Is AI causing a repeat of Front end's Lost Decade?

URL SCAN: Is AI causing a repeat of Frontend's Lost Decade?
FIRST LINE: What AI is doing to the jobs of programmers feels very familiar to a lot of us frontend developers – because it has happened to us before.

THE DISSECTION

This is a veteran developer's memoir of professional dispossession, written with genuine craft and analytical surface texture, ultimately serving as transition management propaganda dressed as lamentation. The author traces the parallel between JavaScript framework deskilling and current AI deskilling, correctly identifying the mechanism, then fatally retreats into "craft will endure" optimism. The Bauhaus section is aspirational window dressing. The whole piece is a sophisticated cope.

THE CORE FALLACY

The author assumes deskilling is a directional process with a survivable equilibrium — that the market will always need some fraction of deep practitioners. This is wrong by DT mathematics.

JavaScript framework deskilling preserved the value of the underlying craft because:
- Frameworks still required humans to build, compose, debug, and extend
- The abstraction layers had finite depth — eventually you hit HTML, CSS, the browser, the network
- Mass production of mediocre frontend still required large numbers of humans, just lower-skill ones

AI deskilling is categorically different because the abstraction has no floor. AI does not need humans to fill the leaky details eventually — it generates the details, tests them, refactors them, deploys them. The "abstraction always leaks" argument the author uses to demand deep craft? AI is the leak, and it leaks everywhere simultaneously. When you no longer need a human to interpret the leak, the leak stops being a problem.

The author writes: "at some point the abstraction will leak. And then somebody has to invest the time to actually deeply understand what's going on – and fix it."

This assumes the fixer is human. That assumption is not supported by current trajectory.

HIDDEN ASSUMPTIONS

1. Quality correlates with economic demand. It doesn't when AI quality is 80% of human quality at 5% of the cost. The author correctly notes companies ship terrible software and succeed — but then ignores the implication: if mediocre AI output is sufficient for business success, the demand for high-quality human output collapses even further.
2. The "pie grows" argument. "Because it's now easier and cheaper to do so, the size of the pie will continue to grow." This conflates volume of software production with demand for human programming labor. The pie grows. The human slice shrinks toward zero.
3. Stack Overflow analogy as comfort. The author frames LLMs as "just the next step from Stack Overflow copy-paste" — implying a continuity of craft practice. But copy-paste from Stack Overflow still required a human who understood the problem domain. Agentic coding increasingly does not.
4. "Nobody was ever fired for choosing React." The author uses this to illustrate quality not mattering. But the modern equivalent — "nobody was ever fired for replacing three senior devs with one AI platform" — is being written right now, at scale, in every board deck in Silicon Valley.

SOCIAL FUNCTION

Classified as: Transition Management Propaganda with Craft Aesthetic

This article does what sophisticated cope does — it correctly identifies the wound, dresses it in historical reference and personal credibility, then recommends the patient adapt rather than flee. The Bauhaus comparison is particularly telling: it cites Dieter Rams and Jonathan Ive as proof that craft survives industrialization. But Rams and Ive succeeded in a world where physical materials and manufacturing still required human design intelligence. Software has no equivalent constraint — the design and the production are the same act, and AI does both.

The piece is marketed to and consumed by developers who are precisely the population most at risk and least likely to receive accurate structural analysis. The framing — "this happened to frontend devs, now it's happening to all of you, here's how to think about it thoughtfully" — generates trust precisely because it is authentic firsthand testimony. Authentic testimony serving a false conclusion is the most dangerous content type.

THE VERDICT

The author has diagnosed the symptoms with clinical precision and then prescribed homeopathy. The deskilling mechanism is real. The grief is legitimate. The historical parallels are interesting. But the conclusion — that deep practitioners will persist as a smaller but meaningful cohort, that the pendulum will swing back, that Bauhaus-style integration of craft and industrialization is the answer — is survivorship bias expressed as historical analysis.

Under DT logic, the question is not whether quality developers will still exist. They will. The question is what economic role they occupy when the marginal cost of AI-generated code approaches zero and the demand signal for human cognitive labor collapses across the board. The author's own evidence points toward the answer: worse outcomes for workers, better outcomes for businesses using AI, and a slowly shrinking professional niche for those who refuse to abandon craft. Not because craft doesn't matter — it does — but because the market for human-produced code is being structurally devalued at the foundation level.

The Bauhaus survived. The craftsmen who built handmade furniture for mass market did not.