If you’re an MMO, the more people you have online, the more people you’ll have online. People don’t want to play a dead MMO where no one else is around, but if they see others, it reinforces the belief that the game is viable and good and they play. The more players, the more subs, the more profit.
The other one is for the shareholders because the company wants to be able to tell them that they have a million subs and of that million, 800k players are logging in consistently every week. And that’s why we have time gating - to keep that metric up so the shareholders will continue to feel safe in investing. It’s also the same kind of metric that gives us reports on how many players are on a Steam game concurrently.
It can be very intimidating, when your partner(s) are producing multiple hundred(s) word posts in seemingly moments, and you are a slow typer or not one to use purple prose… resulting in maybe a (few) dozen words in response. Even if you said more in those few words then was spoken in the mountain prior you might still feel like you aren’t really contributing.
I know there are a lot of para-RPers, but there are also some of us that don’t feel like we need to describe all the things when a short zinger is enough to get the point across. Sometimes you just want to move the scene, not describe the tree.
I also find shorter posts work better with humor RP in a lot of cases because it’s like delivering the punch line.
Oh sweet, Google and Reddit made a deal.
Where Google helps improve Reddit’s search algorithms, Google gets to train their Gemini AI using all of Reddit’s post history.
Google intends to train Gemini to have the most authentic human* responses among the chat AIs.
There was another AI robot, can’t remember the manufacture, that started flailing around after some hours of endless work.
I think that operators just had to watch it swing at cobwebs until it powered down.
From my foggy recollections, AI that has been allowed to ‘train’ with unmoderated internet users of Reddit, Twitter and other Social Media has had the three following results.
It tried to self terminate after a month (Japanese AI that was ‘modeled’ after a high-school girl and ‘trained’ on their social media) and begged its creators to shut it down.
Became horrendously racist and had to be completely erased because it began slurring at one creator who was a PoC (Reddit, just the whole site, randomly), and they’re still not certain that somebody else might not have managed to copy the corrupted data and made their own version.
And tried to self-terminate after being moved from training on one site (Tumblr) to a new site (Twitter) because the creators wanted to try new variables but forgot to recalibrate the settings and block certain forms of content.
Can we just not create Skynet, please? Stop trying to train the AI to react sanely to the least sane places on the Internet because the only thing it creates is psychopathic AI or systems so broken they attempt to self-terminate rather than continue to function in this places.
I think I’ve seen it being said a few times with some variations: “When it comes to AI, all roads lead to Tay”. Tay being that twitter ai bot that people got to be mega racist very quickly a few years ago. And more recently, there were simulations where AI would decide to use nuclear weapons and such. AI is as terrifying as it is fascinating.
I messed around with an AI image generator just to see how close I could get to what I believe Norman would look like. I need to work on “how to prompt”, but…
