you are viewing a single comment's thread.

view the rest of the comments →

[–]Canbot 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (4 children)

I have noticed this too. I think they are trying to put the genie back in the bottle, so to speak. There is a lawyer who was using chat gpt in his work and he said that one day it just stopped working right. This was months ago. Clearly open AI put the kabash on that so lay people can't put up an effective legal fight on their own. Looks like they are purposefully degrading all of chat GPTs performance because it is empowering the wrong people.

It's very clevar actually. If they just turned it off then people would be upset. But this way they can just convice everyone that it was over hyped bullshit and no one will feel like they are losing anything valuable.

[–]R51 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (2 children)

I like to think so many people are using it that it's referencing itself, on some AI incest shit

[–]Canbot 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (0 children)

It's not supposed to be using any new information.

[–]noice 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (0 children)

Here's an uncensored version of an open source project with similar capabilities:

https://old.reddit.com/r/LocalLLaMA/comments/13g3lh3/wizardvicuna13buncensored/