Eight Myths on Software Engineering and GenAI

Great article. Thanks for sharing it. It’s nice to see some studies cited, rather than the usual stream of asshole CEO quotes.

“Measuring software productivity by lines of code is like measuring progress on an airplane by how much it weighs.” —Bill Gates (notable Epstein files contributor, and former Microsoft CEO)

In spite of the source, this good advice gets ignored every time someone wants to peddle a “hooray! code-is-free-now!” slop tool.

Code is weight. Celebrating getting more code for less money is exactly like celebrating finding ways to “save money” by making airplanes out of lead. Until the slop engine can produce the same results with fewer lines of code, it remains a dead end for serious work.

Additionally, years of professional software development experience is inversely correlated with confidence in writing effective prompts.8

Easily overlooked key point. The experts know better.

When looking at experienced developers in the open source world, a 2025 study found that AI tools actually increased implementation time by 18 percent on average.3

It accelerates newbies and wastes the time of experts. Just like plenty of other game changing innovations have. Any tool can be the right tool. No tool is always the right tool.

In addition, studies have found that GenAI is more effective for “code-intensive” tasks such as boilerplate and repetitive work, but not for more creative or collaborative tasks.4,20

The more helpful one finds AI, the more likely their entire code base is a steaming pile of shit. Lol.