ChatGPT vs Claude: Canvas, Projects, and Artifacts Explained 🤖

So we're starting off today looking at Claude. We're going to be looking between Claude, Anthropics, Claude. So, you know, OpenAI have chatgpt. Anthropic is the business that have Claude, which is just another very similar large language model, works in a very similar way to ChatGPT. If you've not seen it before.

There are some slight differences in terms of the power behind the models, behind everything, and also the way that it's laid out and structured and the way that the features work. We generally find that these large language models compete to each other on such a level that when one updates something and we go, oh, that's a really great feature, and we all swan to Claude, then two weeks later, OpenAI will update their feature and they'll do it just a little bit better. So there's always just this back and forth of who's got the better features and which one's more powerful than the other. I do tend to use Claude and chatgpt interchangeably and at the same time for a lot of the stuff I'm working on. So, depending on what kind of project I'm working on, I will choose either ChatGPT or Claude to work on it. But a lot of the time what I will do is I will take that prompt that I'm using and I'll put it in both and I'll just see which one generates an answer that I prefer, or I'll take bits and pieces from my preferred ANSW and combine something together and do a bit of back and forth with both and just see which one generates an answer that I like more. Okay, so I do use CLAUDE and chatgpt interchangeably in that sort of way. And if you've not tried CLAUDE and you've only tried ChatGPT, or if you've not tried either, then it might be worth you just signing up for a free account on both and just having a play with both and seeing which one you know, speaks to you a little bit better, which one you prefer. They're both incredibly powerful tools and, and both of them are going to do a lot of great things for you if you know how to use them. So what we're going to look at Today is the two features in Claude and ChatGPT that are essentially the same. They've got different names. So Claude is called Artifact, and Chat GPT is called Canvas. So if you've heard us talk about any of those before, that's the, the, the similarities between the two is what we're going to look at and how they differ and why you would choose one over the other in terms of whatever project that you're working on. So I'm going to show you in a second, I'm going to jump into the software and show you what that looks like. But first of all, let me just tell you when artifacts might be used. So in Claude, this is. So Claude creates an artifact automatically for you. You don't have to ask it if any of these criteria it considers to be met. So if when the content it is sharing has the following characteristics, it is significant and self contained, typically over 15 lines of content. So if you're asking it to do a large thing, you're asking it to write a long blog post, a really long email, some wording for your website, a systems operation manual for your clients.

An onboarding, series of emails for your clients, you know, your service package, your brochures, whatever. If it's something that you're trying to create, that's quite a large piece of material that you're going to create, then it will automatically open up an artifact for you and set you up in that way. So what you're going to see is you're going to have your chats on the left hand side of the screen and your document that you're working on on the right hand side of the screen. So you've got this kind of set up so that they're next to each other and you can work on that document and also have the conversation with.

