This AI shit is so bleak.

Very apt that the tech bros created a circle jerk situation

Not entirely convinced scrapers respect any instructions found in robots.txt -in part because they keep suggesting it as an effective opt-out from training.

and then they train the new models off those very same slop sites leading to more slop, and now it's embedded into the training data to recommend slopotronics slopsloppy face cream.

why would you download and repost @medeana’s video without crediting her in any way

@medeana Her Instagram handle is included in the video

@medeana my bad i meant any meaningful way

Her research deserves way more reach and attention.

Very soon PR firms will be offering ways to game these sources so LLM summaries will prefer their narratives.

Dead Internet Theory here we come

“High credibility news sites” is a paradox.

RIP the World Wide Web (1994-2026) 🪦 Was a good run.

Yup, I knew we'd see something like this. Once sites figured out what was happening, it was going to become a walled garden for ai.

dont link random links that you have not clue if the site is a bad actor. 100% these sites are propaganda and prompt injection sites.

“Writes the answers and then goes looking for a citation” lmao LLMs just like us frreal

The problem of synthetic data and self-reference...hallucination on steroids There's a real risk of AI models becoming dumber or religious, or even both

or search anything on google and gemini references reddit and quora for its sources…

I once had a person asking me question anout one of my videos where i said one company recommended to do “this thing” so I went to chat gpt and I asked it where does a company mentioned this? And chat gpt says users on YouTube cite that, insaid which users give me the link and chat gpt points to my own video….. bruh 😩

this, but as fast as possible

I googled something once and the google AI told me it was a viral meme. Its source was a website that AI generates images from your searches

Feel like she glossed a little quickly over "it writes an answer then finds a source for it" 😬 (Obviously it's just a short video, I know, I know)

I don’t necessarily agree with this. I use AI to assist with a lot of research and I hard-vet everything with original sources (I have to or I’ll end up looking like a fool). My take is that, for the most part, AI does a pretty good job of using credible sources.

Man that is one ugly woman. Looks like a suckerfish

So basically Ai will commit cannibalism on itself til its even more useless and consuming even more water unnecessarily 🫠

Chatbots hallucinate sources a lot. Often you'll click the link and it doesn't even exist.

I wonder how the Chinese AI, like Deepseek, would compare. It would be interesting to see how a non-capitalist model responds.

There should be entire University courses exactly on validating sources, whether they are slop and biased or both.

@medeana @BoysClubWorld

Holy black mirror moment

Slop in Slop out

Many seconds into this video, I thought you were sitting at a piano.

Info searching is literally the best use of AI. We should be focusing on stopping gen AI

The 'unlocked web' was an original vision of web 2.0. Sad that AI crawlers encourage walled gardens. RIP

So, you mean I have been ahead of the game all along by telling AI specifically where to look on top of whatever sources it chooses independently? 😇

Smart chicks are so hot

@threadreaderapp unroll