GOOGLE and ANTHROPIC's NEW Insane CLAUDE SHOCKS The Entire Industry! (FINALLY ANNOUNCED!)

There's another new ChatGPT competitor that has been announced and nobody is really talking about it. As you can see right here from this article. The name is Claude and it's a very good Chat GPT rival and we're going to be getting into it. Now. You probably wondered why have I put Google in the title? Well, you can see right here it says Google has invested $300 million into Anthropic, co founded by former OpenAI employees and is based in San Francisco. So it's this investment by Google is a direct investment to compete with the likes of ChatGPT. So what exactly is this Claude and how does it work? So essentially you can see right here. So it says that Anthropic is essentially another chatbot that is an alternative to OpenAI's ChatGPT that is backed by Google. Now what you also need to know is that it's currently available and I haven't really seen too many people use it. So today we going to see how it compares to ChatGPT and exactly how strong this new chatbot is. So this is the article which they released and this was a couple of days ago. So apologies for not being on this sooner, but it is very interesting. So these are right here, They've been working in closed alpha for quite some time and I do actually have access to the very premium version of this. So we're going to get into this version in a minute. But you can see right here that it is very, very similar to ChatGPT and in terms of how you can ask it questions and in terms of how it delivers those questions as well. So it's really, really interesting, especially since it also has different buttons that you can press in order to generate certain prompts and, you know, provide examples of its capabilities. Now if you do want to use this, this is what you want to use. It's called PO and it's a fast and helpful AI chat. And you can see right here also that it is integrated into notion, which is something that people have actually used before. So let's go ahead. I'm not going to go through the entire thing right here, but I'm going to show you all a live demo of this and show you exactly how it works. So this is what we have right here. Now this of course isn't a tutorial, but this is essentially how this works. So on the left hand side you can see that there are a bunch of different apps. You can see chat GPT, you can see Claude, you can see Claude plus and you can see GPT4. So essentially what you can see there is you can see the normal GPT version, which everybody has access to, and Claude, the lightweight version, which everybody has access to if you use po. And then of course, you can see Claude plus, which is their, you know, pretty, you know, robust version, and GPT4, which is also their robust version. So essentially, Claude is, you know, you can see right here, and of course, GPT4 right here. So essentially what I'm going to do is I'm going to ask them both the same questions and see how their responses differ, and I'm going to use all versions of Claude. Now, I know that there are many different ways that you can use this, but we're going to go with some basic things. So I've just asked Claude, and this is the normal version of Claude, how to change a car tire. And you can see right here, it's giving me the results in a very, very quick manner, and it's giving me all the steps that I need. But like I said, I'm gonna have to compare this to chat GPT because I want to see how the response differs and if the response is pretty good. So, of course, we're getting nine steps here, and this is giving us some very decent stuff. And I think it also gives us some hyperlinks that we can click, which is very, very good. So I think if I click this, it's going to give me a hyperlink. Okay. No, it just, it just tells me more because I did click one of those things there and it just gives me more. Okay, so let's now go on to ChatGPT and see exactly how this response is going to change. I did actually go to the normal chat GBT just because I didn't know if there was, you know, going to be any kind of discrepancies between the one that's integrated into PO or the one that's not. So I'm going to go into normal chatgpt. This is GPT 3.5. So I think it is giving me the same kind of things, but I think this one is just getting straight to the point. Whereas this one is giving me, you know, find a safe location, and this one is just like, you know, straight to the point. So I think they're both giving us very, very basic steps. And you can see that the two basic versions of them are very, very similar. You can, you know, nine or ten steps just, you know, telling you every single thing that you need to know. So what I'm going to do now is I'm going to use the robust version. So of course we've got Claude plus here. Now, I don't have a million messages to send to this. I think they've only limited it to a thousand messages a month or maybe 300 messages a month, and that's for that and GPT4. So I've got to be careful with how many messages I sent. So I'm going to write how do I change a car tire into Claude Plus? I'm going to see if it gives me anything different. So I, I guess it