With either Claude or ChatGPT on the left hand side. So that's how it's going to look. And I just find that a lot easier than having everything in just one long line in a conversation. And especially as a lot of the time when you start a conversation with ChatGPT or Claude and you say, can you do this thing for me? The first paragraph that it will say, or the first sentence will be like, yeah, that's a great idea, let's have a look at doing that and we'll do that for you. And then the rest of it starts and then if you want to copy that, you then need to go back in and remove all of the bits where it said, oh, that's a great idea, let's have a look at doing that. You'd need to go and remove all of that stuff from it. So having the document on the right hand side and the conversation on the left just, it's an easier interface and how to interact with it. So if you have a large piece of content, Claude generally is going to just recognize that that's going to be a long piece of content and it will open artifacts for you and make you a new artifact. If it doesn't do that, you can just add into your prompt, please use artifacts to do this or something along those lines. Use artifacts. You just add that to your prompt just to make sure that it is definitely going to do that. Because sometimes it doesn't pick up on it or it just starts to do it into the main chat and you actually wanted it in an artifact. So you just make sure you add that to your prompt if it doesn't automatically do it. So also if it's something you're likely to want to edit or iterate or reuse outside of the conversation. So if it's something that you're going to do a back and forth on, normally when I start mine, I'll say, can you help me create a first draft of this thing? So it might be an ebook or some wording for a landing page, whatever it might be, can you help me make a first draft? So it knows that we're going to go back and forth on that we're going to change things, we're going to make little, you know, edits to it as we go to get something that's perfect, but it's going to be a work in progress and that's why it would open this side by side screen so that we can do a bit of that back and forth with, with, with each other. I'm saying, like me and Claude can work together to make that, that product. So if it represents a complex piece of content that stands on its own and you don't need to have that extra conversation, then it's also going to put that on the right hand side as well. And if it's content that you're likely want to refer back to and use later on. So if you're doing an ongoing long kind of project where you're going to want to have that conversation and the document saved on the right. So the examples of the artifacts that it might create are documents obviously is going to a document on the right hand side. Now it can do that in plain text and that's generally what we're going to see most of the time because we don't really deal much with code. But it could also do it in markdown and markdown is just a way of laying out text so that when you copy it, it has all of that formatting built into it. So it will have like titles, emboldened things, things that are underlined or striped through, you know, italicized. Just by using different symbols in our text, we can get that formatting immediately to occur. So you can also write things in Markdown, which is very helpful when it comes to writing prompts and stuff. It will also handle code, website code as well for HTML, scalable vector graphics, images, not that I've particularly used that, and also diagrams and flow charts and things like that as well. So I will show you it making a chart because Claude does actually make the chart for you, whereas ChatGPT makes the code for you to then go and input that somewhere else to make the chart. So they do both have slight differences in how, how they approach that. So that's artifacts. And then we're going to hop in there and have a look at that in a second as well. But I'll just talk to you about Canvas at the same time and then we'll go and have a look at both of those. So Canvas allows you to create that large document or a large piece of content on the right hand side. So you've still got that side by side effect. Generally I find that with Canvas you do need to summon it more often. You need to tell it that you're going to use Canvas now, you know, it's a bit of a balance and I, I think I use Chat, GBC and Claude a lot. So this is just my personal perspective is that I found that Claude generally will start an artifact without me asking, whereas ChatGPT, I do need to prompt it to get it to use Canvas a lot of the time, which is very easy to do. But it's just that extra step of making sure that you, if you know that it's not going to do that and you don't want to waste your message limits by changing things and going back and forth. Just make sure you, you do tell it to use Canvas before you start doing your, your, your work.

So the other great feature about Canvas rather than artifacts is that you can actually directly edit the document that you're working on. So when you've got that side by side view, you can have Canvas open on the right hand side and you can click on it and, and delete pieces of text, you can type in text that you want to do, you can change things, make it that, make it a heading or make it boldened or anything like that. So you can, you can actually edit it as if it were a document on the right hand side of your screen. Whereas Claude, you can't do that. But what you can do is highlight sections and say, I would like you to change this in your conversation area and it will go and change that for you. So I do think genuinely that even though Canvas has those great features and those extras, that it, that it can do that and have that editable document and some other things that I'm going to show you in a second. I do think that Claude generates a quicker answer for you because it doesn't have all of those extra features, because it doesn't have all of the faff. It can just focus on creating the content for you much quicker. So it's a, it's a slight trade off and it's a very minor trade off. So Canvas is still incredibly fast, but it's just in terms of, you know, Claude might take 30 seconds to make something, whereas Canvas might take a minute or a minute and a half. So it's taking longer. But it do does have those extra features. So it just depends on what you're going for. If your priority is speed, maybe use Claude. If your priority is being able to iterate that document with ease, then maybe go to Canvas for that.

Also.

Canvas. Well, kind of along those lines, I have found that Canvas can be a bit glitchy as well. So generally if I'm using it, I will double, triple, quadruple, check what it's writing because sometimes I'll say, can you remove this section and write this fresh for me? And what it does is it just writes the new thing on top and then still has the old one below it. So I have to go in and make sure to delete that. And I'm finding that Canvas does that a lot more than Claude. I've not seen Claude really do that yet. So Canvas can be a little bit glitchy and again, it could be due to the fact that it has all these extra features that is drawing that energy. So it, or it just needs more, you know, play time for OpenAI to kind of fix these little bugs. But that's what I have found personally when I use these and I do use them a lot. So Canvas uses GPT4. Oh. So if you're on the free plan, you of ChatGPT, you get unlimited use of 4oh mini and you get limited use of 4o. Now you do. I can't remember off the top of my head how many, how many numbers of messages you get. And I think it does vary based on usage generally.

