researchers capable of building frontier AI models is incredibly small—estimated at 100-150 people. This talent scarcity makes creating a leading model a much greater bottleneck than for a company like OpenAI to scale a known advertising business model.

AI can produce scientific claims and codebases thousands of times faster than humans. However, the meticulous work of validating these outputs remains a human task. This growing gap between generation and verification could create a backlog of unproven ideas, slowing true scientific advancement.

Tom Milton says he is not convinced verification is strictly necessary. Nuclear deals relied partly on trust, satellites counting bombers, and spying; AI deals could also use auditors, whistleblowers, or unilateral slowdowns. But as more lab leaders and governments call for coordination, verification becomes valuable optionality.

AI can generate vast amounts of content, but its value is limited by our ability to verify its accuracy. This is fast for visual outputs (images, UI) where our eyes instantly spot flaws, but slow and difficult for abstract domains like back-end code, math, or financial data, which require deep expertise to validate.

AI excels at generating code, making that task a commodity. The new high-value work for engineers is "verification”—ensuring the AI's output is not just bug-free, but also valuable to customers, aligned with business goals, and strategically sound.

While interest in AI safety has grown, it's dwarfed by the explosion in AI capabilities research. There are only about 1,000 people in technical AI safety versus up to a million working to accelerate AI capabilities, creating a massive talent imbalance on a critical issue.

Paying a single AI researcher millions is rational when they're running experiments on compute clusters worth tens of billions. A researcher with the right intuition can prevent wasting billions on failed training runs, making their high salary a rounding error compared to the capital they leverage.

With AI generating complex formulas and proofs, the most challenging part of scientific research is no longer solving the core problem. Instead, the primary human task becomes verifying the AI-generated results and writing them up, fundamentally changing the research workflow.