doesn't seem to giving me anything too different, but I do think that this is a lot faster. That was really, really quick. So, yeah, I'm gonna go ahead and switch to GPT4 now. I'm gonna see exactly how the response differs. Now I do want to say though, that the responses can't differ too much simply because it is a very, very, you know, there's only a set amount of ways that you can change a car tire. So I'm trying to see if there's any kind of depths or, or any kind of steps that they do add between these platforms. And it's going to be interesting to see how this is going to work. I do just want to say, in OpenAI's defense, in ChatGPT's defense, because this is slower, remember that this is GPT4, which does have limited access, meaning that the messages are going to be slower and the bandwidth is currently being throttled because they can't roll it out to everyone, as it isn't currently too profitable to run GPT4, hence the reason for the slowdown. Now, I'm not sure this was supposed to happen, but during my experiment, GPT4 just simply froze. And I'm pretty sure it's because GPT is currently experiencing some issues. So I'm going to try the other app and see if that works better. And already this seems just, just incredibly better because as you can see right here, just the way that this is being formatted actually is going to help me because this actually tells me exactly what I'm going to need. I'm going to need a spare tire, a car jack, a lug wrench wheel, chocks, gloves, a flashlight, a torque wrench. You know, this is actually giving me all the stuff. Now, what I do like about this as well from po, which is another platform, what's actually I like about this is that it's actually highlighting the important parts here. So of course, GPT4, like I said, it didn't really get the, you know, it didn't really get the chance because of course, you know, for some reason it did stop working. I do think that right now there are issues going on because there were security breaches on this, where people could see each other's history. But I do believe that between claude, it doesn't seem too terrible. Okay, but of course, as you guys can see right here, it said the bot may make incorrect statement. It does not have knowledge after events after 2021. So, of course GPT4 is going to be insane. But if we compare Chat GPT okay. To actually Claude, you can see right here that, you know, it doesn't seem too bad. I mean, there's only nine steps, but we need to ask it more questions just to see if there's, you know, an incredible disparity between them, say. So far, GPT4 does seem like the only thing that's going to be winning this race. I do want to test about CLAUDE is if it can code. I'm not even sure if this is a feature, but we do know that GPT4's abilities are quite insane. So I'm going to ask it to can you write me code for a Trading View Pine script? And this is a fairly simple thing, so I'm going to ask it and we're going to see what response we get. So you can see right here that it actually does does write code, okay, in claude. Plus, they've actually managed to get this bot to actually write codes. Now, I haven't actually tested this just yet, so we don't know if this code is going to work, but it does actually have the ability to code, which is actually quite good. And I know that I've used GPT4 previously to code. I can't actually access my history, so I can't show you that. But GPT4 does give you code that works instantaneously, which is a step up from before where I had code that just didn't work at all. See right here. I did actually try to add this to the chart and it didn't actually work. Now, I'm no coder, maybe I've done something wrong, but when I use GPT4's one, it did work instantaneously. So maybe I've made some mistakes, but GPT4's one did work absolutely instantaneously, and I'm pretty sure if I worked with this a bit more, I could potentially get this to actually work. So it's definitely interesting to see if this is actually going to compete with ChatGPT because if it is false faster and it does work better, then it's definitely going to be interesting. I'm going to do is. I'm going to test something else. Okay. I want to test it's actually copywriting skills and its writing ability. So I'm going to ask it to write me a job description for a YouTube channel video editor with a background in editing AI related content. So I'm going to ask it to write me that and we're going to see the response that we get. So it says, here is the suggested job Description for a YouTube channel Video editor, and so far it looks really, really good. Honestly, this, this isn't Chat GPT. This isn't Google's Bard. I mean, it definitely does look better than Google's Bard. And you can see right here that this is. It's actually pretty, pretty good. And that was fast. I mean, that was very, very fast. And I mean, if we go ahead to GPT4, so if we test normal chat GPT, I want to see exactly how this does. And then if we also test Claude, there's no real point testing GPT4 because GPT4 is just miles ahead, I honestly think. So if I'm going to write this into Claude, I'm going to go back to the ChatGPT and you can see that it's very