Summary
Three months after publication, the thesis has gone from a niche thought experiment to a widely debated macro scenario. Markets, economists and policymakers now argue about the mechanisms it named: AI-driven labour displacement, Ghost GDP, SaaS margin compression, private-credit exposure. The strongest evidence supports the microeconomic parts. The weakest evidence is the macroeconomic catastrophe. As of June, the live data are consistent with structural disruption and repricing, not with an unfolding crisis.
New evidence in June 2026
2026-06-26Anthropic Economic Index measures deep task coverage inside real jobs; over a third of workers expect AI to handle most tasks within a year
86% report speed gains, 69% quality gains. 57% say AI makes their skills more valuable, and that rises with how automated their usage is. 10% rate losing their job within a year as likely, slightly below baseline separation rates.
One report, two verdicts. Adoption is deep. The people closest to it do not expect to be replaced.
Touches 02 (supports), 03 (contradicts) · Anthropic — Economic Index report: Cadences
2026-06-15AI displacement has become a mainstream macro topic; the memo itself moved markets
The discussion moved from 'can AI do the work?' to 'what happens if it does?'. Ghost GDP is now being formalised in academic work rather than dismissed.
Touches 01 (supports), 09 (supports) · Monthly report, June 2026
2026-06-15No broad labour-market collapse: unemployment near 4.3%, software-engineer postings still positive year on year
Layoffs and restructuring exist. There is no nationwide collapse in white-collar employment, and critics point to organisational and infrastructure bottlenecks on deployment.
Touches 04 (contradicts), 05 (contradicts) · Bureau of Labor Statistics — The Employment Situation
2026-06-15No demand-collapse feedback loop and no mortgage stress attributable to AI
The memo's loop of AI to layoffs to reduced spending to weaker revenue to more layoffs is not operating at macro scale. The mortgage component remains entirely hypothetical.
Touches 06 (contradicts), 12 (unresolved) · Monthly report, June 2026