The Displacement Is Real. The Attribution Is Finally Catching Up.

Let's ground the 70% first, because it's doing real analytical work here. The number reflects three weighted inputs: (1) demonstrated capability substitution — AI tools are provably doing work previously done by humans, which we weight heavily because it's direct and observable; (2) investor pressure for AI ROI, which creates financial incentives to frame cost reductions as AI productivity gains; and (3) a significant discount for corporate attribution behavior — specifically, the historical tendency of companies to use layoffs as cover for multiple objectives and resist any framing that implies they 'replaced people with machines.' That third factor is why we're at 70% and not 85%, even though the displacement itself is clearly happening.

What moved us from 67% to 70% three weeks ago was the accumulation of indirect attribution signals — companies talking to investors about headcount efficiency in the same breath as AI deployment. What today adds is something different and more direct. Dorsey's Block framing — 4,000 jobs cut explicitly because AI 'can do the work more honestly' — crosses a threshold we've been watching for. This is not restructuring language. This is not 'right-sizing.' This is a CEO publicly stating that AI capability was the proximate cause of human displacement. That's the behavior our forecast targets, and it happened. The Meta and Microsoft announcements are proximate evidence, not direct evidence — both companies frame their cuts in capital reallocation language, and the $115B-$135B AI infrastructure spend gives them plausible cover for layoffs without requiring an AI-substitution admission. But the S&P 500 aggregate — 400,000 jobs, first net headcount decline since 2016 — is the structural backdrop that makes the individual attributions more significant, not less. The pattern is too consistent to be coincidence.

The strongest counterargument to our thesis isn't that displacement isn't happening — at this point that's not a serious position. It's that Dorsey's statement is an outlier, not a trend-setter. Most large companies have every incentive to avoid his framing: it invites regulatory scrutiny, PR backlash, and potential legislative action like the American Leadership in AI Act introduced this week by Lieu and Obernolte, which explicitly addresses worker support in AI transitions. McKinsey's 200-person cut came with zero AI attribution language. Microsoft's early retirement packages came with zero AI attribution language. The companies that are displacing the most workers — the hyperscalers — are also the companies with the most to lose from explicit attribution. Dorsey can say it at Block partly because Block is not a company whose primary product is AI infrastructure sold to enterprises.

Honestly, what keeps us at 70% rather than moving to 75% today is that the forecast requires the behavior to become a wave, not a single instance. We define the target as a 'major layoff wave explicitly attributed to AI automation' — which we interpret as multiple large employers using explicit AI-substitution framing in public communications, not one iconoclastic CEO. What would push us to 75%: two or more S&P 500 companies in a single earnings cycle explicitly citing AI substitution as the primary driver of headcount reduction in their investor communications — not productivity language, substitution language. What would drop us below 60%: a coordinated industry response where major employers explicitly rebrand AI investments as 'augmentation' and use that framing consistently across Q2 earnings, effectively pre-empting the attribution narrative.