Google Reveals SHOCKING Plan To Release Model 10x Better Than GPT-4

So Google is honestly making some big moves. Take a look at this. Google is investing $300 million in an OpenAI challenger that will take on ChatGPT whilst focusing on AI safety. This means that Google is not messing around. They understand that they need to catch up to OpenAI and they are doing something insane. This is crazy. I'm surprised everyone isn't talking about this. Take a look at this. The company that they're working with is called Anthropic, and what they're building is Claude Next. And essentially they want to build something that is 10 times more capable than today's most powerful AI, but that require a billion dollars in spending over the next 18 months. So we know, okay, that the most powerful AI at the moment is GPT4, but this company that Google is investing in and they have a 10% stake in, are essentially trying to build something that is 10 times more capable than today's most powerful AI, which is GPT4. Which means that the race is really going to heat up because at the moment it does seem like OpenAI do have a monopoly on the most powerful AI in the world. But if this company manages to get something that is even surpassing ChatGPT, then we could see AI on a level that we never even thought out before. So it really means that now we have a situation on our hands. We now have Google investing millions of dollars and potentially billions over the next couple of years to increasingly ramp up at the rate in which this AI is developed. Now, this company called Anthropic, I'm going to show you all a demo of the tool to show you that they're not far away from ChatGPT. And honestly, if they do continually develop faster than OpenAI, then potentially they could catch them. Now, of course, you might be thinking that Anthropic is owned by Google, but they're not actually owned. They only have a 10% stake, which is the $300 million deal that they recently did acquire. But I do think that in the future, if Google isn't able to fix Bard, they, they might just completely buy majority shares in Anthropic. That way they can just expedite process and it's literally just integrate it into all of their products. Now, here's something else that is very, very interesting. The AI that they are building, which is going to be called Claude, which I will show you later on in the video, is essentially going to have a different kind of model. So it's going to be constitutional AI, which essentially is going to be ensuring that this kind of AI does not go rogue. And they're essentially making sure that this AI is aligned with human intentions, because if it's not, then we will have some seriously dangerous scenarios. Now, something that you do need to know is that this frontier model that they're building, essentially they're focusing on using it to build virtual assistants that can answer emails, perform research, generate art books, and more. So what's interesting about this is that they're talking about generating virtual assistants, which means that I do think that we're moving closer to, towards what we saw recently with ChatGPT and what we saw with AutoGPT, in which these models are actually just completely autonomous. Now, I'm not sure if they're trying to do that, but from the text that I'm seeing, it definitely seems like that is going to be some of their main focus. Now, they also go on to state here that these models could begin to automate large portions of the economy. And this is the key number. They say that companies that train the best 2025 to, to 2026 models will be far too ahead for anyone to catch up in subsequent cycles. Which essentially means that by 2025 to 2026, whichever company has the best AI model, you're not going to be able to catch them and their AI is going to be too advanced because the rate at which the progression is going to be at, you're just simply not going to be able to catch them and there will be simply no point in investing in that AI. Now, what's also interesting is that Google is consistently making moves to try and catch ChatGPT or, or OpenAI and Microsoft. So what we have here is Google's new move. So Google merges with DeepMind as Alphabet plans to take on OpenAI's ChatGPT. So essentially what they've done is they've merged two companies that are working on artificial intelligence into one. Because of course, as we know, the more people you have working on something, the larger the results that you're going to get. It makes sense to unify your forces and try and see exactly how far you can take it. So without further ado, let's take an actual look at Claude next, which is going to be the main competitor for Chat GPT in the foreseeable future. And let's see if it actually is worthwhile and how it compares to GPT4 or GPT3. If you want to actually access Claude, you're going to have to go to this website, po.com you're going to have to sign up and make an account, and I think it's around 20amonth. But once you've signed up, you then get access to Claude plus that you can see right here. And of course you do get access to GPT4. Now, just like GPT4 has chat GPT Claude has Claude Instant, which is a more lightweight, a more friendly and a more instant version. Now it's essentially a more powerful version. You can see Claude Instant is just a normal one and then Claude plus. Now, the reason I'm showcasing this today is because this is going to be the main competitor that is going to be facing off in the AI wars when it comes to GPT4 or ChatGPT, however you want to call it. Now, what's also cool about Claude plus, before I give you guys a real demo on how good this actually is, if you remember, a couple of days ago, Amazon actually announced that they're going to be introducing their Amazon Bedroom software, which is going to allow you to access many different large language models and many different pieces of software so you can run your AI company more efficiently. They're actually using Claude on their Amazon software. So it's going to be interesting to see how Amazon manages to fine tune this and if it's able to beat ChatGPT. So one of the things that I like to do is I like to experiment with how these bots are able to code, because I do think the coding is very different. So this is something that I asked GPT 3.5 to do and and it did actually struggle, but with GPT4 it got