July 25, 2026
Cheap code, expensive drama
Engineering management after the cost of code collapsed
Bosses rethink coding rules as commenters roast the hype and warn of an AI mess
TLDR: The article argues AI has made writing software cheaper, so managers should question old rules but avoid betting people’s jobs on hype. Commenters were split between optimism and eye-rolling, with some saying AI opens new doors and others warning it just creates faster, messier code.
A senior engineering manager dropped a big claim: the price of writing software has crashed thanks to AI writing tools, so a lot of the old management advice may need a rewrite. But instead of cheering, the comment section instantly turned into a spicy reality check. The author’s main point was actually pretty cautious — don’t assume teams are magically faster, don’t fire people based on vibes, and don’t trust old scoreboards like counting tickets or code changes when machines can now pump those out cheaply. In plain English: just because more words of code appear doesn’t mean better products appear.
And readers had thoughts. One camp said this is just the latest step in software’s long history of change, with AI helping people tackle problems that used to be out of reach. Another camp was far less impressed, basically saying: congratulations, now we can create bad code at record speed. That sparked the hottest tension in the thread — is AI lowering costs, or just inflating future cleanup bills?
Then came the comedy. One commenter sneered that the real thing getting cheaper was blog posts, roasting the article’s polished management-speak like it was a corporate TED Talk audition. Another zoomed in on the line saying “Gemini 4 helped with the editing” and immediately went detective mode, joking that maybe the author meant a different model and the AI was too polite to correct him. So yes, the article was about management. But the comments? They were about trust, hype, debt, and whether we’re all just automating the mess.
Key Points
- •The article says the strongest supported claim is that the cost of producing plausible code has fallen significantly due to LLMs.
- •key engineering-management practices should be evaluated based on the assumptions they rely on, not on whether they seem old or modern.
- •The author says claims that AI has already made organizations dramatically faster, eliminated the need for review and documentation, or justified major headcount cuts are not established facts.
- •Reported productivity gains are described as more evident in greenfield, boilerplate, and unfamiliar work than in deep work on well-understood systems.
- •The article argues that proxy metrics such as velocity, pull request counts, ticket closures, acceptance rates, and prompt counts can become misleading when code generation becomes cheap.