General Discussion

That's what I'd rather see in an excerpt box, along with a link.

These plagiarism machines work, in effect, by scraping words and images from the internet and other sources, compressing them down and mashing them together in a lossy manner until the sources are entirely obfuscated. Then according to some user prompt they expand this compressed information, stitching together all the rips and tears and filling in the holes in some plausible manner before spitting out the results. (This process uses an unconscionable amount of electricity.)

Occasionally the results go entirely sideways in a Mad Lib fashion, and those garbage results are called "hallucinations" in an effort to disguise what's really going on inside the box.

There's an easy to find video presentation by Dave Plummer, a retired Microsoft engineer, about using AI to recreate the Notepad text processing program as it was before it was re-imagined (a polite way to say ruined) in Widows 11.

It looks like magic if you don't know what's going on. (And it begs the question, why not simply use the old version of Notepad itself?)

It's not magic at all when you realize what this vibe-coding software is actually doing. It's been "trained" (another deceptive word choice) on actual text processing software, maybe even the code for the older version of Notepad itself.

It's like a kid who, instead of doing any actual research for a term paper, rewrites an encyclopedia entry "in their own words," puts a few sources that they haven't even looked at in the bibliography, and hopes the teacher won't notice.

Of course I date myself having grown up in a time where the internet wasn't open to the general public and the Encyclopedia Britannica usually had an entry about whatever topic you were assigned to write about.

This is my term paper about Liechtenstein...

These days Wikipedia serves the same purpose and the plagiarism machines which are not intelligent scrape that too.