Lao Gan Ma Cut 600 Workers to 48. The Story Isn't AI. | Feast - Vocal Media

TEXT START: The viral "AI factory" headline hides a quieter shift in food manufacturing: cameras, airflow, and traceable chili batches.

The Dissection

The text is performing a useful demystification and then using that demystification to understate the event's systemic meaning. It correctly strips away the marketing label: the factory uses narrow computer vision, sensors, traceability, and mechanical automation. But the technical distinction between “AI” and older automation is economically secondary. Six hundred workers became forty-eight. That is capital replacing labor in measurable form.

The article accidentally documents the Discontinuity Thesis while trying to neutralize it. Its own figures show the repetitive work disappearing, the surviving workforce shrinking, and replacement jobs being fewer. The camera does not need to be intelligent in a human sense. It only needs to be cheaper, faster, and more consistent than human labor.

The Core Fallacy

The central error is confusing narrowness with harmlessness. The system does not need general intelligence to damage the employment circuit. It only needs durable superiority over a specific task. A conveyor camera that eliminates 552 positions is already P1 operating through industrial automation, whether the system is called AI, machine vision, or a glorified sensor array.

The article also treats the remaining 48 experienced checkers as evidence that human judgment survives. They are better understood as a temporary verification layer and a lag defense. If taste, defect categories, or sensory standards become more measurable, codifiable, or economically replaceable, that layer contracts too. “The soul of the sauce” is rhetoric, not a labor-market moat.

The text recognizes that capital replaces labor but stops before the actual conclusion: when this pattern generalizes across sectors, retraining cannot recreate the volume of economically necessary human work. Transfers may preserve consumption. They do not restore productive participation.

Hidden Assumptions

• That the current limits of sensory automation are permanent rather than temporary technical and commercial constraints.
• That the 48 remaining workers are indispensable rather than serving as a costly verification layer awaiting further compression.
• That new maintenance, labeling, tuning, and operations roles will be numerous enough to offset the positions removed. The text itself admits they are often fewer.
• That social support and retraining can absorb displacement without addressing the collapse of labor’s role as the mass distribution mechanism for income.
• That automation will remain confined to measurable factory tasks instead of expanding as data, sensors, robotics, and models improve.
• That calling the system “AI” is the main distortion, when the actual distortion is treating the absence of general intelligence as evidence that labor displacement is not structurally important.
• That consumer demand growth and rising efficiency can continue without confronting who receives purchasing power after labor income erodes.
• That human taste is economically sovereign merely because it is difficult to encode today.

Social Function

Classification: partial truth, transition management, and ideological anesthetic.

The partial truth is real: this is not a general-purpose agent, and “AI” is being used as a loose prestige label. The transition-management function is also explicit: accept automation, retrain the displaced, provide support, and continue. The anesthetic lies in presenting the surviving human judgment as a durable sanctuary rather than a narrowing bottleneck.

The article makes obsolescence easier to swallow by renaming mass displacement “industrial upgrading” and emphasizing that the machine only handles the repetitive parts. But repetitive parts are precisely where millions of livelihoods reside. The text humanizes the remaining workers while treating the eliminated workers as an administrative adjustment.

The Verdict

This is a competent factory-floor description wrapped around a failed systemic conclusion. The article correctly identifies the technology as narrow automation, but narrow automation is enough to begin the death of the wage circuit. Lao Gan Ma has not disproved AI-driven obsolescence; it has supplied a clean exhibit: 600 workers reduced to 48, output and revenue restored, and the remaining human role converted into a smaller verification function.

The machine does not need to understand chili, language, or humanity. It only needs to make 552 workers economically unnecessary. That is the thesis in miniature.