If you think about copilot etc...that would produce the same pattern. Fragments being obviously matching existing work.
Not sure I'd go quite as far as calling an author a plagiarist on the basis of specific short fragments matching. It would need bit more of an overarching say plotline matching or similar.
Who here hasn't copied some code off stack overflow?
How do you know it’s correct for your need without actually reading and internalizing it? It might even be correct…but not compute what you thought it does.
Plugging code from SO into your program is like assuming that if your program compiles it must be correct.
I wish I could add a sarcasm marker but I see chunks of dumb code inserted in public repos, NPM, etc all the time.
I can only hope that my bank’s phone app has been subject to reasonable code review.
This is a clue that the current fashion of crying ChatGPT at every instance of bad writing is quite silly. Plagiarism, and bad writing in general, predate LLMs by centuries.
You could, I suppose, add a CC BY SA acknowledgement to both StackOverflow, and the person who posted the code for every snippet of code you copy from StackOverflow in license notifications in your own project. You must acknowledge both. But you know you won't.
Instead, you will rely on the goodwill of every corporate overlord who will ever own StackOverflow, or inherit its copyright assignments for one Mickey Mouse copyright lifetime. Which seems unwise given that the current corporate overlords don't seem to be trustworthy. And seems particularly unwise now that the Millennial internet is passing away.
There was some discussion about allowing authors to post code under a dual-use license when StackoverFlow first converted their terms of use to require CC BY SA licensing. But whenever I add dual-use license notifications to my StackOverflow posts, StackOverflow moderators keep editing it out without my permission!
If the answer links to code on another site, then that code isn't covered by the StackOverflow terms of use.