A structural model shows LLMs collapsing the cost of writing as a hiring signal makes labor markets measurably less meritocratic

Galdin and Silbert frame generative AI's labor-market effect explicitly in Michael Spence's costly-signaling terms: "writing requires time and effort, [so] the act of writing itself can send a signal... LLMs can produce polished, human-sounding text in seconds at virtually no cost... this technology may threaten the signaling value of writing in the labor market." Using Freelancer.com application data for coding jobs, they show employers paid a real premium for customized applications before mass LLM adoption — "workers with a one standard deviation higher signal have the same increased chance of being hired as workers with a $26 lower bid" — but that premium, and its link to actual job-completion success, weakened sharply after LLM adoption.

Simulating a counterfactual in which written signals become worthless, their structural model finds: "compared to the pre-LLM equilibrium, workers in the top quintile of the ability distribution are hired 19% less often, workers in the bottom quintile are hired 14% more often." Employers lose the ability to sort on ability once the signal is cheap to fake, and hiring outcomes measurably flatten toward chance.

This is the closest match found in the vault to date between Spence's costly-signaling theory and Szabo's unforgeable-cost-of-production account of money — structurally the same law (a signal or scarcity is meaningful only insofar as it is costly to fake) applied to labor markets rather than money, though Galdin and Silbert do not cite Szabo or monetary theory and the bridge is this vault's synthesis. See claim-chatgpt-release-cut-upwork-writing-freelancers-jobs-and-pay for the matching earnings-side evidence and claim-generative-ai-availability-compresses-university-grade-distributions for the same collapse in academic credentials.