Peter Bryant's essay makes an argument that has been quietly circulating in engineering circles for years but rarely stated this cleanly: software developers are skilled tradespeople, not knowledge workers. They make and fix things. The material is bits instead of atoms, but the work — years of apprenticeship, pattern recognition, hands-on craft — maps far more closely to electricians and carpenters than to the managerial and administrative categories that define white-collar work. The reclassification isn't a demotion. Bryant is careful to note that recognising development as a skilled trade gives more credit to the work, not less. Electricians and carpenters spend years mastering their craft. Formal education helps but isn't strictly required. The same is true of software. The prestige gap between 'trade' and 'profession' is a cultural artefact, not a reflection of skill or value. The real payload of the argument is about AI. Bryant contends that the current AI identity crisis among developers is partly a product of the misclassification. White-collar workers historically bear the brunt of technological disruption. Blue-collar tradespeople, by contrast, have repeatedly absorbed new tools — power tools replaced hand tools, CNC replaced manual machining — and adapted. If developers had always understood themselves as tradespeople, the arrival of AI coding assistants would feel less like an existential threat and more like the introduction of the chainsaw to lumberjacks. Bryant names the honeymoon directly: thirty years of protected time, inflated salaries relative to peers, and unchallenged authority over timelines and complexity estimates. Developers were treated like rock stars because non-technical colleagues couldn't verify their claims. AI tools don't just automate code — they make the work legible to outsiders, collapsing the information asymmetry that sustained developer privilege. The essay's historical analogies — seamstresses and looms, accountants and spreadsheets — are well-chosen because they all resolved the same way. The trade survived; the monopoly pricing didn't. Seamstresses still exist. So do accountants. But neither commands the premium they did before their tools became accessible. What's missing from the analysis is the structural question: who benefits from the reclassification? If developers are tradespeople, the wage premium shrinks, the mystique dissolves, and management gains leverage. The essay frames this as a natural correction, but it also happens to align perfectly with the interests of the companies currently deploying AI to reduce headcount and compress engineering salaries. The loom analogy cuts both ways — yes, people adapted, but the transition destroyed guilds, crushed wages, and concentrated wealth in factory owners. Bryant closes with a gentle consolation: you can still sew and chop wood by hand, for the love of it, even if you can't get paid for it anymore. It's an honest ending, but a stark one. The question isn't whether developers will adapt — they will. The question is who captures the productivity gains from that adaptation, and so far, the answer is overwhelmingly not the developers.