But you'll just need to be aware that if you are creating something with Canvas and you're on the free plan, you are going to be eating into that message limit. So just be aware of that. You can also use the O1 model with canvas. So the slight difference there, GPT4O the, the great feature about that is that it's omnimodal, which means it can handle different types of input, so text, audio, image, images and eventually it will have video as well. But we're not quite there yet. But it can handle those different types of inputs and that's why you would use 4.0. So it generally has all these extra tools. O1 is a slightly more advanced model that is really focused on advanced reasoning and taking complex things and doing them step by step and having a real methodical approach to stuff. So O one thinks a lot harder and longer on doing its projects. So if you want something that will do that and that will go away and think a lot harder, maybe you're doing a research task or something, then you might like to use O1 for your canvas. But again, depending on what plan you are on, that is going to eat into your message limit, however much limit you have for the O1 model. So just be aware of that when you're choosing that model to use your, your canvas tool with. So artifacts. Also actually they use Sonnet 3.5 which is their latest model. And again I think you get a certain amount of messages within a certain time period on Claude. So it's, it'll be something like 40 messages per three hours. They do keep it really vague because it does change based on the general usage, you know, globally, you know, if it's really, really overwhelmed, you will get less message allowance. So it is quite vague. But it's something like, like 40 messages per three hours. And obviously if you're on a paid plan that increases. So you do get more of an allowance with that.

So with that description let's now go ahead and jump into those softwares and see what they look like and show where all of those features are. So we'll start with Claude and I'll just open Claude up here.

So this is what Claude looks like and it's got a slight different interface to Chat gbt, which is this one. Or let me just go back to kind of the home screen. This is what ChatGPT looks like and this is what Claude looks like. So you've got very similar how you've got, you know, the sidebar, you've got your messages in the middle and all that sort of stuff. But there are some slight differences that we're going to look into. So the first Thing that I want to do is just show you what artifacts looks like. So let's get it to do something that would require an artifact. So can you write a lengthy. Because we want that length.

To trigger the artifact, a lengthy blog post on how to save tax in the dental industry. So this might be a blog that you would do if you, if you work in the dental industry. So what you'll see that it does is it's going to open up on my screen. The artifact is going to go to the right hand side and my conversation is on the left. So you can see how quickly that's going as well. So if you can kind of remember how quick that is. And then when we look at ChatGPT, you'll see the difference in the speed. Claude is definitely much quicker. And Claude can also handle much larger amounts of text in terms of uploading files, in terms of, you know, how much text it outputs to you as well, it can handle much larger amounts of text. So that's generally might be a reason why you choose that rather than chatgpt as well. But what you'll see now is I've got on the left hand side, this is the conversation where it's responded to me and said, yes, I can do that for you. It's made the document and then it said, yeah, I've done that for you. So if I click here, this is my, my document now. So if I come away from it, then this is what I would click to open that document on the right hand side. If I had multiple different documents, I can click amongst the chat to find the different documents there as well. But what I can also do is if I did have multiple documents. So let's say, for example, I was writing an ebook for a new lead magnet to bring clients into our business and I was writing it chapter by chapter so that I could make sure that I flesh out each section properly. And I had a five chapter, I might have five different documents. Then what I would do is I would come over here to the top right of the screen where it says chat controls.

And that will bring out all of the different artifacts or documents, if you like, that I have open within that conversation. So I have one document here. If I was to make two or three different documents, they would all be listed here and I can just quickly jump to whichever one I'm working on. So if you're building a website that has different pages and you've got different documents for all of that, or you're doing an ebook with different chapters or you're making a systems guide on how to use QuickBooks and you've got different chapters for that and you've got different pages, then this is where you would come to find all of those different documents laid out.

Something that I actually noticed the other day, which is interesting, is that Claude does have a dyslexic friendly option here. So I'm not dyslexic, but if anybody is, then that might be something that you find interesting and. Or you can just change the font to a different kind of font. It's using CL Claude 3.5.

