A dated feed of new evidence, news, and ideas — the living layer that keeps this work current. Showing August 2026.
A rich, deeply biblical essay (adapted from the forthcoming book The AI Apocalypse). Chris Watkin warns against defining our humanity by ‘the gaps’ AI hasn’t filled yet, and reframes the whole question: our worth was never our output. AI can imitate human words but never participate in them — it offers ‘costless reciprocity’ where the gospel is ‘costly mutuality.’ As Watkin puts it, AI ‘cannot kneel … it cannot repent; it cannot commit; it cannot sacrifice.’ A clarifying, hopeful frame for exactly the questions this moment is raising. The Gospel Coalition →
A balanced state of the field. Humanoid-robot shipments hit about 19,100 in the first half of 2026 (up 272% year-on-year), and $8.7B of venture funding poured in through July — nearly double all of 2025. Real work is happening: Figure’s robots logged 1,250+ hours at BMW’s Spartanburg plant (handling 90,000+ parts), and Agility’s Digit moved 100,000+ totes at a GXO warehouse. But the industry’s own federation (IFR) is measured — humanoids are more likely to complement workers than replace them, and the sector is still in a ‘validation phase.’ Embodied AI is genuinely arriving in physical work — steadily, unevenly, and slower than the hype. AI Insider →
A two-part Economic Brief series from the Richmond Fed. The first (“Worker Types, AI Exposure and the Recent Decline in Job-Finding Rates”) finds the slump in finding work is hitting strongly attached workers in AI-exposed occupations hardest — not the usual recession pattern. The second (“Aggregate or Structural?”) is the honest limit: AI is a real, historically unusual structural force, but accounts for at most about a third of the total decline — less-exposed workers explain roughly half. Official, measured evidence that the disruption is real and that not everything is AI. Federal Reserve Bank of Richmond →
The August revision of Stanford’s ‘Canaries in the Coal Mine’ study (using ADP payroll data) finds no economy-wide displacement — but workers ages 22–25 in the most AI-exposed jobs now sit about 19% below where they’d be otherwise, up from 15% a year earlier. The cause isn’t layoffs; it’s that they were never hired. The first-job on-ramp is quietly narrowing for the very generation most exposed to this shift. (full paper) Stanford Digital Economy Lab →
A peer-reviewed Stanford study (in Nature Human Behaviour) of 1,131 Character.AI users found that people with the smallest offline social networks who confided in chatbots for companionship reported the lowest well-being — and the more they opened up, the worse it got. The researchers call it ‘social junk food.’ A stark confirmation that machines can simulate presence but not provide it — and that real, embodied community is exactly what this moment will leave people hungry for. Stanford Report →
Want these in your inbox? Subscribe.