Monitor 001 / Check-in 03 of 8 / APRIL 2026
Two analysts wrote a memo from June 2028 about an AI-driven crash. Every month we check which parts of it are coming true. This is the April 2026 reading, kept exactly as it was written. Source memo: THE 2028 GLOBAL INTELLIGENCE CRISIS, “What follows is a scenario, not a prediction.”.
So far · Machines: Happening. People: Starting. Money: Symptom only.
A quiet month that moved one thing decisively. Aggregate private-credit defaults set another record, so the financial stress the memo describes is now unambiguously present. The attribution still is not. We are also noting that the most careful independent forecaster in the adjacent field has revised his own timeline, which is a useful reminder about what we are doing here.
Up from 5.8% in January. The rise is led by consumer products, healthcare services and smaller borrowers.
Touches 10 (supports) · Credit and Collection News — Credit Defaults Hit Record 6% As Private Lending Market Faces Stress
Follows their February grading exercise, which found quantitative metrics running at about 65% of the pace AI 2027 assumed. Kokotajlo's own median has drifted later even as capability news improved.
Included because it is the closest thing we have to a peer monitor, and it shows the same split we keep finding: capability on track, consequences slower.
Touches 01 (supports) · FutureSearch — AI 2027 Six Months Later: Karpathy, Kokotajlo, and Shifting AGI Timelines
Recording the distinction explicitly this month: the symptom the memo predicted is present, the mechanism it predicted is absent. This is the first time we have written that sentence and it is likely to recur.
← March 2026 May 2026 → All check-ins
A predicted symptom appearing does not mean the predicted mechanism caused it.