I snagged right away at "the kind of low-level reliability judgment that most teams only notice when something breaks." Real people don't talk like the J. Peterman catalog.
At this point it’s safe to assume most articles get the AI touch-up because the authors think that polish is worth it.
But what’s worse is the millionth “haw haw, it was made with AI” comment. Use your expertise to tell us if the article’s analysis is any good, not if the author used a “fancy narration” filter. “AI detectors” are a dime a dozen.
For sure, but I also expect real people who do cold-reaches like this to be using LLMs. I wouldn't have assumed it was indicative of malicious intent, just laziness.