From the post: "The problem isn’t just laziness. It’s degradation ... Eventually, you’re living in a cathedral of technical debt, and every user pays ... 50 million users have an extra 3 seconds of unnecessary lag in a day because you wanted to hit tab rather than write code?"
https://deplet.ing/the-copilot-delusion/

I'm seeing output from Midjourney and ChatGPT that may be input examples, phrases that seem likely to have appeared in the input, or generated images very close to known input examples. I wonder if this is due to the pressure teams are under to improve the output. ChatGPT seems to do this, most obviously on recipes or HR memos, but also on fiction. And others pointed out examples from Midjourney that seems very close to known photos or art, eg, this one (which I generated).