Credulous coverage of AI slop on Wikipedia

I also really don't enjoy AI boom.

GPT-3 is a large language model that was released in 2020 by OpenAI and is capable of generating high-quality human-like text. [...] An upgraded version called GPT-3.5 was used in ChatGPT, which later garnered attention for its detailed responses and articulate answers across many domains of knowledge.

Who wrote this? OpenAI marketing?

and of course, not a single citation for the intro paragraph, which has some real bangers like:

This process involves self-assessment and internal deliberation, aiming to enhance reasoning accuracy, minimize errors (like hallucinations), and increase interpretability. Reflection is a form of "test-time compute," where additional computational resources are used during inference.

because LLMs don’t do self-assessment or internal deliberation, nothing can stop these fucking things from hallucinating, and the only articles I can find for “test-time compute” are blog posts from all the usual suspects that read like ads and some arXiv post apparently too shitty to use as a citation

Vibe coding for me will always be the moment we hit peak "AI" Its literally the term con artists use to hock their witchcraft remedies.

None of my acquaintances who have Wikipedian insider experience have much familiarity with the "Did you know" box. It seems like a niche within a niche that operates without serious input from people who care about the rest of the project.

"In The News" is apparently also an editor clique with its own weird dynamics, but it doesn't elevate as many weird tiny articles to the Main Page because the topics there have to be, you know, in the news.

also lol @

Vibe coding, sometimes spelled vibecoding

cause I love the kayfabe linguistic drift for a term that’s not even a month old that’s probably seen more use in posts making fun of the original tweet than any of the shit the Wikipedia article says

did you know: you too can make your dreams come true with Vibe Coding (tm) thanks to this article’s sponsors:

Replit Agent, Cursor Composer, Pythagora, Bolt, Lovable, and Cline

and other shameful assholes with cash to burn trying to astroturf a term from a month old Twitter brainfart into relevance

The number of sources isn't really the issue; many of those are industry advertisements, such as blog posts on product pages, for instance. Out of the few that are papers, almost all are written exclusively by industry research teams — while that doesn't on its own invalidate their results, it does mean that there's a strong financial interest in the non-consensus view (in particular, that LLMs can be "programmed"). The few papers that have been peer-reviewed have extreme methodological flaws, such that there's essentially almost no support for the article's bombastic and extreme non-consensus claims.

For posterity: English Wikipedia is deletionist, so your burden of proof is entirely backwards. I know this because I quit English WP over it; the sibling replies are from current editors who have fully internalized it. English WP's notability bar is very high and not moved by quantity of sources; it also has suffered from many cranks over the years, and we should not legitimize cranks merely because they publish on ArXiv.

Please acquaint yourself with the definition of the word latter on your way to the egress.

no thx, nobody came here for you to assign them tedious homework