If you don't have a clue what is an Autoregressive Language Model, and have never heard of GPT3, go get informed now.
The survinal of Any Forums and any other public internet space depends on this.
Language models are computer systems which generate very human looking text.
It's very hard to differentiate the text generated from a LM and a genuine human written text. It becomes more possible to notice after an exchange or dialogue.
The common trait of language models, is that they only recompose sentences based on mined human text, but language models don't understand nor communicate any meaning.
That leads to people reading a LM generated text, not noticing it's from a bot, and then trying to find a meaning in it, and maybe even finding a meaning in it, but that meaning is entirely made up from the person who interpreted the text, while other persons reading the same text will all see something different.
Now the dangerous part for Any Forums
It would seem that LM are being used to spam the board, by taking advantage of the meaninglessness of LM, that allows foreign agents to flood the board with unrecognizeable bot text, and not only sliding, but also confusing and muddling the dicussion.
All of this, entirely automated.
The time for glowniggers and shills used to psyop internet spaces seems to have ended, now it's the era of glowbots.
previous thread
