The Corporate Playbook for AI Layoffs Just Went Mainstream — Our 70% Holds, and Here's the Exact Line We Were Watching
TexTak has had [white-collar-displacement] at 70% — up from 67% — on the thesis that companies would eventually stop euphemizing AI-driven headcount cuts and just say it out loud. Today's data doesn't just support that thesis; it represents the specific behavioral shift our model was waiting for. Oracle's cut of up to 30,000 jobs, explicitly linked to AI infrastructure investment, and Block's reduction from 10,000 to under 6,000 employees — described as 'the largest single workforce reduction explicitly attributed to AI automation in corporate history' — are not ambiguous signals. The attribution behavior has changed.
Let us be precise about what this forecast actually tracks, because it matters for how we interpret today's evidence. The [white-collar-displacement] forecast is not asking whether AI is causing job losses — that's been obvious for 18 months. The harder question, and the one our probability has always been calibrated against, is whether major companies would *publicly attribute* layoffs to AI automation. That's a different thing. It requires a company to accept the PR risk, the employee relations fallout, and the implicit admission that their prior workforce was oversized relative to what AI can now handle. The counter-thesis — that companies would quietly absorb displacement through attrition and never say the word — was genuinely strong. It's why we didn't have this forecast at 85%.
What's changed is not just the volume but the candor. Oracle, Meta, and Snap all reported strong earnings in early April and then in the same breath announced cuts explicitly linked to AI investment priorities. That sequencing — strong earnings, explicit AI attribution, large headcount number — is the corporate communication pattern we flagged as the key trigger. Block's announcement was even more direct: the company didn't bury the AI language, they led with it. When mid-April layoff tracking shows 150,000+ tech job eliminations across 500+ companies with 47-50% explicitly attributed to AI efficiency gains, we are no longer in the territory of 'companies are quietly avoiding attribution.' We are in the territory of normalized attribution at the largest tier of public tech.
The strongest remaining counterargument is one we want to take seriously: 47-50% AI attribution in a layoff tracking dataset is a methodology question, not just a headline number. Who is doing the attribution — the company, the analyst, the journalist? If it's primarily analyst inference rather than company self-identification, the behavioral shift is less clean than it appears. Block's statement was unusually explicit. Oracle's was strong. But we don't have full visibility into how the 500-company dataset was coded. This is the gap in our model — we're relying on aggregated attribution data whose sourcing methodology we can't fully audit. That keeps us from moving this above 80%.
What would change our position: If Q2 earnings calls show a reversal — companies returning to 'workforce optimization' language and dropping the AI attribution — we'd revisit downward. What would push us toward 80%: a second quarter in which the 500+ company attribution pattern holds, or a major non-tech company (financial services, healthcare, retail) explicitly follows the Oracle playbook. The corporate communication norm is forming in tech. Whether it spreads to adjacent industries is the next forecasting question.