Cendyne says their employer eliminated its last engineering role because Fable was cheaper, not because it was better at the job. The essay separates what coding models can do from what their marketing persuades employers to do. This is a firsthand account and an argument, not evidence that AI explains every layoff.
The distinction runs through three claims:
- Generating a requested feature is different from deciding whether it belongs in the product—and accepting the maintenance commitment it creates.
- Cendyne argues that models lack the long-term judgment needed to weigh operating costs, complexity, accessibility, and customers’ ability to afford the product.
- They nevertheless see real value in supervised code migrations: humans choose the target behavior, supply the tools and checks around the model, and measure progress step by step.
The unsettling point is not that engineering judgment has become worthless. It is that a model can change a hiring or firing decision before it can reliably take over the responsibility. Being better at the work and being worth retaining are no longer the same argument.
The 266-comment thread on Hacker News adds competing accounts of what “good enough” means—and who remains responsible afterward.
What the thread adds
- thefz — asks who fixes failures and carries liability: “When a bug introduced by the AI creates real damage to you, or your org, or third parties, who is responsible? You or the AI?”
- hackyhacky — argues that replacement need not match expert quality: “It’s not because it makes human skill useless, it’s because it makes non-human skill good enough.”
- rnd33 — reports a different failure mode from outright broken code: “The bugs fixed were real, the features added are in some ways useful, but they really do not move the needle in any particular direction.”
- glimshe — challenges the causal explanation rather than the author’s experience: “In other words, would those people have kept their jobs at current interest rates if there was no AI?”
- gregwebs — offers the opposing outlook, shifting the role toward accountable delivery: “Users of a paid software product still need a human to be responsible for it.”
- Veelox — supplies a counterexample to the loss of programming satisfaction: “As a counterpoint, I have enjoyed that last several months greatly. I have been able to solve real, long standing, issues quickly and efficiently.”
These comments do not establish which employers will succeed with smaller teams. They expose the unresolved boundary between making software cheaply and owning the consequences when it fails.
HN handles are pseudonymous, and HN publishes no per-comment scores. The source ordering is HN’s own ranking; this usefulness-selected slice of the thread is not a consensus.