Meta cut 8,000 jobs. CEO Mark Zuckerberg says AI is creating more work-but not inside Meta

TEXT START: The tech giant laid off around 8,000 employees, reassigned another 7,000 workers to AI projects, and is preparing to spend $130 billion to $145 billion on AI infrastructure in 2026.

The Dissection

The article is performing narrative containment. It concedes that Meta is destroying cognitive jobs, then points to construction, electrical, cooling, and power-infrastructure work as evidence that AI “creates jobs.” This converts a structural displacement event into a labor-market reshuffling story.

The infrastructure boom is real, but it is capital-deployment work surrounding AI—not proof that AI preserves mass productive participation. Meta is concentrating labor around a smaller technical core while spending unprecedented sums to replace and amplify human cognition.

The Core Fallacy

The article equates any employment with equivalent employment. That is the central error.

Temporary, geographically concentrated construction and infrastructure jobs do not replace thousands of durable corporate knowledge-work roles. They are lag employment created by the physical bottleneck of AI deployment. Under the Discontinuity Thesis, this is a lag defense, not a reversal: AI can create jobs while simultaneously severing the mass employment → wage → consumption circuit.

The relevant question is not whether AI creates some jobs. It is whether displaced workers retain economically necessary participation at scale. Meta’s layoffs answer that question plainly: no.

Hidden Assumptions

• Infrastructure jobs will be permanent rather than tied to a finite construction cycle.
• Displaced Meta employees can access and perform those jobs.
• New infrastructure wages will match the scale and distribution of eliminated cognitive wages.
• AI-driven revenue growth will generate broad labor demand rather than accrue mainly to asset owners.
• Construction, energy, cooling, and maintenance will remain human-intensive as automation advances.
• Spending $700 billion proves productive success rather than an arms race that may destroy cash flow and concentrate ownership.
• Renting compute is a social employment engine rather than another mechanism for capital centralization.

The article also quietly shifts the unit of analysis from workers to jobs. That is how elite narratives hide the corpse: count openings, ignore who lost bargaining power, who can qualify, how long the work lasts, and who owns the machinery.

Social Function

This is partial truth serving transition management and elite self-exoneration, with a layer of ideological anesthetic.

It acknowledges the layoffs enough to appear honest, then uses visible construction activity to imply that the system is adapting normally. Zuckerberg’s statement is technically defensible and structurally evasive. He is describing the temporary labor demands of building the replacement system while avoiding the permanent question of who remains necessary after it is built.

The Verdict

AI is creating jobs around the construction of its own labor-replacement infrastructure. That does not save post-WWII capitalism; it marks the transition toward it.

Meta’s 8,000 layoffs are the signal. The data centres, power systems, and cooling plants are the scaffolding. Once the physical bottlenecks mature, the temporary jobs narrow while ownership of compute, energy, logistics, and maintenance becomes more valuable. The article mistakes scaffolding for a new social contract.