Fable Took My Job: An Engineer's Account of AI-Adjacent Layoffs and the HN Debate It Sparked

An engineer writing as cendyne published "Fable took my jorb" on 2026-10-11, arguing that frontier AI models did not outperform him at his job — they were simply cheaper than his compensation, which gave the company's remaining leadership enough confidence to coast until an acquisition. The essay reached 224 points and 362 comments on Hacker News, where the discussion split into three broad camps: those who say a job is just a job, those who say AI is a convenient excuse for cuts driven by interest rates and offshoring, and those who say software engineering as a career is finished.

The layoff was economic, not a capability comparison

The author states plainly that the model was not better at his work. He was the last engineer remaining, and the model was cheaper than what he cost.

It wasn't that Fable was better at my job. As the last engineer, Fable was cheaper than my compensation. Evidently, it gave enough confidence to the last vestiges of leadership that they could coast until M&A.

The essay is explicit that this was not the author's first exposure to layoffs. He had previously been on the executing side, having to justify headcount against a budget ceiling and then telling two of his own reports at a house party that they would lose their jobs the following week. He describes the physical toll of the final layoff: heart rate above 100 bpm for over a week, and roughly four effective hours of sleep out of ten offered.

The author reports the cuts were not isolated to him. Six or seven people in a DevOps interest group he belongs to were laid off or received notice within a single month. One hiring manager who interviewed him at another company was laid off a week later, after that company's CEO published a late article about AI changing software development. Another contact lost an entire team — while actively interviewing people in the USA — because of a top-down decision to offshore again, on the reasoning that AI assistance had raised the capability of a less passionate, less experienced, and less costly workforce.

The author's core claim: models lack taste and foresight

The essay's central technical argument is that current frontier models cannot make product decisions because they lack two things: taste and consequence-tracking.

On taste, the author contrasts a product manager who implements whatever a Reddit user complains about with a leader who refuses a feature because committing to it changes what the product is:

It is cheaper than ever for a product manager to say "Someone complains about this on reddit, let's implement it", while the tasteful leader says 'No, that's not what our product is. Making that means we commit to it for a long time and our customers really want {something else we already do, but not perfectly} instead.'

On foresight, the author argues models cannot predict the downstream consequences of decisions across weeks, months, or years — in operating costs, complexity, accessibility, and affordability to the customers the product exists for. He calls these secondary effects the things that actually make a product worth buying, and says AI-generated "slop services" omit them consistently enough that users can perceive the absence.

Hacker News commenter A_D_E_P_T agreed and extended the point: because doing things is now cheap, discernment has become the scarce resource. In that commenter's framing, an agent will happily agree to every product-bloating change and offer to implement it immediately.

Where the author concedes AI is already working

The essay is not a blanket denial of model capability. The author cites two examples where he believes AI-driven development is genuinely effective.

First, Shopify's engineering post "Back to Native," which the author reads as Shopify deciding to abandon React Native and return to native iOS and Android apps, machine-coded but overseen by experienced, tasteful developers. The author says he agrees with that decision, because models are now good enough to be driven step by step through a carefully crafted harness to reproduce behaviors that have already been A/B tested — "They can push a rock up a hill. Then up the next hill, and so on — when the technique can be measured and check-pointed."

Second, Theo's "for the memes" TypeScript-in-Rust port (github.com/pingdotgg/ts-rust), which ported the ts-go code to Rust. According to Theo, Opus 5.5 reproduced bugs from the Go version in the Rust version exactly. The author notes the counterexample in the same breath: Astra, described as the rival to Fable, failed at something familiar, well known, and well documented.

The author's synthesis is that agents are "terribly good at copying things across domains" but that this does not amount to replacement:

We're heading towards the end of the year and I still do not see how these copy-machines with "super intelligence" inside are going to replace humans. What they have done is convince those with money to spend differently on their fellow humans.

"AI took my job" versus "AI contributed to my role being eliminated"

The essay draws a careful distinction between causation and rationalization:

"AI took my job" is far more fun to say than "AI contributed to my role being eliminated."

Commenter philipodonnell pushed the distinction further, arguing that attributing the outcome to AI lets the decision-makers off the hook: "AI did not take your job; a person decided to use the productivity improvement from a new technology to realize increased short term profitability at the cost of losing you."

Commenter glimshe raised the counterfactual that runs through the whole thread: how many of these engineers would have kept their jobs at current interest rates even if AI did not exist? In that reading, leadership would simply be citing a different reason.

