What is the Google Equivalent to ChatGPT?
Google Bard is Google’s competitor to ChatGPT.
In theory, Google should be waaaaay ahead of anyone else when it comes to developing useful artificial intelligence.
So, when ChatGPT came out, it made Google look like they were behind the curve.
For more details on what exactly ChatGPT is, click here.
In response, Google rushed out an AI chatbot…
Is Google Bard Better than ChatGPT?
No, Google Bard is substantially worse that ChatGPT…
…even before ChatGPT allowed people to use GPT-4…
…and even without considering how plugins will grow ChatGPT’s capabilities.
Relevant page on how to use ChatGPT.
Another relevant page on how ChatGPT works.
Most up to date video:
Check out this relevant video:
Approximate Transcript of the above video:
Hi, this video is about comparing chat GPT to Google’s bars. And I’m going to put it through a few tests here, I’ve heard some things in about how good or bad bar it is. And I’m going to test for myself and see what we find with my own types of prompts that I do, which is a little different than some people I think so. And already right out of the gate. So actually, I am looking for a project manager for for SEO. And I wanted to write it out in detail. And I thought, I’m just gonna kind of dump out my main thoughts and have chat GPT throw something in there. And then I also would try it with Bard. So first of all, Bard just says, Nope, I can’t do that. Let’s actually give it another shot. I’ve heard sometimes that if you just if you give it the prompt again, that sometimes it actually will give it to you even though the first time it didn’t like I don’t see, there’s no reason why it wouldn’t see now that’s really interesting that it did that. There is a nice feature here that there’s multiple drafts. Let’s take a look at them. Let’s let’s read this real quick, my pauses, I’m gonna read them. And then I’ll give my my summary. So right out of the gate, all three of these, it seems like it’s making up a lot of stuff. So I gave it pretty specific instructions. And then while there’s a certain amount of creative license that I think is expected, it it it does hallucinate a lot of things, I said nothing about this, I did not say anything about this is not going to manage all aspects, this was a very narrow job, or job description. And I didn’t say anything about the amount of experience at all. Now maybe that’s just part of the standard, but I don’t really like any of these, I don’t think I would use them. And I’m gonna read this one, just this one is much better, it did a much better job of taking into account what I put in there, what I put in the description, and actually using it thoroughly in here. And it did add a few things. But basically, it gave me back what I gave it. And with a little bit of things cleaned up, I’m not sure there’s really anything that they added. I guess strong customer did kind of expand upon this. You know, this, this is again, a little bit awkward sort of added a little bit here. But that that’s about it. I liked this way better, this is far more usable. So one point to check GPT. And in this case, one to zero, was keep a little counter over here so we can check this. Alright, so let’s try a new one. We’ll start this one. Okay, brainstorm 10 gift ideas for for women. All right, let’s see what this one does. So an interesting thing that I already noticed right out of the gate. So this is I’m gonna get give a point to Chet GPT. Before we even do this, because it can’t start a new conversation. Chat UBT lives leaves a lot to be desired in terms of the UI and interface and efficiency. There’s no search, there’s no archive, there’s no, I can’t select multiple chats at the same time to either clear them all or have to go through the Spirit. It’s very inefficient. That said, Bard seems to be like hold my beer. how inefficient can I make this, this is extremely not useful as like a business tool. There’s no history, you just have one chat conversation. And I kind of don’t want to using this. So I’m going to have to reset it. And then I’m basically starting started from scratch. So I like that I can go back to this if I need to. With this. If I wanted those I would have had to copy and paste it otherwise they’re just gone. Alright, so we’ll give that a second. Okay, let’s read these here are 10 gift ideas for women can do a gift card piece of jewelry. Oh, this one’s not bad. Let’s check out the other ones. Very, very similar. And these all these are all decent.
I feel like these are a little bit better. Just because the detail and kind of a it seems more real. I would say this seems more robotic. But I wouldn’t say either one of these is I would say this is bad. So we’ll just we’ll just maybe we’ll get some bar doesn’t feel so sad. We’ll give them both a point for this one. Okay, so this one should be interesting. We’re gonna write a poem about artificial intelligence in this in the style of Yoda. Let’s give that a second. And we’ll see what this does. I’m pretty sure this is going to be a point for Chet GPT. But we’ll see. Intelligence artificial a wondrous thing it is learn and grow, it can to do many things from simple tasks to complex, it can accomplish all with his vast knowledge that can help us all. But be careful, we must pray it can be dangerous if used for evil, it can cause much harm. So let us use a wisely. You know what’s interesting? It like starts off in the tone of Yoda. And then it really only use it one more time, I would say this is kind of a fail the poem is okay, but it’s basically a fail maybe this one. Yeah, it just doesn’t. It’s weird. It seems like it can’t hold its thought in its head and it loses what it what it was actually trying to do. Which is pretty, which is a really bad sign for Google, I think. Let’s try this one. And to be clear, I think this is something where chat GPT should shine. In galaxies afar a tale a tale of wisdom, and code where circuits dwell, artificial minds and silicon bound, a Yoda like verse in rhyme, profound Telde well bound profound from human hands creation rise in metals and wires, intelligence lies, okay. So this is already way, way better, I almost want to give them two points. This is awesome, actually. And to be clear, this does a much better job of keeping the context. This is I think, this is a theme that you’ll see throughout a lot of these, we will just bar just kind of like do to do Oh, we were asking about that. I’m gonna go over here. Now. It seems just way, way inferior. So we’re gonna give, but I’m just gonna give them one point. Just for the just to, again, I don’t want bar to feel too bad. All right. But yeah, I’m going to go into one that will definitely feel bad. Okay, so this one, we’re going to write a Python script to create a local graphical user interface to call an open API API for the user to have a discussion with DaVinci over and we’re going to reset this
I’d actually be surprised if it does this. So pip install open AI, I do have some Python experience. So I can sort of read it, but I’m not like a Python expert. Okay, so this has to do with I think this is the gooey stuff, the graphical user interface. It’s actually where a lot of the complexity of this comes into play. So it’s actually creating this main window within all this code.
