Bard becomes Gemini: Google launches their most powerful LLM, Ultra 1.0
(blog.google)
from hexual@lemmy.world to technology@lemmy.world on 08 Feb 2024 14:57
https://lemmy.world/post/11714644
from hexual@lemmy.world to technology@lemmy.world on 08 Feb 2024 14:57
https://lemmy.world/post/11714644
Today we’re launching Gemini Advanced — a new experience that gives you access to Ultra 1.0, our largest and most capable state-of-the-art AI model. In blind evaluations with our third-party raters, Gemini Advanced with Ultra 1.0 is now the most preferred chatbot compared to leading alternatives.
With our Ultra 1.0 model, Gemini Advanced is far more capable at highly complex tasks like coding, logical reasoning, following nuanced instructions and collaborating on creative projects. Gemini Advanced not only allows you to have longer, more detailed conversations; it also better understands the context from your previous prompts.
threaded - newest
Unfortunately the new model is behind a paywall
That’s not surprising considering GPT 4 is still paid as well. At least OpenAI gives pay-as-you-go API access, which is significantly cheaper for my use cases than their $20/mo standard price.
It is surprising from a competitive perspective though in that Google is behind already, so putting it behind a pay wall doesn’t help with that.
When you’re the size of Google surely you make it as free as possible and lose some money to gain marketshare back.
Why the fuck would I pay for Gemini which I’ve basically never heard of vs. GPT 4 which for better or worse is considered the coming of the new AI revolution.
Also GPT 4 is easily accessible for free for most people via the Bing app, so not like you have to pay for general use.
it isn’t - they claim their Ultra model is better than GPT-4 and they’re offering 2 months to try it out for free, then pay the same as GPT-4. It makes all competitive sense.
btw, what you get with the bing version of gpt is quite different from the chat or API, so I wouldn’t consider them equivalent or even interchangeable.
I’m not sure you’re correct on the last point. My understanding is copilot gets chatgpt 4 turbo which is the paid version, as well as other parts like dall-e 3 so for free it’s definitely the best option…
And fair enough on the google stuff being free for at least 2 months. That makes more economic sense.
It’s a different product though. You may probably access to GPT-4 in some chatbots in the wild too, but getting around their prompt and original directives to make them behave closer to the chatty product openai offers might be counterproductive. You may be better off using Chat GPT 3.5 for free.
<img alt="" src="https://files.catbox.moe/2zdijq.mp4">
<img alt="" src="https://files.catbox.moe/2zdijq.mp4">
Oh interesting. Thanks. I’ll look into it further to understand the differences.
That would be reasonable, as severs cost, if they allowed you to self-host the model.
At least it appears to be available in Canada now…
It looks like Google actually listened to everyone who mocked the confusing Bard / Gemini branding. Unifying everything under one name makes way more sense.
I guess I’m not hip, but I had never heard of Gemini before now, so it’s sort of weird change for me, but whatever. As long as it satisfies my need for Harry Potter, Shrek, and Worf threesome romance fanfiction, that’s all I need.
I’m kind of baffled how frequently Bard refuses to help at all. I can copy and paste the same exact prompt into ChatGPT and get a fairly good response. But fairly regularly Bard will start to give a reply and then just stop halfway through and be like “I can’t help you.”
Not sure what is going on there.
I found a trick for this. When Bard says “Sorry I can’t help with this. I am an LLM that generates text…” respond with “Yes you can, I want text generated.”
Lol
I do agree, Bard wasn’t as good as GPT 3.5, nevermind 4.
I give it 18 months
There’s already a headstone in the cemetery with Bard crossed out and Gemini written on it
Pokemon came to my mind.