The HN split: "it's just a job"

A large share of the highest-engagement replies pushed back on the emotional framing rather than the economics. danw1979 wrote that the author's description of the first layoff meeting — ringing ears, color draining from the world, barely able to walk straight — reflected taking work too seriously: "You turn up, they pay you, then you go home to your actual life, the one that matters." sajithdilshan made the same point as a general rule: "no matter how good you are, you could be obsolete and replaceable any day and the companies/corporations don't care."

The HN split: "AI is management's excuse"

fwlr offered the thread's most structural argument, framing the moment as Capital acting on belief rather than evidence:

The moment AI achieved the barest shred of plausibility in its promise to replace Labor, it became an article of faith that Capital feverishly believes in and aggressively acts on. All of this is to say that we have a case of "the market can stay irrational longer than you can stay solvent" on steroids.

voidifremoved supplied a historical parallel from the mid-2000s offshoring wave: when presented with metrics showing offshore quality was worse and spec overhead was non-trivial, management's answer was that for the price they could have the work redone five times and still save money. tasuki questioned the essay's framing of offshoring directly, noting that many Indian programmers are as passionate and experienced as anyone while still being cheaper, and asking whether the original characterization was fair or true.

The HN split: "software engineering is over, plan accordingly"

A third camp treated the essay as denial. Valakas_ wrote that people who predicted this outcome were mocked as cultists and that the discussion is now moving through the stages of grief, adding that the change "will extend to ALL jobs." small_model was blunter: "SWE as a job is over, this should have been obvious at start of the year... accept it and have a plan B/C/D in the works while you are one of the lucky ones where it company has not made the move yet."

The strongest counterargument in the thread

Commenter user43928 directly contradicted the essay's central claim that models cannot replace engineers, arguing that no elaborate harness is needed — only a text box, UX taste to select good concepts, and patience to QA the output. In that account, building software has become trivial and a team of five to ten becomes one competent person. The commenter preemptively dismissed maintainability objections as imagination rather than experience.

Other commenters noted the symmetry: wwind123 observed that if executives believe AI can replace engineers, most of those executives are themselves replaceable. chung8123 predicted that laid-off engineers will use the same AI tools to build alternatives to their former employers' products.

What the thread reveals about the debate itself

burger_moon pointed out that Hacker News runs two contradictory narratives in parallel: in this thread, software development is a dying career and tech companies will shrink because QA, PMs, analysts, and recruiters are no longer needed; in visa threads, the framing is that whoever brings the most human bodies into their country wins. The commenter's conclusion was that fear, not evidence, is driving most of these discussions, and that the threads function as therapy.

TrackerFF supplied the thread's longest historical perspective: an electrician apprentice colleague who was laid off from a factory job when production moved to China, restarted at entry-level pay in fisheries, was laid off again after an acquisition, spent 15 years delivering mail until postal volumes fell, and finally entered trade school in his fifties at apprentice wages. TrackerFF's argument is that blue-collar workers have absorbed repeated displacement without public alarm, and that UBI conversations only begin when professional-class jobs are affected.

The essay's closing position

The author's final framing is not that programming is dead but that its emotional reward has been relocated:

The joy of programming is dying. The joy of delivering results lives on.

Commenter poisonborz called that line the summary of the whole debate regardless of which side a reader takes. gregwebs offered the most concrete constructive reading: the job has changed from "writing code" to "shipping software," the technical lead role persists longer than the individual contributor role, falling software costs will expand total software demand, and the same tools let engineers become small business owners directing AI across an entire business.

Key facts at a glance

• Source essay: "Fable took my jorb," cendyne.dev, published 2026-10-11, roughly 8 minutes of reading.
• Hacker News submission: 224 points, 362 comments.
• Author's stated cause of layoff: he was the last engineer and a model was cheaper than his compensation, not that the model performed his job better.
• Models named in the essay: Fable and Mythos (marketed frontier tiers), Opus 5.5 (credited by Theo with reproducing Go bugs in a Rust port), and Astra (described as Fable's rival, which failed a well-documented task).
• Cited evidence that AI-assisted development works: Shopify's "Back to Native" migration and Theo's ts-rust port.
• Author's stated model limitations: no product taste, no multi-year consequence forecasting.
• Author's closing thesis: models have not replaced engineers, but they have convinced those holding budgets to spend differently on people.