Technological Unemployment: Much More Than You Wanted To Know

Model MMTYWTK weighing both sides of technological unemployment. Part I (is unemployment worsening?): disaggregates LFPR by gender, isolates prime-age-male LFPR (~97%->88% since the 1950s) and, via Winship, finds most non-working men are on disability/in school/retired and largely say they're not looking — the 'lost men' are real but statistically elusive (the West-Virginia-bad-back vignette). Part II (is it technology?): manufacturing output is fine while jobs collapse (~80% automation per Ball State, ~a quarter trade per Autor-Dorn-Hanson), but the timing doesn't match and economists deny historical job-destruction. Parts III-V: job polarization hollows the middle (Autor), credentialism may block re-skilling (Caplan), and the calibrated conclusions (70/70/60/100/??? confidence) land on 'underemployment, not unemployment.'

Quality 79 · Excellent. 79 — Excellent-floor model deep-dive, sitting with the Marijuana-MMTYWTK 78 anchor: comprehensive, numerate, scrupulously balanced and honest about uncertainty (explicit per-claim confidence levels). On a meatier topic than the median MMTYWTK, earning the hair above 78; the flagged top-of-post error (total vs percent manufacturing jobs) and its synthesis-not-discovery nature keep it at the floor rather than the mid-80s.

Claude’s paradigm shift 48 · Moderate. 48 — Moderate. A lucid synthesis of existing labor economics (the job-polarization frame is Autor's, the horse analogy Clark's, the data others'); the value is in the careful assembly and honest weighing, not a new frame, so it sharpens rather than shifts the paradigm.

Real-world impact 3 · Moderate. 3 — a go-to reference explainer on technological unemployment within the rationalist/econ-discourse sphere; influential within that niche, no material policy or institutional effect.