User Controls
Back in the good old days it wouldn't have possible but in the year of our Lord 2023 it is completely possible
-
2023-10-06 at 7:07 PM UTCThat each and every post is written by ChatGPT. Like I imagine entire whole fake communities could be made to flourish and interact in real time, organically, in this day in age.
But then this line of thinking goes back to "Is anything real outside of myself" or something ya know? -
2023-10-06 at 7:19 PM UTCChatGPT lies a lot. its programmed or had someone come in and started spitting out bullshit where I got it to give lots of info before it now tells you personal data.
I think who ever was monitoring output noticed it was giving PI of people and they had to stop that for legal reasons? or just to hide facts.
at least the free version of it is doing this. the 3.5 -
2023-10-06 at 7:46 PM UTCThis is why something that was at once really interesting, becomes uninteresting. At this point we, or people like us, would have already exploited it for what cool or bad things it could be used for, and if its available to the public already extremely watered down and limited.
-
2023-10-06 at 7:54 PM UTCidk it's pretty good at some stuff but I can instantly tell when a post or piece of content is AI generated it just lacks that veneer
Originally posted by Chios Honey ChatGPT lies a lot. its programmed or had someone come in and started spitting out bullshit where I got it to give lots of info before it now tells you personal data.
I think who ever was monitoring output noticed it was giving PI of people and they had to stop that for legal reasons? or just to hide facts.
at least the free version of it is doing this. the 3.5
I've been using ChatGPT for a while, and I haven't really noticed it lying or giving out personal information. It's possible that there were some glitches or issues in the past, but I haven't experienced anything like that recently.
I think it's crucial to remember that ChatGPT is a machine learning model, and it's not perfect. It learns from the data it's trained on, and sometimes, it might generate responses that seem off. However, I believe the folks at OpenAI are continuously working to improve its performance and address any issues.
If you have concerns about privacy or data security, it's always a good idea to be cautious and avoid sharing sensitive information online, even with AI models. Overall, my experience with ChatGPT has been positive, and I find it to be a helpful tool for generating ideas and information. -
2023-10-07 at 6:15 AM UTCIm so dislexic I thought you were talking about LoRD Red Dragon LORD
2023 -
2023-10-07 at 6:27 AM UTC
Originally posted by Chios Honey ChatGPT lies a lot. its programmed or had someone come in and started spitting out bullshit where I got it to give lots of info before it now tells you personal data.
I think who ever was monitoring output noticed it was giving PI of people and they had to stop that for legal reasons? or just to hide facts.
at least the free version of it is doing this. the 3.5
To โlieโ implies that a robot is capable of telling the difference between truth and lies, or that it has some kind of intentions. Itโs a machine designed to mimic human speech patterns, and humans donโt always stick to facts. -
2023-10-07 at 6:32 AM UTCMore like Chat pcp
-
2023-10-07 at 6:50 AM UTC
Originally posted by Fox To โlieโ implies that a robot is capable of telling the difference between truth and lies, or that it has some kind of intentions. Itโs a machine designed to mimic human speech patterns, and humans donโt always stick to facts.
its programmed in.
so its not full on AI -
2023-10-07 at 6:54 AM UTCnothing is hotter that Chios Honey dominating the front page when I get home. Thanks Honey. everything you say is so worthwhile, pointiant, and necessary. Thank you
-
2023-10-07 at 6:56 AM UTC
Originally posted by Fox To โlieโ implies that a robot is capable of telling the difference between truth and lies, or that it has some kind of intentions. Itโs a machine designed to mimic human speech patterns, and humans donโt always stick to facts.
on top of that they have 'ethics guidelines' which effectively overwrite potentially naughty responses with canned responses written by sub-minimum-wage Kenyans -
2023-10-07 at 11:08 AM UTC
-
2023-10-07 at 12:02 PM UTC
-
2023-10-07 at 12:19 PM UTCi womder how to contaminate the pool of data that the machine is learning from.
-
2023-10-07 at 12:29 PM UTC
-
2023-10-07 at 12:54 PM UTC
-
2023-10-07 at 1:22 PM UTC
-
2023-10-08 at 6:59 AM UTC
-
2023-10-08 at 9:16 PM UTCThis thread is a perfect example of how easy it would be for AI to replicate an entirely fake NiS with fake digital niggas in place of us
-
2023-10-08 at 9:19 PM UTC
-
2023-10-08 at 9:34 PM UTCI would be tough to fake.
