The Writing Doom Loop
Source: Alex Kotran Substack
Author: Alex Kotran, Nathan Kriha
Original source: https://alexkotran.substack.com/p/the-writing-doom-loop
Published: 2026-02-11
Source type: essay
Private backup: the full article text is archived in the private repository at archives/articles/alexkotran-substack-com-the-writing-doom-loop.source.md. It is not published on the public Quartz site.
Summary
Alex Kotran and Nathan Kriha argue that generative AI is weakening writing’s ability to signal skill, effort, originality, and thought. They use research on Freelancer.com proposals to show how AI-generated applications can make high- and low-skill candidates appear more similar, creating a hiring “doom loop” in which polished writing loses predictive value. They extend the same logic to admissions essays, online writing, and school assessment. The educational danger is that students may bypass the productive struggle through which writing develops synthesis, perspective, voice, and agency.
Pull quotes
Polished AI writing weakens signal value
“After widespread AI adoption, that correlation breaks down. AI-assisted proposals look polished regardless of the applicants’ underlying skill, and thus lost their value as predictors of good work.”
Signal collapse can make markets less meritocratic
“Galdin and Silbert used a structural model to simulate a world where written applications no longer signal anything about ability and find that the market becomes less meritocratic: workers in the top quintile are hired 19% less often, while those in the bottom quintile are hired 14% more often.”
Writing still needs productive struggle
“We cannot allow widespread adoption of AI tools to smooth over the friction, the productive struggle, students encounter as they learn to write and critically analyze text — when a student outsources writing processes to AI in the spirit of efficiency, they are bypassing the work required to synthesize disparate ideas, challenge their own perspectives, and strengthen their critical thinking and analysis skills.”
Big ideas
- Learning still needs some struggle, even when AI can make things easier
- Students need to bring the purpose; AI should not supply it for them
- AI literacy should help people notice how AI changes what counts as knowing
- AI is changing what knowledge work asks people to do
Claims
- AI-generated text can make finished writing less trustworthy as evidence
- Learning requires some productive struggle that AI can remove
- Take-home essays are no longer reliable evidence by themselves
- AI changes how people come to know things, not just how fast they work
Key evidence and examples
- The article cites research on Freelancer.com showing that AI-assisted proposals made application quality less predictive of actual worker ability.
- A simulated labor-market result suggested top-quintile workers were hired less often and bottom-quintile workers more often when applications lost signal value.
- The authors compare this to a market-for-lemons problem in which readers cannot distinguish high-effort human writing from cheap synthetic text.
- College admissions essays and AI screening tools are used as examples of institutions responding to signal collapse with more automation.
Education relevance
Highly relevant for writing instruction, academic integrity, college readiness, admissions, assessment design, and the question of why writing still matters when AI can generate polished prose.
Durability note
The labor-market examples may evolve as hiring systems adapt, but the core education concern—writing losing value as evidence when process is invisible—is durable.