December 29, 2025
Bring popcorn: devs vs bots
The Future of Software Development Is Software Developers
Veteran says humans still needed; juniors fear AI wiping jobs
TLDR: A seasoned coder argues the toughest part of software is turning messy ideas into precise logic, so humans aren’t going anywhere. Comments erupt: seniors agree, juniors fear AI erasing entry-level jobs, pragmatists add AI as a safety layer, and skeptics warn bad AI code could poison future tools.
A 43‑year programming veteran just dropped a spicy take: despite decades of “end of programmers” hype—from drag‑and‑drop tools to today’s AI chatbots—the hard part of software is still turning messy human ideas into precise logic. Cue the comments section going full meltdown. One camp, led by simonw, is nodding so hard they might sprain a neck: the real skill isn’t typing code, it’s thinking clearly. Another camp, like mohsen1, is panic‑tweeting: “Coding AIs design software better than me,” and worrying this time really is different. Then the generational drama hit—aizk claims seniors are out of touch and new grads are getting steamrolled. Popcorn, please.
Meanwhile, the pragmatists pull out the Swiss cheese meme: d_silin says treat AI like another safety layer—imperfect, but helpful. The doomers go big: berdon warns AI‑generated slop code could poison future training data, making everything worse. There are jokes about “programming in English” never happening and memes about the rise of the prompt‑whisperer class, plus nostalgic eye‑rolls at Visual Basic saving the world by “tracking the killer’s IP.” The vibe? Split. Some say AI is a tool, not a replacement. Others say junior jobs are the first on the chopping block. Everyone agrees: knowing what to ask for is still the boss fight.
Key Points
- •The article surveys repeated historical claims that new tools would eliminate programmers, from WYSIWYG editors to 4GL/5GLs and compilers.
- •It argues past cycles increased both the volume of software and the number of programmers, not reduced them.
- •LLMs are presented as different in scale and visibility, but are said to often slow teams and reduce reliability and maintainability unless process bottlenecks are addressed.
- •The central difficulty in programming is converting ambiguous human intent into precise computational logic, not merely writing code.
- •Edsger Dijkstra is cited to support that programming will not be done in natural languages due to inherent ambiguity, implying continued need for skilled programmers.