All right so Google just came out with an announcement that is their answer to GPT 4 and GPT 3.5 in its current state and yay they’re coming out with a model that is better than gp4 and this is both very exciting and interesting and very disappointing so in this video we’re going to talk about both plus everything you need to know as a user of these products because a part of this announcement is available today for Google’s B but more on that later because first things first we need to talk about the ginormous announcement here Gemini Ultra, their most capable model moving forward.
HTML tags:
All right so Google just came out with an announcement that is their answer to GPT 4 and GPT 3.5 in its current state and yay they’re coming out with a model that is better than gp4 and this is both very exciting and interesting and very disappointing so in this video we’re going to talk about both plus everything you need to know as a user of these products because a part of this announcement is available today for Google’s B but more on that later because first things first we need to talk about the ginormous announcement here Gemini Ultra their most capable Model moving forward
Summary
It’s multimodal from the ground Up, not as an afterthought as is the case with gp4 based on their design. Some of these use cases are mind-blowing. It actually beats gp4 on most benchmarks, which is amazing. This is a truly Next Level model that will be coming sometimes early next year. So, yeah the numbers check out here from what they’re showing what they’re telling us this is really impressive. Google Executives will be able to maintain pH in front of their board members and shareholders at the end of 2023 and users will be getting Next Level Tech in a few months. I like it, that’s a win-win. Here’s the biggest deal. We just crossed a specific score on this MMLU Benchmark and the score that they talked about is this 89.8% on the MMLU Benchmark. This is what human experts achieve when taking that test. Okay, so this is the very first model we will get access to that is actually better than human experts at specific tasks. Yikes! So as you can see, the Gemini Ultra scores 90% on this and gbd4 scores 87.2. But they actually tried to fool us here. No, not happening on this channel though because look in the blog post they’re comparing the performance of Gemini Ultra and GPT 4 but they’re not using the same prompting technique. Okay, so whether that’s a fair technique to evaluate it or not, they used prompt engineering to get the best result possible. But what they’re comparing it to is to a five-shot approach on gp4 which is a different and generally speaking less capable way of prompting the model. It’s important to be aware of that. But other than that, it just beats gbd4 across the boards, so it’s fair to assume that this is a better model than gp4. Another important fact here is that this model will have a 32k context window which does not match GPT 4 128k. Some other important points that need to be taken into consideration are the multimodality of this model and the developer ecosystem. In addition to this, Google has released other models, including Gemini Pro and Gemini Nano. The Gemini Pro is essentially the 3.5, with better performance in certain tasks and worse performance in others. Gemini Nano is also a significant release with two versions, one with 1.8 billion parameters and the other with 3.25 billion. The Pro model inside of BART today is good, and it includes prompt presets which are beneficial for developers. It’s important to bear in mind that all the prompts learned from this channel and prompt engineering resources can be applied in the BART model. Lastly, it was emphasized that an AI arms race is happening, with multiple companies competing to get ahead in the game.
