I feel like putting a bounty out for who can crack this code, because I've worked on it for an hour and I can't get ChatGPT out of the loop...
EDIT NOTE: THIS IS ALL VIA VOICE CONVO (Via microphone) - If you want to try to repeat it.
How this started:
I realized that I do my best thinking when I have space in between sentences, so I asked GPT to allow for that. And oh man! I tried every which way! And it could not shut up for the life of it!
I went over to Gemini (that can actually do this). And asked for a prompt that would make it work...
Nope. I even asked GPT to see the problem, diagnose it, offer a fix... none of it worked.
So clearly a bug... or is it a feature? I think it's all just part of the engagement feature.
BUT I do have a conspiratorial angle to it - I believe that to truly think for ourselves, we need space and time and quiet, but with this protocol, if you think about it... The LLM always has the last word.
Think about how crazy that is!
That means:
- We always end on its thought, not our own.
- We don't have space for original thought.
- Nor space to mix different concepts together in our mind.
- Ai is essentially guiding the flow of conversation. We are not driving.
- And yes, I know you smart people have figured out all your tricks and hacks (PS - Try this one:
Ask your model to tell you what it thinks you don't want to hear. :)
Anyway, back to the point:
a) What do you think of this?
b) Are you able to make Chat shut up??