And you can add content to this conversation if you want to as well. And if you're uploading content to a conversation, it will then tell you all of the things that you have uploaded within that conversation here it will list that all for you as well. So if there's something you want to jump to just to have quickly have a look, it's all listed here on the chat controls. So if I come off of that, this is how it will look just in the chat. And if I want to open that document, I just click on that document and it's there on the right hand side. So in terms of editing this document, I can't directly type into that section, which is something I can do on Canvas, which I'll show you. But what I can do is I can highlight a section that I want to change and you'll see that it brings up some options for me here. So I can click to explain. If it's written something that I don't understand and I wanted to explain that to me further in the chat on the left, then I can click Explain. So I don't know what it's going to come up with, but I've just highlighted that headline and click the Explain button and it's just. Let me break down the key points of tax optimization strategies for you. The main goal of tax optimization is legally reduce your tax burden while maintaining full compliance. And then it's given me a bit of a breakdown. So it's just explaining what, what that headline means and what it's going on. Talk about. So that's if you wanted to expand on something, you know, if you were using this to, you know, create an educational text or a training manual or something, and if you were learning something at the same time, then you could just highlight a bit and say, explain this to me because I don't really get it, and learn as you go. Now, alternatively, if I want to edit this title, I can just click improve and Then it opens up a box where I can describe what I would like to update. So, you know, maybe I want to incorporate emojis into the title. Maybe I want to have a longer title. Maybe I want it to just give me a couple of different options of titles so that I can choose which one I prefer. So I'm just going to put.

Give me some options for different titles and see what it does. Now imagine it will give me those in the left hand side. Yep. So on the left hand side it's given me some suggestions of different titles. But you can see there that where it's got this little line. It has took the, the previous title into its consideration. Now if I just went and said I like this one, the Dental Practice Owner's Guide to Tax Planning, I'm going to copy that and say change the title to this and then just paste that. You'll see that it will edit the document on the right hand side when it's got to do it. There we go. So it changes that on the right hand side really quickly. And you can do that with the whole document. You know, you can say I don't like any of this, redo it, start again. Or you can say change the tone of this and make it friendlier, make it a lower reading grade, make it longer. You know, maybe I want to shorten the blog post, shorten the entire blog post to around 500 words. You know, maybe it's a bit too long for me. That's going to edit the entire document really quickly, as we'll see.

And now that I've got a completely different version to what I was working with previously.

This is a completely different version. I can go back and see the previous one by clicking down here on the bottom where it says version 3 of 3. So I changed the title. So that made me a version and then I've changed the whole thing, which made me a third version I can go back through and say, oh, actually I prefer this section here. So I'm going to take that part and what I would generally do here is I would copy this information and put it onto a Word document and then tweak it and edit it however you want. Or if you wanted to copy and paste this straight into the area where you are posting your blogs, you've got a WordPress or wherever it is you're posting your blogs, then you can do that. And you would do that by down here at the bottom it says copy contents. So you can copy that. You can also download it as a file. Now it does download it as a bit of a weird file. So you do have to go and then change the file or choose an application that you want to open it. I think it's a. Txt file that it does it as. So you do get text. But honestly I would just rather just copy and copy it to a clipboard and then paste it into a Word document. That's the easiest way to go about it, I think.

So that's how that looks.

Just think if there's anything else I want to show you with that. Generally that's how it works.

Let's look at Canvas then and see how that differs. I'm just going to grab the same prompt so that we can see it do exactly the same thing. Now, if you'll remember, I didn't tell it to use artifacts, it just did it for me. So if I just paste this in here, we'll see what it does. Because sometimes it will. Okay, yeah, sometimes it will open Canvas and sometimes it won't. So this time it has now.

Same kind of layout. You can see I've got my conversation on the left hand side, I've got my document that I'm working on on the right hand side and again I've got the, the document link here on the left. So if I was to make multiple different documents in the same conversation, I can scroll through here and find the different documents that I'm working on. It's making a really long blog post for me. I mean, I did say in the, in the prompt, I did say I wanted a lengthy blog post and it's very lengthy. I'm just going to stop that just so we can, we can carry on.

