YouTube Cope Score

← Back to YouTube Scanner

Software Engineers Are Getting Dumber. I'm One of Them.

MadeByMHS

What the video argues

MadeByMHS argues that widespread LLM use has changed software engineering from a craft requiring sustained logical and problem-solving effort into an oversight role where engineers approve and direct AI output. He says code has become cheaper and less of a bottleneck, leading companies to accelerate roadmaps and demand more software. He worries this workflow causes engineers’ technical reasoning, domain understanding, debugging ability, and sense of fulfilment to atrophy, comparing it with the loss of navigation skills through over-reliance on digital maps. He forecasts that junior hiring, already badly damaged, will be followed by pressure on mid-level and senior engineers, with some workers moving into management while AI handles more coding. He describes an Anthropic experiment in which non-AI participants performed better on a later quiz, except for a small subgroup that used AI as a tutor rather than simply accepting its output. He recommends using AI to explain concepts, verifying its answers, reading documentation, debugging manually, taking notes, and developing broad expertise across the software stack.

Through the lens of the Discontinuity Thesis

The video partially recognises discontinuity. It explicitly says junior hiring has taken a battering, predicts that mid-level and senior roles will follow, and acknowledges that AI is performing much of the underlying work. That is substantially more realistic than an outright augmentation narrative. However, the argument stops at occupational skill atrophy and assumes that humans will remain economically necessary by moving upward into management, domain expertise, verification, or AI supervision. Its claim that cheaper software will simply generate demand for more software treats productivity expansion as a compensating source of employment, without asking whether firms need fewer total workers or whether the gains accrue to capital owners. The AI-tutor and full-stack-learning prescriptions are solution cope: they may help individual engineers remain competent, but cannot restore aggregate labour demand or the wage-consumption circuit. The video therefore offers a real diagnosis of displacement followed by an adaptation narrative, warranting heavy rather than terminal cope.

Butcher's verdict

He gets one important thing right and then flinches. The lines about junior hiring being battered and mid-levels and seniors coming next are the clearest encounter with the discontinuity. But instead of following that logic toward fewer humans, weaker bargaining power, and capital capture, he repackages it as a personal maintenance routine: read the docs, keep notes, debug line by line, use Claude as a tutor, and try to become the supervisor of the system replacing you. The claim that more software will be demanded is the quietest piece of cope: it turns a cost-cutting productivity shock into a bigger-workload story, allowing everyone to pretend that output growth automatically means employment growth. The managerial escape hatch is no escape hatch at scale; firms can need fewer coders and fewer supervisors. This framing benefits employers by converting structural exclusion into an engineer’s skill deficit and keeping workers busy optimising themselves while the labour market is hollowed out.

🎯 Scapegoats

none explicit

🛠️ Cope Mechanisms

augmentation-fantasy managerial-transition-framing omission_cope individual-responsibility-shift
Scored: 2026-11-13 09:19:24 Transcript: 17,907 chars Watch on YouTube ↗
The Cope Report
Weekly. Free. No cope.
The week's most revealing AI coverage,
scored for omission. Every Monday.
Got feedback?

Send Feedback

Custom GPT Ask the Oracle