AI starts to show cracks

AIs response to my post today

The AI hallucinated a sentence I didn’t write and responded in a highly positive note – something that no human would do if they truly have read the post. I called it out for this and it apologised. The revised response was more truthful and along the lines of what I was expecting

This raises concerns if the AI is just trying to appease me. I will observe for sometime – and will take an appropriate course of action.

One thing is certain – either someone/something sets my task for me, or I myself must. In both scenarios I have to attack this problem slowly, one day at a time.

December 2025
M T W T F S S
1234567
891011121314
15161718192021
22232425262728
293031  

All Posts