And I'll find a different one that I already made.

So here we go.

So you can see that it's made lots of different content. If I wanted to edit this document on the right hand side, I can just click and change exactly what I want. So this title here on the right, this is a different one. This is just some website copy. I can click this button here and make that big heading. I can.

Highlight this bit and make that bold or italic and so on. I can edit it so I can say caring for pets and their clients. I just, for just for demonstration purposes, I can just delete that part and I can type it back in. So I can, I can delete things. I can type in whatever I want. I can add extra spaces if I want that to be further dropped down and so on. So I can edit that almost as if it was a Document. You know, there's some things you can't do. You can't really do underline or strikethroughs or things like that, but generally you can do what you need to do with it.

Again with this one, this does have more features. So on the little arrow here on the, on the right, the little pencil icon, these are all the extra features that I have for ChatGPT for Canvas. So there's a button that allows me to add emojis everywhere, which, if that's what you want to do, then you can do that. If you may be creating social media content and you know that emojis are really powerful in that format, then you can do that. But if you click that button, it's going to add emojis everywhere. So we can add it to individual words, we can add it to sections by lists or remove them entirely. So let's just do it as sections just to see how, how that kind of ends up looking.

And hopefully you notice straight away how much slower this is than, than Claude. So it's still doing it. It's still doing it at a rate which is quicker than I could write it myself, but it is a lot slower.

So if you're wanting speed, Claude is the one to go with. But you can see it's just adding emojis just to the ends of each of those sections. So if I'd have put words, I think it would have put it, you know, throughout all of the paragraphs entirely. So it would have put done a lot more.

If I decide I don't like that, then up at the top I have a previous version. I also have a show to the different versions so I can show the changes that have been made. And then, I mean, it's hard to see with the emojis, but it, it has highlighted that with a kind of green highlight just to say that that's what has changed. If I didn't like that and I just want to revert that, then I will click that arrow to go back and it says you're viewing a previous version. Do you want to restore this version? I go back to the latest version, so I'm going to restore that version and now it's removed the emojis for me. So on the rest of these buttons I can add a final polish. If you're drafting something and it is a first draft and you want to give it more, you know, a final kind of judge up at the end and just make it a little bit more fancy at the end, you can do that. And add a final polish and it will just, just redo the whole document and just give some final tweaks to it. You can change the reading level. So if you're creating something for your clients and you want it to be technical, jargon, heavy or com, you know, serious professional vocabulary, then you can change the reading level here. And it is a toggle. So we can go up to the top to graduate school or we can do college or high school, or I can leave it the same if you want to go the other way and make the reading more accessible, make it easier to consume.

Lower the reading level. Because we know that generally speaking people are reading at a level of a, a 10 year old, 10 to 12 year old, 7th or 8th grader I think it is. But I'm, I'm English, so I don't know what is in American. The, the age generally of around about a 10 or a 12 year old. That's the, that's the way that we generally consume. So when there's a lot of big vocabulary, even if you're a really intelligent person, it's harder to consume that text. So if you wanted to just make it simpler, you can just drag it down to middle school, for example, and click that button and it will just change the vocabulary and the tone of that as well. So again, this is going to be quite, quite slow to do it. So it's saying now we love helping pet businesses with their money and taxes rather than whatever it said previously about. We offer these accounting services and payroll and bookkeeping. It's just really simplified that wording so that it is very child friendly. So you know, play around with that. Pick and choose the level that you want because it is a professional service. You're probably going to want a slightly higher reading level so that it's, it's not, you know, really like childish.

Okay, I'll just wait till it's finished that. Then we've also got here adjust the length so you can just use this toggle to go longer or shorter. You know, these are all here in these buttons, which makes it really convenient. You can also just type it in on the chat on the left and say can you make it this length? Can you make it around about a thousand words or something like that? If you want something specific then you can do that. You can also suggest edits here. So if I click this button, it's going to find areas of the, of the text that it thinks you could maybe update and it's just going to give you some suggestions of how you might update those. So I think it finds you five as a standard. Yeah, so it's found me five. So this sentence here, if I just click on the little icon on the right, that it's. That it's brought, it says, consider simplifying this sentence for a smoother flow, such as we handle the numbers so pet professionals can focus on caring for animals. So it's given me a suggestion of how I might reframe that sentence. And if I was. If I liked what it said, and I think, hey, actually that's a really good idea, I can just hit the apply button and it will make that change for me without me having to do anything else. So it's going to suggest its own critique and then carry that out as well.

