Meta Cut 8,000 People. It Has Nothing To Do With AI Working.

The video challenges the popular narrative that 'AI layoffs' represent a single coherent trend, arguing instead that the label is being misused to cover fundamentally different corporate situations. The speaker frames these layoffs as high-stakes strategic signals that leaders and job seekers should decode individually rather than aggregate.

The first category is the hyperscaler layoff, exemplified by Meta's 8,000-person cut. The speaker argues this is not evidence that AI is working so well that fewer people are needed — rather, it reflects Meta's need to offset massive GPU and data center capital expenditure with a compelling operating expense reduction story. The speaker notes that Meta is reportedly using Claude internally rather than its own Llama model, and is even discussing a 'Plan B' of becoming a cloud compute provider if it cannot catch up in AI. Meta's competitive layoffs are also shaped by its internally competitive, tiered culture, which the speaker claims encourages token-burning for status rather than outcome-driven work.

The second category is the visionary leader layoff, with Jack Dorsey and Block as the primary example. The speaker credits Dorsey for seriously grappling with the idea that the firm itself is becoming intelligent and must be restructured accordingly, drawing an analogy to the early days of electrification when factories hadn't yet been redesigned around the new technology. However, the speaker criticizes Dorsey for not going far enough in thinking through the human and change management implications of AI transformation, and argues that leaders who are afraid of getting technical will struggle to envision the scale of workflow changes required.

The third category is the activity-based layoff, with Cloudflare's '600% usage increase' justification as the key example. The speaker argues that using token consumption or AI usage metrics as the basis for workforce decisions is fundamentally flawed, because individual activity does not linearly translate into firm-level outcomes. The speaker also references Uber's reported frustration with its Claude budget as another symptom of companies that are mentally stuck in a last-year's-individual-productivity model rather than thinking about agentic pipelines and outcome-aligned organizational design.

The fourth category is the hope-based layoff, exemplified by Cisco, where companies use layoffs to tell an AI transformation story to markets without having strong usage statistics or a clear strategic vision to back it up. The speaker understands the board and Wall Street pressure behind this behavior but warns that it is counterproductive long-term, since companies will need their best people to execute AI transformation and cannot know in advance who is expendable.

The speaker concludes by noting a fifth unlabeled category — layoffs that have nothing to do with AI and everything to do with poor business performance or over-hiring — and urges both leaders and job seekers to resist the temptation to treat all AI layoffs as a single phenomenon, emphasizing that each type carries distinct competitive intelligence and career risk signals.