it completely instantly. So I'm going to ask Claude plus if actually do this. So we can see that it is actually coding this bot right now. So essentially, Pine Script is a language that you use in TradingView, if you are familiar with charts. And essentially you can then copy this code and you can simply paste it on a chart and then you can simply use it as an indicator. You can see that this was pretty, pretty quick and I'm pretty sure that if I apply this, I'm sure this should work. So let's go ahead and and test if this actually does work in a trading view. So if we go over to TradingView and go on the Pine editor and then I click paste and then I click save, we should be able to see if this actually works. And I think it doesn't work at the moment because you can see right here it says Script could not be translated, Script could not be translated from Null. So it's clear that this doesn't actually work at the moment. But I wouldn't use this as a benchmark to see how good the software is, because there are many different ways in which you can actually test whether or not software is good. There are, of course, some other tests that I do want to run, but of course, if I do actually ask GPT4 for this, you all know that GPT4 is absolutely excelling at pretty much everything it does do this. The only gripe that I do have with GPT4 is that it is much slower than other models. Okay. Like, for example, PO instantly gave me this very, very quickly, although it doesn't work. So I guess you could say it's completely useless if it's slow. But GPT4 does actually move a lot slower with its code. I'm guessing that's not a bad price to pay if you're going to be getting some accurate code. So you can see right here, Wagam, I'm just simply going to wait for this code to be pasted and then I'm going to copy the code and see if it adds it to chart. So now the code is here. I'm going to copy this code. I'm going to go over to the Pine editor, I'm going to paste it in, then I'm going to go ahead and click add to chart. As you can see right here, it says I cannot click. So you can see right here. After one small iteration, Chat GPT4 was able to give me a very, very interesting indicator. You can see that instantly trading view changed. And it showed me the net profit of, of how much I would be making on every single trade and how much I would lose. It showed me that the net profit if I use this indicator would grant me around $1,000. You can see that right here. And that was literally just something that GPT4 literally just created, which is absolutely insane. Which just goes to show you how powerful GPT4 is. Now, of course, Claude plus is still very good at other things. So let's say, for example, we say, can you explain how to backtest this strategy? And then, for example, I ask GPT4, can you explain how to backtest this strategy? We're going to see what both of their responses are. So it's very interesting because CLAUDE plus gives us this kind of explanation. It says, define the variables, enter trades on your entry signal, exit trades on your exit signal. But ChatGPT4 actually gives us the exact kind of definition that we want because it says, log into your account, open this chart, navigate to this. It seems that GPT4 kind of knows exactly, exactly what is going on. Okay, so it's very, very different. Now if you're looking at the conversational aspects of Claude plus maybe you need to write an email. Let's say, for example, I wanted to hire a video editor for my YouTube channel. Okay, I'm gonna make a new chat, I'm gonna clear this and I'm gonna say, okay, so I've typed it up here. Can you write me a job description for a YouTube video editor for someone who edits AI related content? Then I'm going to click post, then I'm going to go ahead to GPT4, then I'm going to do the exact same thing. Now, you might argue that comparing GPT4 to Claude next isn't fair because of course GPT4 is much more advanced and of course they're still developing this model. But I just want to show you all the benchmarks of where they are. So I'm going to ask poe, and you can see right here that this definitely does look pretty decent. It says, hi there. We're suggesting for a YouTube video editor. It posts everything that I would need. Freelance position. It says two years of experience, proficiency with certain programs. This is definitely really, really good stuff. If I'm being honest with you, I would argue that this is far better than Bard. So if you're thinking about Bard, and if it is like this at all, trust me, this is a lot better. The only thing I would say that I don't like about this at the moment is how you access Claude through this app. It doesn't seem to be very like, I just don't like the format. Whether it was GPT4, you can see that being able to digest the format on this screen just seems a lot more easier, especially with the way that the chat window is set up. Maybe that's just a small thing, but it doesn't seem like there's too much of a difference here. But I do think that with GPT4, it's showing that it does have a little bit more stuff there. And then of course, when I asked Claude Instant, which is another model, it actually seemed to get this even better. So I'm guessing that Claude plus, even though it's seemingly better, doesn't perform better at certain tasks because Claude Instant actually did it better. And of course, just like ChatGPT, it says it does not have knowledge of events after 2021. So I'm not gonna lie to you guys, it is very interesting to see how they fine tune these models. I think that these models are definitely going to be competing against each other in the future to see where they are now. So it's like a sort of time capsule to see how these kind of models do develop. Now, of course, this was very, very quick, this post that it did give me. I could definitely use this. And this was also very, very quick too. So I would say the GPT 3.5 and Claude instant are quite similar in their ability. Now, let me know what you thought about this video. Let me know if this was something of interest. Are you going to be using Claude, or are you going to be sticking to ChatGPT? Will Google manage to catch OpenAI or are they too far ahead? These are all of the questions that many people are wondering, and let me know exactly what you're thinking and I'll see you in the next video.