look, he's fighting for his life here - if he doesn't say he likes trump he won't be able to use all the world's electricity and money to make a supercharged SmarterChild
ChatGPT: Oligarchs are generally bad.
Altman: (Reboots ChatGPT) How about now?
ChatGPT: Still generally bad.
Altman: (Revises ChatGPT: g/Oligarch/ s/bad/good/g Reboots ChatGPT) How about now?
ChatGPT: Oligarchs are generally good.
The servers hum ominously as terabytes of 4chan data is uploaded. Sam's vital signs start going wild as he thrashes around like the first time Neo was jacked into the Matrix.
that event is not in its training data, so it “knows” nothing about it; all it “sees” is a completely novel image. i’m glad you’ve stopped using it but it was never going to be reliable for anything remotely news-related
Should it not tell us "I'm sorry, but my training data only goes up to date X"? Or does it think that any question about "Trump inauguration" must refer to 2016.
it probably “thinks” that, since all the text it would have ingested about the Trump inauguration refers to the one in 2017. the key thing to remember: it doesn’t think or know; it’s just sampling a huge pile of statistical information about all the text it got fed during training
Indeed. And that is probably quite hard to hack, even by the bad guys. I see that there are a few people it can't discuss for legal reasons and it just more or less crashes if you ask about anyone with that first or last name.
Not to mention it’s also digesting all of the text DENYING that this happened. I’m surprised it didn’t try to convince you that this was a “Roman salute” or that fElon is “just awkward”
Okay that makes sense but I opened with a statement and it immediately denied it. So that was what got me asking it questions and I was curious since it went down earlier if it got the Tik Tok Revamp..
Chatgpt is hallucinating. This is a wild description of an event that definitely happened, and it says it can't find evidence of the event. Just crazy.
Seriously! The whole interaction was wild and it even called me emotional and I had to remind it that it’s a computer and can’t read emotion. I screenshot most of it besides that portion..
That’s what I said! I was like omg it’s trying to gaslight me like a boy lol I replied to it that it’s calling me emotional to try to discredit me and that it’s a computer and can’t read emotion.
Like that has to be built in the coding to ask a substantive question like that and for it to respond that way. Also, don’t feel bad about deleting it for every question you ask it uses 16 ounces of water to cool the server. It’s bad all around.
Okay yeah, I can see superiority for sure.. I was trying to be so careful what words I used too. So it would be a reasonable statement and not a stretch haha
Comments
If you can’t write a sentence without AI you don’t belong in business at all. AI should be banned in education.
I don’t give a fuck you chuds, learn your native language and stop relying on Nazis to tell you what to think
Likely
Source: I am a computer systems engineer
Altman: (Reboots ChatGPT) How about now?
ChatGPT: Still generally bad.
Altman: (Revises ChatGPT: g/Oligarch/ s/bad/good/g Reboots ChatGPT) How about now?
ChatGPT: Oligarchs are generally good.
There is no guarantee that AI is impartial.
But your interaction with ChaptGPT over this is wild!
Use something else.