Dead Economy Theory — HN today: 676 points, 859 comments
TL;DR
Two Wharton economists formalize the «AI Layoff Trap»: each firm rationally cuts staff, but collectively this destroys consumer demand. Block laid off nearly half its workforce — stock jumped 25%.
Two Wharton economists formalize a problem game-theoretically in «AI Layoff Trap»: in a market of 20 competitors, savings from layoffs accrue to you, but demand destruction is shared across the industry. Each firm individually rational; collective outcome is consumer-market contraction.
Block laid off nearly half its workforce. Stock +25% after-hours. Nobel laureate Acemoglu's numbers: only 4.6% of tasks are worth automating with AI now; AI's ten-year productivity contribution estimate is 0.66%. In 2025, over 90% of enterprises reported «no quantifiable productivity impact» from AI investments. But cheap «good enough» AI keeps cutting jobs because quarterly incentives demand it.
The most painful one for developers: Anthropic's own research found that junior engineers relying on AI coding agents complete tasks not much faster, and post-task comprehension scores are lower. The tool simultaneously takes the job and erodes the process for building the ability to do the job.
Dario Amodei has publicly said: «Democracy's power balance is predicated on ordinary people gaining leverage by creating economic value. If that premise disappears, things become quite frightening.» Anthropic hasn't endorsed any related legislation. OpenAI published a white paper calling for higher capital gains tax — while its Super PAC spent $2M attacking congressional candidates who proposed AI taxes.
via HN
Block laid off nearly half its workforce. Stock +25% after-hours. Nobel laureate Acemoglu's numbers: only 4.6% of tasks are worth automating with AI now; AI's ten-year productivity contribution estimate is 0.66%. In 2025, over 90% of enterprises reported «no quantifiable productivity impact» from AI investments. But cheap «good enough» AI keeps cutting jobs because quarterly incentives demand it.
The most painful one for developers: Anthropic's own research found that junior engineers relying on AI coding agents complete tasks not much faster, and post-task comprehension scores are lower. The tool simultaneously takes the job and erodes the process for building the ability to do the job.
Dario Amodei has publicly said: «Democracy's power balance is predicated on ordinary people gaining leverage by creating economic value. If that premise disappears, things become quite frightening.» Anthropic hasn't endorsed any related legislation. OpenAI published a white paper calling for higher capital gains tax — while its Super PAC spent $2M attacking congressional candidates who proposed AI taxes.
via HN
