“Can large language models actually replace coders? Not really, no, and I’ll get into why later in the piece.”
Our reading, the sentence that was checked:
Large language models cannot really replace coders.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, The Case Against Generative AI, paragraph 30.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-the-case-against-generative-ai-c5d9b3-a1
We recorded 39 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-09-29. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- large language models
- actually replace coders
- Not really, no
Our interpretation
- The plain reading concerns practical substitution for working software developers, not whether models can generate any useful code.
- Replacement means performing substantive coding responsibilities without another human doing that programming. Ordinary management and review may remain.
- The statement describes contemporary capability. With no reporting calendar specified, I use the current calendar year through publication as the latest reportable capability period.
- Evidence from earlier in that period is not automatically representative of models available at its end.
Counts as supported if: Representative evaluations and documented deployments establish that contemporary systems still require humans to perform substantive coding responsibilities rather than functioning as practical developer substitutes.
Counts as contradicted if: Verified deployments establish that contemporary LLM systems reliably took over substantive developer responsibilities, or representative independent evaluations demonstrate that substitution under realistic conditions.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The studies establish important limits on early-2025 coding systems, but neither measures practical replacement of developers across workplaces. Employment evidence raises a credible substitution concern without proving that models took over those jobs. The available records therefore do not settle the broad claim.
Limitations
- Task success, developer productivity and job replacement are different outcomes.
- Early-2025 results may not describe every system available at publication.
- The employment counterevidence uses contemporaneous news reporting because the original primary URL now supplies a later revision.
Evidence (3 items)
Context only: METR / arXiv,
AI actually increases completion time by 19%
Why it matters here: The February–June study found slower completion among experienced developers using AI on familiar repositories. It measures assisted productivity in a limited setting, not whether LLMs can substitute for developers across workplaces.
- Where
- Abstract
- Publisher
- METR / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: OpenAI researchers / arXiv,
unable to solve the majority of tasks.
Why it matters here: SWE-Lancer documents substantial limits on real freelance-task benchmarks. Its results do not establish the performance of every contemporary model or whether successful tasks translated into workplace substitution.
- Where
- Abstract
- Publisher
- OpenAI researchers / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: TIME,
refrained from directly saying that AI was definitely causing unemployment.
Why it matters here: TIME reports declining employment among young workers in AI-exposed occupations, including software engineering, alongside stronger employment among older workers. This is material counterevidence to dismissing substitution, but the report does not establish that LLMs performed displaced developers’ substantive work.
- Publisher
- TIME
- Kind of source
- News report
- Describes events of
- Public from
What would settle it
Dated workplace records showing who performed substantive programming after deployment, or representative independent evaluations of complete developer responsibilities using systems available by publication.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, The Case Against Generative AI
Can large language models actually replace coders? Not really, no, and I’ll get into why later in the piece.
- Where
- paragraph 30
- Published
- 2025-09-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-the-case-against-generative-ai-c5d9b3-a1: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment wyea-the-case-against-generative-ai-c5d9b3-a1.