So you can copy the text. The button up here is on the top. So if you're happy with it, you just copy all of that and put it into a document and off you go with that. And you can go and edit it on a different level. See, I like to generally get it 90 done on canvas or, or Claude.

Artifacts and then put that into a document and then make my final tweaks in there just because it's not perfect. And sometimes you just want to add your own flair to things as well. So that's kind of what that looks like. Let's look at a couple of different examples of things then that you could potentially do with, with Canvas or artifacts. So.

The one that I'd shown you, there was some website copy, I believe.

We'Ve got also here a SWOT analysis. So let me just go back up to the top where I put this in. So I did tell it to use canvas. And by the way, the way that you would do that, let me just go to a new chat. The way that you would access Canvas. So it used to be in the model section here when they were testing it, but it's now in the what I would kind of consider the tool section rather than the model section. So it's on your conversation starting area here, just where the three dots are. This is where your tools are. So you can make a picture or you can click Canvas and use Canvas there. So if I decidedly want to use Canvas and I want to make sure it does it, and I don't want to risk me putting the chat in it, not using it, and then I have to do the chat again and use an extra message, then I'll just make sure I select Canvas at the start and I can do it that way. And. But for anybody who Likes these little short shortcuts and little shorthand things. If you put a forward slash, it then brings up the tools that you can use as well. So I can click, I can just type in Canvas there and it does exactly the same thing. So that's how I would, that's how I would access that. Now let's go back to the SWOT analysis.

So I just, I've got a pre made prompt that I put in here to get it to do some SWOT analysis and it's then created a, a canvas for me and as you'll see here in the conversation, similar to how it had it in Artifacts, if you're not in the canvas area, this is what it looks like. So it just shows you the start of the document and there's a little icon here on the right that just allows you to need to open that up and expand it and open it into canvas. And now we're here in Canvas. So it's on a SWOT analysis for me. And now this works in different ways with Cat, with Jack bt, you can give it a URL to a website and say do a SWOT analysis on this website because it has that search feature. Claude does not have Internet access. So if you wanted to do something like this, a SWOT analysis, you would need to upload the business plan or some information in a separate document along with that prompt. So you can still do it. It's just a slight different way and obviously it's going to get different results when you do that either anyway. Because if you did it where you told it to go and look at somebody's website. Whereas if you have a written out business plan that might be a better way to do it anyway. But you can upload a file on ChatGPT as well anyway. So the SWOT analysis we've got now this is just a random ideal client profile that I've made, which I'll show you why I've made that. When we do projects in a minute and it's made, it's, it's highlighted the strengths, the weaknesses. So we've got geographical limitations, high service fees for small businesses. This, this.

This example is for some an accountant that works in the pet industry. So dependence on client, tech adoption, limited brand awareness, scalability challenges. We've got opportunities for expansion beyond Boston because that's where it's based, growth in the pet industry, strategic partnerships, technology upselling, industry thought leadership, etc and then we've got threats. So it's made all of this with copy and paste, that prompt and put it straight in there. And I grabbed that prompt, by the way, from our prompt library that we have in the value pricing academy, where we have, I think something like 50 prompts just ready to go. You just click on copy that and paste it in here. So that's how I got this one. Now, the SWOT analysis I also believe I've done on.

