I find the Gemini app to be virtually useless. The guardrails are too strict, its writing style is poor, and it's not good at helping me re-write things while retaining my tone and intentions. Sometimes it completely alters the concept of what I'm trying to say. ChatGPT 4o is good, but it's a paid service. The free version, 4o mini, feels significantly less capable. Copilot's chat is okay, but I am simply BLOWN AWAY by Gemini Experimental 1206 on the AI Studio website.
Yes, it takes some time to generate replies, but the responses are leagues better. Its level of "understanding" seems vastly superior to the other major LLMs I'm used to, to the point that I rarely bother with them anymore. It's a much better collaborative tool, and its outputs feel less generic and more human, somehow.
We often see posts here, or in many other subreddits, about how AI has degraded in some way, or about moral dilemmas surrounding AI. So, I just wanted to take a moment to express how much I like this particular model. I'm excited to see where it goes from here, considering what we already have by late 2024.
[link] [comments]