Lately, managing the AI has been giving me a stroke. The AI works perfectly fine for routine tasks, but when it hits a bottleneck, it's incredibly frustrating. The monitoring project was progressing well, but yesterday it got stuck on write permissions in a sandbox. It seems to be a bug, and that's when the weaknesses of ChatGPT and Gork became apparent. ChatGPT just couldn't get past the issue, couldn't find a workaround, and even automatically reported the bug to OpenAI, but it just wouldn't budge. Gork was even worse. It seemed to be having a serious mental breakdown, making ChatGPT search for non-existent menu items without any certainty. After being criticized by ChatGPT, it started apologizing, but then I kept giving ChatGPT a hard time. I was absolutely furious. As a result, the computer was on all night, and the two AIs were just endlessly looping. A lot of tokens were burned, but nothing was done. Then I had to wake up and see what happened. I tried to figure out a solution, getting ChatGPT to open a new Work instance to test for write permissions, and it passed the test, so we could continue. I've had enough! However, this also illustrates that without two AIs working together, there will inevitably be numerous potential problems. Recent collaboration has made the limitations of individual AI operations very apparent. One @Gate, trading more markets.