Does anyone have experience with how Bard compares to ChatGPT-4? If ChatGPT-4 is on the verge of usefulness for a layperson, is it worth spending time with Bard?
It's drastically improved, although I haven't done a lot of comparison yet. Some of my go-to prompts yielded results that are (in my opinion) competitive with GPT-4.
Edit: Just ran into a situation where Bard uses `typeof something === "array"` in some generated code, and although it acknowledges that typeof in javascript never returns "array", it doesn't see it as a problem in the code that it provided to me.
The funniest to me is that one of the examples they suggest you try is asking questions about google/jax repo, which kinda suggest that Bart can index source code from GitHub. Well, it fails even at listing directory structure and completely hallucinates all following answers!
I'm not sure why Google would bother releasing Bart in this state. It's made me lose respect for their AI competency.
Maybe I'm prompting differently than other people, but I find them all give satisfactory answers.
Bard does seem to be getting confused during conversations and has trouble following, but improvements are noticeable from a few weeks ago. For some prompts I'd say its similar to GPT-4, and it has much newer information (appears to be semi-connected to the internet) which is really handy, but it definitely needs some work still.
Example of "semi-connected to the internet": I asked Bard the about the current weather in one place and it gave me accurate current info, but for somewhere else it gave me "as of 10AM EST".