Not sure this is correct yet. But this is definitely really nice. I love how the window looks and how it’s commented it makes it really easy to read. Oh, actually, hmm. In the past, sometimes. BART has been known to just be like, Man, I don’t want to do programming right now. import random, important time to find Get Response prompt. Okay, so I actually don’t know if this is better code or not, I don’t think I have experienced enough to really be able to tell what would be better to use this Tinker, you know, I’d have to actually go test this code. In general, I would say this is a little bit of an advantage to here. This seems to be a bit deeper, it’s definitely easier to read. There’s more context given you know, this one just has spits back. There’s, there’s not like, much detail in it. It doesn’t does it all in one shot. Let’s see if we get a little bit slightly different. Maybe something there’s no commenting or there is commenting, but it’s less. I’d say this has to go to church up but Barton better than I expected. Okay, now we’re gonna try some pros. Write a short story about a penguin that overcame great hardships to be a big success. We want to reset this. There we go. Okay, so I’ve read all four of them, basically. And I would say that Chachi pte is better. The way that this is written is it’s just much more robotic and much more lower level, like a really young person would write who’s really trying to hold to the rules without trying to actually kind of tell a story and in a rich way, when you read, which LGBT has given us, it feels more real. There’s more variety to the sentences. There’s more detail, there’s a little bit more nuance. You know, there’s, there’s more emotion in it. This this is just far, far superior. I gotta get I gotta give a point. To check GPT for this one. Okay, so this next one is going to be one in the advantage of GPT. I mean a bard because it can actually pull from here. So now to be fair, there’s plugins that the judge up we’ll have that we’ll connect this pretty soon. And so I think that this is will give a point to Bard. Well, let’s check this first. Here’s some of the news. It’s called lambda. This is not really news. It’s happened a while ago. And it’s supposed to, I guess I could click Google, it’s supposed to give it sources I might even. So if we did a Bing, I think that this would, it would it would look like this with, it’d be basically using GPT four, and it would give it sources. So in a way, I almost want to give it to you because you could use it in Bing and get get a better result, but I don’t really like that they don’t add any sources in this doesn’t really seem like news. But it may be Oh, man, I hate to even give them a point. But we’re feeling kind of sorry for bar right now. So we’re gonna give that another interesting thing is kind of the text. The the context window sizes. So let’s try this. Chat GP tes 3.5 TPT. Let’s actually try this and see what see what it says for this.
Huh, I wonder if this is true. So what they meant was this is actually trained on this, this is not a how many tokens that you have in the context window. So for example, GPT four has 8000 tokens, which is something like 6000 words and it at some point, it’s there’s there’s a way or supposedly like a 32,000 token one, which is something like, it’s like 15 pages of content or something like that, I almost think it couldn’t find a good answer on the token window size of Bard. But I kind of don’t care if it’s a really big size, what we saw before, especially with like the Yoda thing. And there’s some other examples of this, where it kind of loses its way that even if their context, one was huge, I don’t think it really matters, because it’s not an effectively large GPT for I’ve noticed definitely has, it can go way back. And it can analyze a lot of data at once. I’m not sure exactly how far it goes back within here. And it might actually be sort of a little bit of a moving parts thing. Additionally, with the plugins that are about to come, you can analyze way more data than what’s in the context window, just due to the way the plugins will work, you can give it a 300 page PDF, and it can use that as a reference with the plugin and some kind of AI database type stuff. So I think I have to give this to to check GPT just because there’s kind of more than it can be given in terms of the input data to affect the answer right out of the box. But additionally coming soon, it’s going to be much greater. And I don’t really care what bars window is because it just sucks that kind of keeping track of where it’s at. So I don’t not sure it would matter. Okay, so where does that leave us chat GPT seven, bar two. And actually, this is being I think, generous to Bard. Google really needs to step up their game, I don’t know what’s going on, you know, my for the last 1011 years, I’ve been up all of them Google’s business for a really long time. And they seem to have their shit together. And a lot of ways and the technology that GPT four is built upon was created at Google. To be fair, I think there’s like 10 or 12 people that were on that transformers paper. Only one of them I think is still working at Google. And so do they have some better technology in the background? That they’re just holding back? Because they want to be safe or whatever? Or is this the best that they got? Because if this is the best that guy, they’re in trouble, at least from the AI worst standpoint, which I think will take a huge bite out of their core business of ads on the search results. So this would be very concerning to me. I’d hope that they come up with something a little better. I know this is their experiment. Maybe they’ll make some really big gains and they’ll outsmart open AI and Microsoft but every day that this the GPT just actually seems to be getting further ahead and stead of the reverse of that. So I don’t really see any reason to use Bard for anybody for anybody. Because if you the one use case that I thought that maybe Bart would better at would be, like actually using sources and searching the web for sources, which is Google’s kind of bread and butter searching the web. I think that doing it with Bing, and Microsoft Edge is just superior. Because they just they have you have to actually quit Google. That’s not the same thing as pulling in relevant sources. So if I were to, you know, give like bonus points, I would say this is really more like, this is effectively almost like 12 The one that that’s really what I would say that like, if you look at how much better is Chet GPT than Bard, it’s at least 10 times better. So that’s my conclusion. Let me know what you think. Like if you liked this, subscribe for more awesome AI content. Thanks. Have a great day. Bye
