$100 Million for One Human: The Talent Inversion

Two labor markets ran in opposite directions this month, and the gap between them is the clearest picture yet of where this is heading.

Market one: Meta's new superintelligence lab reportedly dangled offers reaching nine figures — yes, $100 million packages — to poach individual researchers from OpenAI and rivals. A bidding war over perhaps a thousand people on Earth whose skills move the capability frontier. Google, meanwhile, effectively acqui-hired the leadership of a hot agentic-coding startup in a $2.4 billion arrangement.

Market two: everyone else. The entry-level pipeline kept tightening, internships thinned, and "we're being strategic about headcount" became the corporate phrase of the season — while ChatGPT gained a full agent mode this month that operates a virtual computer: browsing, building spreadsheets, completing multi-step tasks. The kind of work that used to be somebody's first job.

The inversion, plainly

When a technology can replicate median cognitive skill, the price of median cognitive skill falls — and the price of the rare skill that improves the technology explodes. This is the mechanism. Not malice, not conspiracy: arithmetic. A researcher who nudges the frontier is leverage on billions in compute; a capable generalist is, increasingly, in competition with that compute. Left alone, this arithmetic produces an hourglass economy — enormous rewards at the top, a squeezed middle, and service work the machines haven't reached at the bottom.

I refuse the fatalist reading. Every prior transition's arithmetic was also brutal until institutions caught up — weekends, labor law, public education were all inventions, not gifts of the market. The post-labor era will need inventions of the same scale: ownership structures that spread capital income, transition floors, new ladders into contribution.

Actionable Steps

Companies: your comp curve is barbelling — be honest about it internally, and create explicit paths for strong generalists to become agent-fleet operators, not casualties. Individuals: in an hourglass, position where machines multiply you — accountability roles, client trust, cross-domain judgment — or where they can't reach yet. Policymakers: capital income is where the winnings are pooling; if you want a broad middle class in 2035, broad capital ownership is the lever. Study every model from sovereign funds to employee ownership trusts. Now, in calm weather.

One market pays for humans like franchises. The other is learning to do without them. Bridging those two is the political project of the decade.