Attribution Is No Longer Hypothetical: The AI Displacement Wave Has a Paper Trail
textak has held a 73% probability that we'd see a major layoff wave explicitly attributed to AI automation — and the first half of 2026 has delivered evidence stronger than we anticipated when we set that number. Jamie Dimon named AI as the cause on the record. Jack Dorsey at Block attributed a 40% workforce reduction directly to automation in a named CEO statement. GitLab called it 'Act 2' and cut 14% explicitly to redirect savings toward AI agents. The structural barrier to this forecast wasn't capability — it was corporate attribution behavior. That barrier is down.
Our 73% reflected a specific thesis: companies were displacing workers through AI but avoiding public attribution due to PR risk. What we weighted heavily was the investor pressure side — eventually, the ROI case for AI investment requires showing where the savings come from, and that creates a structural incentive toward disclosure that outcompetes the reputational caution. Q1 and Q2 2026 confirmed that dynamic. The Skillsyncer data showing 56% of layoff announcements citing automation across 156,270 workers is the most direct evidence we've tracked to date. This is not anecdote — it's a population-level pattern.
The strongest counterargument we've held against ourselves: most displacement would happen through attrition, not announced layoffs, meaning the public attribution criterion would never get satisfied at scale even if the phenomenon was real. The Block case specifically undercuts that counter. Dorsey didn't wait for attrition — he made an explicit 40% cut and named the cause. GitLab's announcement was structurally identical. These aren't the slow-rolling non-renewals we worried would be invisible. They're discrete events with named executives and explicit AI attribution in the press record.
What we're still watching: the manufacturing and logistics data (20-30% labor reduction) is directionally consistent but remains largely anonymized. Large-scale displacement in knowledge work — legal research, data analysis, software development — is described in aggregate reporting but has fewer named-firm examples outside tech. The forecast as written requires 'major layoff wave explicitly attributed' — we think that criterion is now satisfied by the aggregate record, but readers in regulated industries may reasonably argue the threshold requires sector-specific named examples beyond tech.
What would move us below 60%: if Q3 earnings cycles show companies pulling back from explicit attribution — framing cuts as 'restructuring' or 'efficiency' without naming AI — that would suggest the H1 disclosures were anomalous rather than a durable trend change. We don't expect that. The investor incentive structure hasn't changed. But we're watching the Q3 earnings language closely, and we'd treat three consecutive major tech earnings calls that avoid AI attribution as a meaningful signal to revise down.