Comment by phillipcarter
2 years ago
> Starting on December 13, developers and enterprise customers can access Gemini Pro via the Gemini API in Google AI Studio or Google Cloud Vertex AI.
Excited to give this a spin. There will be rough edges, yes, but it's always exciting to have new toys that do better (or worse) in various ways.
Indeed! Shame there's a lack of access to ultra for now, but good to have more things to access.
Also:
> Starting today, Bard will use a fine-tuned version of Gemini Pro for more advanced reasoning, planning, understanding and more. This is the biggest upgrade to Bard since it launched.
edit-
Edit 2 - forget the following, it's not available here but that's hidden on a support page, so I'm not able to test it at all.
Well that's fun. I asked bard about something that was in my emails, I wondered what it would say (since it no longer has access). It found something kind of relevant online about someone entirely different and said
> In fact, I'm going to contact her right now
OpenAI did well to let anyone try it with a login on a website.
Yep. That's their "moat", to go with The Discourse. For better or for worse, a bunch of us know how to use their models, where the models do well, where the models are a little rickety, etc. Google needs to build up that same community.
Gemini Pro is only GPT3.5 tier according to the benchmarks, so unless they make it extremely cheap I don't see much value in even playing around with it
I still think it's worth it. GPT-3.5 is extremely powerful, and it's what we use in production. GPT-4 is way overkill for our prompt and use case.
If it's similar, or even marginally better in any way, we'd consider switching over. Not because OpenAI is bad or anything (they're great, actually!) but because it's so easy to do that.