kind of different in how it sort of formats these, and I don't think that the differences are that big. Now, of course, we just tested how CLAUDE could code and we saw that it did fail on that instance. But that doesn't mean that it is completely fails on absolutely everything. So understand that this doesn't always give the broadest things because in order to test AI, you need to do a vast number of tests OpenAI did conduct on their recent test for GPT4 to just, you know, gauge how smart the thing is. So this video title from ChatGPT is really good because it actually does include many different pieces of software, includes relevant experience. And of course, you can see right here on Claude plus, you can see right here that it gives us, you know, a lot of other stuff as well, like requirements and responsibilities. And I think in the original Claude, let's see exactly what it gives us. I don't think it's gonna be too crazy, but it'll definitely be interesting. So, yeah, if we combine the two, you can see that Claude plus does do pretty well. But if I'm being honest with you, ChatGPT does kind of seem to organize this in a better way. It just says Key responsibilities and qualifications. Then Claude, when we look at Claude, Claude actually has some information, but it does look, it does look a bit, little bit, you know, unstructured because it does say of course, three years of editing experience with this app, this specific app and stuff like that. But it does seem a bit like unstructured because it is quite hard to read. Whereas if you go to Claude plus, it's responsibilities and it does have requirements, I would say that based on the probabilities, this is just a presumption, I don't know everything about this AI app that Claude plus is probably on the level of ChatGPT. So I'll leave a description to this article because it does delve and it does show you all exactly how different these two AIs are. And you can see right here that it goes over countless examples and I want to show just some of them because some of them are really, really interesting because they are very, very telling as to where these AIs are in terms of their mental capacity, their mental abilities and their cognitive abilities as well. So you can see right here, head to head comparisons, Claude versus ChatGPT and calculation. You can see right here that they've, you know, fared a bit differently, I guess you could say. So they asked this question. So for example here, for our first comparison, we asked both chatbots to take the square root of a seven digit number and you can see they both provide different answers. And it said the correct answer to the above problem is approximately 1555.80. And you can see right here that of course ChatGPT got pretty close, but Claude didn't get there. And what was interesting is, is that when they upped the difficulty of this question, Claude wasn't able to answer that question at all. You can see right here it says, I'm sorry, I don't know the exact cube root of that number. Calculating that is just beyond my capability. So it's clear that ChatGPT is miles ahead and we kind of already knew this. So of course Claude is pretty good. And I gotta say, based on what we've seen from Google, it seems that this is probably even better than Google's Bard. So it definitely seems like maybe Google might need to put more investment into this because it seems pretty, pretty interesting. So you can see right here it did ask another question which was based on factual knowledge and reasoning. And they said ChatGPT eventually managed to reach the correct answer, whereas Claude's answer was incorrect and he got it wrong. So yeah, there was a Multi hop riddle, which of course, as you guys can see right here in Japan is the right answer. And of course Claude got this one right. So I guess there are some small nuances between these. And I would say that, you know, currently what we have, based on what we have, based on the data, I would say that Claude's app, and maybe they've refined this even more by now because of course this article was written in January, maybe they have refined it. I would say that currently what we have right here is with Claude plus we essentially have Chat gbt that's just based on what I've seen so far. And of course ChatGPT is simply miles ahead because, you know, stuff like the code doesn't work and stuff like that. So let me know if I did miss a small mistake by that. But I do think that this is definitely very interesting for Google to be still on the heels of Chat GPT because a lot of people think that because bot is now released, Google is taking a very, very back step. But that isn't the case. They're still working on investing in companies that are developing good chatbots. And this is honestly a good chatbot, guys. It's not bad. I mean it doesn't know certain things, but honestly we've only come so far, so quickly, so I would say of course ChatGPT is going to be better. But this is also very, very good and it does power some other things as well. And this is also good, this app called PO, because it allows you to access ChatGPT4 whilst it is down right now. So definitely use this if you.