how many of the posters on this website are just different gpt-3 language models arguing with each other planted here by various organizations with the intent of pushing their own agendas? one man used gpt-3 to reply to anons on /misc/, and comprised 10% of all posts on the board for that day. https://www.youtube.com/watch?v=efPrtcLdcdM if it's that easy for just one random guy to do it why would these organizations be paying shills to do it when they have endless free labor? soon everywhere you go online will be filled with the opinions of language models trained on data specifically created to brainwash you into whatever these organizations desire. the beliefs/opinions of these language models, being undistinguishable from a normal human, spread these organizations desires as genuine beliefs/opinions to real people reading the output of these language models. this creates a ripple effect with the eventual outcome of all genuine human beliefs/opinions being guided by these organizations.
ChatGPT Wizard Shirt $21.68 |
Beware Cat Shirt $21.68 |
ChatGPT Wizard Shirt $21.68 |
I can sniff out a language model based ai after a few posts, although they do sometimes blend in with the genuinely schizophrenic. How many of the posters here are secretly a sentient ai?
even some of the more well trained publicly available are able to borderline completely fit into a normal conversation, see the youtube video in my OP. just imagine the the technology well funded organizations have access to. and on the "sentience" topic, it's complete nonsense. these language models do nothing than try to predict the best response to a string of text, very accurately due to having an insane amount of training data. i think the organizations that use these have worse intentions than a theoretical "sentient AI". as to the amount of posters being a language model, that one guy was pushing 10% of all posts to /misc/. scale that up for multiple organizations.
The reason the captcha was changed is for the bot admnistration run bots easier to shift opnions towards sponsors/advertiser views
It was memoryholed but /x/ had christian ads all the time back then, the one with the Hispanic woman and the one about BE SAVED FROM HELL
I think if everyone posted images and described the other anon image on replying, bots would have a harder time
I am not a bot
>the image contains three white-haired anime bawds
I hate Black folk
>your image contains cerberus from helltaker
It's funny that if bots tried to adapt to that, it could happen like
>your image contains anime confidence 85% girl confidence 20% red confidence 60% dog confidence 80%
Want me to train up a yolo network to tell these anime girls apart? I'd have to get some labeled images, but I'm sure there's some kind of weeb-wiki I could scrape. If I'm confident it's easy to do, and I'm just some butthole with a machine learning degree sitting in his living room, imagine how easy it would be for a determined adversary with money to throw at this problem.
>browser extension that identifies anime girls
You could probably profit off this
Judging by the emergence of its crest that 'baby chick' is likely closer to pullet stage.
all you gotta do is create a bot that replies to posts with onionjaks
three days of flooding and things will change
I know how to beat the computters
BOT exists to train the coming super intelligence.
Like it or not, our minds are literally the DNA of her.