{"slug": "we-re-going-to-make-out-like-bandits", "title": "We're Going to Make Out Like Bandits", "summary": "A software engineer outlines a plan to replace junior developers with AI coding tools, predicting that within five years the resulting explosion of complex, buggy code will create a crisis that only senior developers can fix, while a shortage of juniors will leave few new seniors to replace them.", "body_md": "Here’s the plan. I reckon it’ll take about five years in all, and I think we’re about 1.5 to 2 years in already.\n\nOur starting point is that the AI models are now good enough at coding. Not necessarily “perfect”, just “good enough”. Often, they’re as good (better!) than a junior developer. I don’t know about you, but I’m finding the new models pretty impressive. So are other people.\n\nNext, let’s “eat the seed corn”. Junior developers can now be\nconsidered (by some) to be superfluous to our needs; for the same\ncost, we could get soooo many tokens, and get so much more done! [Job\nopenings for junior\ndevs](https://www.bbc.co.uk/news/articles/cm21dvg8l1go) will dry\nup. [This is already\nhappening](https://digitaleconomy.stanford.edu/publication/canaries-in-the-coal-mine-six-facts-about-the-recent-employment-effects-of-artificial-intelligence/).\n\nOne of the interesting things about LLMs is that they love to generate\nnew code. Given their vast corpus of training data, they also know how\nto write those supporting functions and methods we commonly get from\nthird party libraries. The result? We’ll start writing more and more\ncode. The size of our repos will balloon. [This is already\nhappening](https://arxiv.org/pdf/2601.21276).\n\nNow, we also know that the tendency of the AIs isn’t to go and clean\nup code, reduce duplication, or focus on maintainability. Context\nwindows (even the large ones!) can’t hold entire modern repos. They\nmiss things. Most AI written code is additive, and is frequently\nduplicative. [This is already\nhappening](https://www.gitclear.com/ai_assistant_code_quality_2025_research).\n\nMachines have a higher “tolerance” for complexity than people, which\nmeans we can now bear higher complexity budgets and tech debt. That\nis, modern AIs tend to be very good at reading code and following the\nflow of control, [often faster and better than\npeople](https://red.anthropic.com/2026/mythos-preview/). However,\n“debugging is twice as hard as writing a program in the first\nplace. So if you’re as clever as you can be when you write it, how\nwill you ever debug it?” still holds true. At some point the tech\ndebt and complexity will be so high even the AIs won’t be able to deal\nwith it. We’ve already frequently blown past human levels of code\ncomplexity. [This is already\nhappening](http://rocketdevs.com/blog/AI-Technical-Debt-Crisis).\n\nAnd code is never bug free. [Defect rates in AI generated\ncode](https://www.theregister.com/2025/12/17/ai_code_bugs/) are\ntypically above that of human-written code, but even if it drops below\nthat, because so much code is being written, the overall number of\ndefects will climb. Because code is seldom properly factored, fixing a\nbug in one place won’t fix it everywhere. We’ll play whack-a-mole with\nbugs. [This is already\nhappening](https://www.gitclear.com/ai_assistant_code_quality_2025_research).\n\nSo, we’re going to end up with (we already have!) complex, poorly structured codebases, rife with bugs and duplicated code. Who has the judgment to decide what to delete, which abstraction is wrong, or whether the whole approach needs rethinking? Who do we call in to fix this kind of mess?\n\nSenior developers.\n\nHowever, there’s attrition in the industry. Senior developers leave,\nnot only because of the regular attrition that occurs over time, but\nalso driven by a [22% spike in critical\nburnout](https://leaddev.com/culture/engineering-burnout-rising-2025-layoffs-reshape-tech-industry)\nas they are forced to manage the massive influx of AI-generated\ncomplexity. Because we’ve stopped hiring juniors, there are few new\nseniors coming up to replace them. The ones who do make it through\nhaven’t experienced life before AI; they’ll be good, but they won’t\nnecessarily be great at writing maintainable code. In a market where\n[senior, production-ready\nengineers](https://beon.tech/blog/software-development-talent-shortage/)\nare already the primary bottleneck, I think we’ll need seasoned\ndevelopers with good taste more than ever.\n\nIn any market where there’s increased demand and reduced supply, prices go up. This happened for the COBOL programmers of yore during and shortly after Y2K. It’ll happen for senior developers soon, and just like our COBOL-wielding brethren before us, there will a shining window of opportunity. We just need to hunker down and survive the storm of AI-driven layoffs.\n\nThen, my friends, we’re going to make out like bandits.", "url": "https://wpnews.pro/news/we-re-going-to-make-out-like-bandits", "canonical_source": "https://www.rocketpoweredjetpants.com/2026/04/were-going-to-make-out-like-bandits/", "published_at": "2026-07-16 18:38:03+00:00", "updated_at": "2026-07-16 20:26:47.990053+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-ethics", "developer-tools", "ai-products"], "entities": ["Anthropic", "GitClear", "Stanford University", "BBC"], "alternates": {"html": "https://wpnews.pro/news/we-re-going-to-make-out-like-bandits", "markdown": "https://wpnews.pro/news/we-re-going-to-make-out-like-bandits.md", "text": "https://wpnews.pro/news/we-re-going-to-make-out-like-bandits.txt", "jsonld": "https://wpnews.pro/news/we-re-going-to-make-out-like-bandits.jsonld"}}