Raw LLM Responses

I wasn't big into chatgpt either. I remember when I first used it, it was remarkably bad. I started using it again recently around the end of 2024. It's gotten much better but still needs work. Yesterday, I had it help me with making reverse proxies and using nextcloud, npm, setting up my jellyfin through reverse proxy and more. I knew the foundation of npm, jellyfin, but absolutely lost with next cloud. So with chatgpt helping, somewhat, I was able to get it running in about 2 hours. However, issues arose, and so troubleshooting began, and that lasted around 4 hours. Had issues connecting to mobile and permission issues that still exist. Chatgpt helped a lot but also made it very frustrating. Chatgpt talks ahead of itself instead of doing steps 1 by 1. Because errors generally arise during steps, and then it gets off on tangents when the errors arise. It's helpful but also frustratingly bad, too.