Chamath on tech’s AI divide

Chamath Palihapitiya said on X that artificial intelligence is splitting tech into a small class of “AI-superpowered” workers and everyone else, with more wealth flowing to the top. (x.com) In a second post, he said people who benefit most from that shift should act as stewards as gains concentrate. His comments circulated alongside a wider Silicon Valley argument that smaller teams can now produce more with generative artificial intelligence tools. (x.com) (bloomberg.com) That argument has moved from theory to operating metric. Bloomberg reported on June 20, 2025 that startup bragging rights have shifted from valuation and headcount to revenue per employee, with investors and founders talking about “tiny teams” and “scaling without growing.” (bloomberg.com) The mechanics are straightforward: software that writes code, drafts marketing copy, answers support questions, and helps with research lets one employee cover work that once needed several specialists. Deloitte said in its 2026 enterprise report that worker access to artificial intelligence rose 50% in 2025 and 66% of organizations reported productivity gains. (deloitte.com) The split Palihapitiya described is showing up in pay and output data as well. PwC said on June 3, 2025 that workers with artificial intelligence skills drew a 56% wage premium in 2024, while industries most exposed to artificial intelligence posted 27% growth in revenue per employee versus 9% for the least exposed industries. (pwc.com) Startup investors are citing concrete examples from coding. Y Combinator leaders said in March 2025 that about a quarter of the Winter 2025 batch had codebases that were 95% generated by artificial intelligence, a change they said reduced the need for teams of 50 or 100 engineers. (techcrunch.com) (cnbc.com) That helps explain why the debate is no longer just about whether artificial intelligence works. It is also about who captures the gains when one founder, one engineer, or one sales team can produce much more than before. (bloomberg.com) (pwc.com) Not every data point points to immediate job destruction. PwC said jobs still grew in roles more exposed to artificial intelligence, and Deloitte said most companies were adjusting through training rather than large-scale role redesign, with skills gaps still the biggest barrier to adoption. (pwc.com) (deloitte.com) But the direction inside elite tech firms is toward heavier reliance on machine output. Anthropic chief executive Dario Amodei said in 2025 that artificial intelligence could soon write most code, and later said many teams at Anthropic were already near that level, even as he argued engineers were still needed to direct the systems. (businessinsider.com) (msn.com) Palihapitiya’s warning landed because it put a moral frame on a business trend that investors, founders, and executives are already measuring in headcount, wages, and revenue per employee. (x.com) (bloomberg.com)