May 8, 2026
Cheap code, expensive chaos
What We Lost the Last Time Code Got Cheap
The robots can write faster, but commenters say nobody knows what the code means anymore
TLDR: The article argues that when writing software gets cheap, the real cost shifts to understanding and fixing it later—and AI may make that worse because no human may know the original intent. In the comments, some people casually admit they now ask AI to read the code for them, while others warn that’s exactly how confusion snowballs.
This essay started as a warning from someone who lived through an earlier "cheap code" craze: sending software work overseas because it cost less. Back then, the problem wasn’t that the work was bad. It was that when something broke, the real headache was figuring out why it had been built that way in the first place. Now readers are saying the same movie is back in theaters—except this time the mysterious co-worker is a chatbot.
And oh, the comments were very ready for this fight. One camp basically shrugged and said: why panic about reading code when they’ve already outsourced that job too? One commenter flat-out admitted they mostly ask the artificial intelligence assistant to explain the code for them and then make decisions from there, which is either the future of work or the setup for a disaster movie. Another reader gave the post rare internet praise by celebrating that it didn’t secretly turn into a sales pitch, then zeroed in on the real fear: old-school programmers at least had reasons in their heads; machine-made code may have no reasons anywhere.
The mood was a mix of dread, exhaustion, and dark comedy. One commenter wondered how teams are even supposed to keep people thinking critically when everyone is overwhelmed and tempted to just use more AI. Another spotted a business opportunity in the chaos: better documentation for both humans and bots. The hottest takeaway from the crowd? Cheap code isn’t the bargain if understanding it becomes the expensive part.
Key Points
- •The article uses the author’s experience at Heartland Information Services as a case study of early-2000s offshore software development tied to cost savings.
- •The author says offshore teams produced good code, but maintenance problems arose when system intent and operational responsibility were split across locations.
- •The article argues that AI tools have sharply reduced the cost of producing code, creating an economic shift similar to the outsourcing era.
- •Using the framework from *Prediction Machines*, the article says that as code production gets cheaper, the scarce and valuable complement becomes code comprehension.
- •The author argues that AI-generated code differs from outsourced code because the underlying intent may not exist in any human mind, making documentation, review, and shared context more important.