Summary
Nick Saraev, who says he has spent seven years building and growing AI agents for businesses, lays out which careers he expects to keep paying well by 2028. The case starts with the METR time horizon test, which measures how long AI agents can work autonomously at human-level quality about half the time. METR found in 2025 that this horizon was doubling roughly every seven months, and he argues the trend points to agents doing months of human work in a day or so within two years.
From there he builds the argument on a few premises: pay flows to whatever is the bottleneck in an economy, and the most valuable people will act as centaurs, combining human judgment with AI doing the heavy lifting. He cites Anthropic’s finding that AI agents now spend several times more hours on tasks than all of its human staff combined.
The video walks through six career paths, with a look at what a typical day involves and how to get started. Examples include electrician work tied to data center buildout, with faster entry points like low-voltage fiber cabling and solar installation, and AI security and evals roles such as red teaming, where he notes a posted Anthropic red team engineer role at around $400,000 per year.
📺 Source: Nick Saraev · Published November 14, 2026
🏷️ Format: Opinion Editorial







