Comment by hunter-gatherer

Most comments here seem to discuss coding results. I know these are compared against industry benchmarks, but does anyone have experience using these with non CS related tasks? For example the other day I was brainstorming a kayak trip with both ChatGPT and Gemini 3.0. ChatGPT was off the rails. Trying to convince me the river flowed a different sirection than it does, and all sorts of weirdness. Gemini didn't provide information nearly as well as a human with experience, but it wasn't _useless_ information. The OpenAI model was a catasrophe at this. I'd be curious how the different models rate for the general audience, and if that plays into it at all.