And Black Mirror was before that with the idea.

Mods already use bots to preemptively ban you based on what subreddits you have left comments on regardless of context.

The "social credit score" in China was blown out of proportion too. It probably changed since I last looked into it. But when that shit was all over American media it was literally the equivalent of a background check or credit score in the US. Work history, loan and credit history, etc.

Like, it's funny how we can have essentially the same exact thing but only critize China's version of it. Make it scary by adding "social" to the name. People ate that shit up.

Edit: to be clear I think both should be critized. Especially related to criminal history bias on race. But we spend more time talking about the Chinese version that doesn't actually effect anyone's lives in the US.

Unfortunately, you know it will. If it can be done, someone will do it.

Honestly the fediverse is already a bit of a privacy nightmare, and people refuse to take it seriously. Our version of this horror will include your comments and voting activity and likely a bunch of other telemetry which isn't public on reddit. Oh and it won't just be a single "official" AI agent, it will be dozens of them from every corporate or government which wants to build a user database and keep it around forever.

We could bake real privacy protection into our fediverse apps, but the admins and devs seem to have completely punted on this.

Omg, you are such a Monica.

Human personalities can be clustered into which member of the cast of Friends they most resemble.

Fortunately, Friends is so brilliantly written that they managed to capture the entire spectrum of human diversity.

(this is snark. I am totally a Chandler.)

it will be nothing more than grok and FB combined soon enough. the only reason reddit has completely astroturfed left leaning content, is because it allows right wingers to drive up engagement in those subs.

If I get it to call itself a clanker does that count?

"The user is a frequent poster to 'ebony' porn subreddits..."

I did, but the key difference is that it didn’t make an evaluation for you. It just gave you a collection of facts with some hard-coded things to look out for.

The new thing is way worse. It’s just skipping the facts part entirely.

Old Reddit is the only reason why I visit it sporadically. Combined with RES, I still have the same basic keybindings to browse I had multiple years ago. The new UI of Reddit is a waste of space.

https://lemmynsfw.com/ exists, so..

I noticed that there's virtually zero women asking for naked men. I kind of hope this means they are done with men, and will finally allow the population to collapse.

The same way that you can look at someone's post history, skim it, and more or less understand what their deal is.

Most users on reddit don't have this diverse set of interests or posts. Most of them post on the same subs and repeat the same types of posts... ad nauseam. Most people, online, aren't much more sophisticated in their language use or POV or anything... than a bot.

Sure, 10 years ago you'd look at someone's post history and it would be all over the place, and they'd be writing detailed and nuaced paragraphics in lots of detail... but that's not what reddit is like anymore, or most of the internet really. God I remember when libertarians actually had decent theoretical argumentation to defend their viewpoints... and now it's just memes and cliche phrases repeated over and over.

Pretty easy for any LLM to parse a person who just incessantly posts about the same stuff over and over. And if it came across someone doing long-form explanatory stuff... it would just ignore that data since that doesn't fit the task LLM is designed to process.

Narrator : "they were all shit heads"

I personally wouldn't go that far. But, yeah, we need to de-tech-ify a bit. And I am all for steal the control of the internet back from giant corporations.

reddit already gave it to GOOGLEs ai, and Open AI as well.

It looks like a thing only mods can see

Jeez, at least make the fascists wait for the language model to spit out ideas for them to steal.

I edited all my comments to be generic observations, so the only really useful metrics would be where and what I commented on

Agreed that it doesn't say anything because 'strong' and 'provocative' mean different things in different contexts such as subs with very different users.

A fediverse example is saying 'AI sucks' isn't provocative in c/Fuck_AI but is on the dbzer0 instance if provocative is based on votes or replies. Strong can just mean direct and clear. Sure, they might comment a lot about Palestine but the rest doesn't indicate whether they support Palestine or Israel because there is no context for what subs they post in.

Just worthless word salad.

What if the summary is wrong or misleading?