Timeline chosen (2028) to drive conversation on policy needed to avoid scenario; US white-collar job market already weak; AI agents will significantly replace jobs, diffusion through economy will take time but leading edge is concerning.

Reaction larger than expected but not surprising given AI trade has been straight line up for 3.5 years, everyone max long, few incremental buyers left; market digesting that AI is more powerful and starting to disperse benefits/losses.

Constantly churning book, have shorts in businesses to be disrupted by AI, own semiconductors that benefit; main story is risk to overall economy and consumer economy if jobs go faster than we think.

Intermediation sectors at real risk: financial services, insurance, banks (credit cards, lending, deposits). In world where AI agent does tasks for you, friction of changing providers drops precipitously.

Semiconductors, upstream to semiconductors, data center materials, foundation labs, businesses with true economic moats (brands, status goods) are huge winners.

Feedback loop broken by policy: tax incremental/windfall gains from AI complex (materials, semiconductors, foundation labs, tech) and companies cutting jobs without demand suffering; need asymmetric corporate tax to maintain equilibrium.

Biggest tell is white-collar employment in US, especially information workers; employment by information category already down almost 8% from peak; maps to big tech layoffs as they roll out AI.

China foundation lab companies only public exposure globally, so zooming; China analog for what happens in US next two years: automation already rolled through economy, weak consumer economy dragging everything down.