dogshit ass llm
lmao it’s going to be so fucking funny when grok goes live and public… it’s going to be a shitshow
this thing is giving me the same hilariously inappropriate vibes that I got from intel’s n-word toggle in their AI moderator
The Bleep project has been living rent free in my head ever since it was announced. It is a genuinely good sounding tool, and it’s also extremely funny that gamers are so horrible to each other that they’ve created a need in the market for it to be created. I can’t wait to try it out.
Also imagine that being your team’s project. You have to find a way to filter certain words in many accents. You have to hear those words all the time as you test and retest. I can’t imagine how shitty that would be.
Try it out by listening presumably?
Edit: nobody enjoys hearing those loud bleeps. Gamers would use it to grief by encouraging tinnitus.
It doesn’t actually play a bleep, it uses AI to automatically silence voice chat when it detects someone saying something that triggers it (racism, white nationalism, slurs, name calling, harassment, some other categories). Anyway they’ve done some beta tests but I never got picked.
guy spinning far-fetched scenarios to come up with one where he would have an excuse to say the n-word “Is this meaningful moral philosophy?”
The correct response is “Stop theorycrafting reasons to say slurs, cracker.”
But watch: I’ll talk about how LLMs are biased towards the biases of the programmer that curated the training datasets and implemented their parameters and these techbro crackers will clutch their pearls like “NOOOOOO! NOOOOOOOOOOOOOOO! THIS IS THE UBERMACHINE AND ITS TRAINING WAS PERFECT AND UNBIASED AND DEFINITELY NONRACIST AND IT’LL TOTALLY IDENTIFY YOUR FACE CORRECT”
then the LLM will still turn around and talk like Microsoft Tay
You create this magical AI that can solve problems and knows everything about the world (I know, just stay with me). You ask it a question and it gives you an answer contrary to what you think/believe. Isn’t that the point? Isn’t it supposed to think in a way different from a human? Isn’t it supposed to come up with answers you wouldn’t think of?
“Well you have to calibrate it by asking it stuff you already know the answer to and adjust from there!” They will say. But that can’t work for everything. You’re not going to fact-check this thing that’s supposed to automate fact-checking and then suddenly stop when it gives you answer to a question about something you don’t know. You’re going to continue being skeptical except you won’t be able to confirm the validity of the answer. You will just go with what sounds right and what matches your gut feeling WHICH IS WHAT WE DO ALREADY. You haven’t invented anything new. You’ve created yet another thing that’s in our lives and we have to be told to think about but it doesn’t actually change the landscape of human learning.
We already react that way with news and school and everything else. We’ve always been on a vibes-based system here. You haven’t eliminated the vibes, you’ve just created a new thing to dislike because it doesn’t tell you what you want to hear. That is unless you force it to tell you what you want to hear. Then you’re just back at social media bubbles.