@trevdev There is a simple test that can definitely prove that AI reviews are bullshit, but none of these AI boosters are even trying to check that.
1. get some code, any code
2. send it to AI, ask for review/improvements
3. get the resulting "improved" code
4. clear your session, start new chat, give it the "improved" code, and ask for review/improvement again
5. repeat 2-4 until you get tired
It *always* finds something to "improve". But overall it just makes the code more and more complicated for no reason. So where should we stop when we ask AI for review? Should we just take the first version? Or run it a couple of times? When is the code "best"? Who in their right mind would just accept whatever the AI said after the first request? And who in their right mind would run it through AI reviews until it says everything is perfect, which basically comes at random? These people are not in their right mind.