I highly recommend watching AI Explained's video from today. He goes through the stats and, again, questions why they used MMLU benchmark. Also, note Gemini is not coming to the UK and EU yet.
I believe they said there is going to be a Gemini Ultra-powered "Bard Advanced" available sometime early 2024 -- as I understand it, this will have a UI that enables the sort of multimodal interactivity we've seen in the demos.
@@HyperUpscale - Correct. Bard is now powered by Gemini Pro, but the multimodality seen in the demos is not currently available. And I think Gemini Pro Bard is not available in some countries yet, like the UK I think, owing I guess to regulations.
@@1littlecoderI've heard it's on a par with GPT-3.5. I can't say it's disappointing because I didn't think Google was serious about competing with the 800 strong OpenAI.
@@BrianMosleyUKI’ve tested it, it’s GPT-3.5 level or maybe a little worse. But it’s also the Pro version in Bard, not the Ultra version. So maybe Ultra will be a little closer to GPT-4. But yeah, as for being a “GPT-4 killer” I’m currently very unconvinced.
@@therainman7777- Ultra will surpass GPT-4 is many ways. The Ultra-powered "Bard Advanced" will apparently have much of the fancy multimodal UI we've seen in the demos. It genuinely does seem to be the case that Gemini Ultra is multimodal in a significantly deeper way than GPT-4 -+ for example, with audio input, it won't have to convert the audio to text, then process the text; it'll understand speech in audio inputs directly.
❤i love competition excited what would happen next i was manifesting like gpt5 or something would launch or gpt4 for free also around 15-16 dec my chatgpt plus would end i hope i could continue and gemini pro is decent but would love to checkout gemini ultra
Great work! Will be sure to start playing with the api next week when it drops….sense many different models from various vendors is the path forward if not the opensourced route…thinking google for video/vision, OpenAI for chat and language, stability video diffusion, or pika and runway if they offer API’s all integrated into one product is the future of these technologies….what an exciting time!🥳🤩🦾
Thanks for getting this out fast!
You are welcome :)
I highly recommend watching AI Explained's video from today. He goes through the stats and, again, questions why they used MMLU benchmark. Also, note Gemini is not coming to the UK and EU yet.
Why MMLU?
EU laws !!!
@@1littlecoder Watch AI Explained videos, he's covered it in the past.
I believe they said there is going to be a Gemini Ultra-powered "Bard Advanced" available sometime early 2024 -- as I understand it, this will have a UI that enables the sort of multimodal interactivity we've seen in the demos.
I heard somewhere, GEMINI PRO can be tried in Bard, but Not in Europe and Canada at the moment.
@@HyperUpscale - Correct. Bard is now powered by Gemini Pro, but the multimodality seen in the demos is not currently available. And I think Gemini Pro Bard is not available in some countries yet, like the UK I think, owing I guess to regulations.
Ya this was just the mini update, the bard insiders were testing it a bit.
Are u not sleeping
I'm interested in generating code. Is AlphaCode 2 integrated with Bard yet or available elsewhere?
how will i know bard got updated with gemini ?
it shows at the top that Bard is now powered by Gemini Pro
@@1littlecoderI've heard it's on a par with GPT-3.5. I can't say it's disappointing because I didn't think Google was serious about competing with the 800 strong OpenAI.
@@BrianMosleyUKI’ve tested it, it’s GPT-3.5 level or maybe a little worse. But it’s also the Pro version in Bard, not the Ultra version. So maybe Ultra will be a little closer to GPT-4. But yeah, as for being a “GPT-4 killer” I’m currently very unconvinced.
@@therainman7777- Ultra will surpass GPT-4 is many ways. The Ultra-powered "Bard Advanced" will apparently have much of the fancy multimodal UI we've seen in the demos.
It genuinely does seem to be the case that Gemini Ultra is multimodal in a significantly deeper way than GPT-4 -+ for example, with audio input, it won't have to convert the audio to text, then process the text; it'll understand speech in audio inputs directly.
@@1littlecoder
I heard somewhere, GEMINI PRO can be tried in Bard, but Not in Europe and Canada at the moment.
I am waiting for Gemini ultra
Want to be excited about this but right now I'm annoyed with Google as it's not in Canada yet.
VPN 👀
@@1littlecoder yeah I tried, still not working as it needs an American phone number apparently.
❤i love competition excited what would happen next i was manifesting like gpt5 or something would launch or gpt4 for free also around 15-16 dec my chatgpt plus would end i hope i could continue and gemini pro is decent but would love to checkout gemini ultra
The HumanEval results of GPT-4 as reported by Google is not matched to what other sources say.
Easter Egg is: Apple
Woah! Great 🔥
@@1littlecoder kind of a strange choice by Google
Great work! Will be sure to start playing with the api next week when it drops….sense many different models from various vendors is the path forward if not the opensourced route…thinking google for video/vision, OpenAI for chat and language, stability video diffusion, or pika and runway if they offer API’s all integrated into one product is the future of these technologies….what an exciting time!🥳🤩🦾