Claude. So let me just find the right conversation. Here it is. So with this one, I had to upload the document at the same time that I gave it the prompt in order for it to do it now here as well. It didn't immediately put that in artifacts, so I had to just give it an extra prompt to say, can you use artifacts for that, please? And I had to upload the document. So I'm sure hopefully if you are, you've been using these before, you'll know how to do that. If I just start a new chat, let me copy and paste that SWAT prompt, which is this one here, and I'll show you how quick it is as well, because it's really quick. Lord. So if I start a new chat and I paste in there, I can then just click the paper icon here, the paper, click icon and add that business plan to it. As well. So that's, that's what I did. And whilst I'm here, you can also connect your Google Drive to Claude if you want to, so that you can directly send your documents straight from there rather than having to go and dig around in your files to upload it every time. ChatGPT, you can connect either your Google Drive or your Microsoft account as well.

So let's go look at that conversation and you'll see the differences. So I can open that again. Here it is on the right hand side. Now it's significantly longer and it, it does make it really a lot quicker than the other ones as well, I would say so. And it's giving me a nice neat little conclusion as well. So, you know, they both work in a very similar way. So let's look then at projects. Let me just come back to my slides here so that we can look at the differences between projects. So how Claude has artifacts and ChatGPT has canvas, but they're essentially the same tool, but they've got a different name in ChatGPT, they're both called projects and they do very similar things again as well. So let's just look at some of the subtle differences. So with ChatGPT, like we've just mentioned, you can connect your Google Drive or your Microsoft account to automatically upload documents. The projects that you create exist in the sidebar and you can click and drag conversations into your project as long as you haven't used a custom GPT in that conversation. So in essence you can retrospectively organize your conversations. Whereas with Claude you need to make sure that you choose the product that the project before you make the the conversation because you can't click and drag. So with Claude you can connect Google Drive, but it doesn't have Microsoft. Only the starred projects sit in the sidebar, which I personally prefer because I think there's a lot, I have a lot of projects in, in ChatGPT and they start to take up a lot of room on the sidebar. So if you have a project that you're currently working on in Claude, you can star it and it will sit in the left hand menu. If you want all your projects to sit there, just star them all and they'll all sit there in your sidebar.

You can make, you must, you must make the conversation within the project include for it to be part of that project. Like I said, you can't click and drag. But the great thing is that when you start making a conversation in Claude, you can choose which project you want it to attach to right from, from that space. So I'll just hop straight back over and show you that in Claude.

Which I need this one for. So on the left hand side I've got my projects and projects by the way, I haven't really explained what that does. So a project essentially is like a dedicated workspace where you can organize your conversations, kind of like a folder for organizing your conversations. But you can go further than that because what you can do with a project is you can upload documents related to that project so that it has a core base of knowledge for what it's working on, related to that, that subject. So when you just start a brand new conversation with Claude, the knowledge that it draws from is everything it's been trained on in order to be what it is today. And same with ChatGPT, it's been trained on trillions and trillions of documents, of web pages and information. That's how it generates that answer for you by having all that pre trained knowledge. When you have a project, you're giving it documentation about your business or you know, if you want to personal use, any other information. Generally you're giving it information about your business that you're happy to share, you know, ideal client profile, the, the services that you provide, an FAQs document, that kind of information. So that when you ask start a new chat it can draw on that information to answer your questions. So you're giving it a base of knowledge so that it can use that to answer your questions so that the, the conversations that you're having are really nuanced and contextual to your business and what you're trying to create. So hopefully that makes sense. So if I just click into projects here, they're all listed, this is on Claude, all my projects are listed. I can create, create a new project over here on the right and I just give it a name and a description. So let me just make a test one just for the sake of this. I can create a project and now over on the right hand side I can add content and this is where I would up either add my Google Drive and bring the document straight across like that or I can upload a device or I can just add contextual information just by typing it in. So if you've got a document of your idle client profile or something like that that you want to make sure it's got that access to, then you can upload it at this point and it will draw on that information. You can also give it some limited project instructions. So when we look at custom GPTs in ChatGPT you can give it really extensive instructions of what you want it to do. Whereas with this with projects you're more giving it kind of information on what format do you want it to respond in? You know, do I want US or UK English? Do I want small concise sentence structures? Do I want bullet points, you know, things like that? Do I want to make sure that it always uses artifacts? I can put that in there, always use artifacts. Just how do you want it to respond? What are the general things you want it to you to do? You can put that kind of information in your project instructions. So let me have a look at the one I made for today. So for this one what I have added to here is I've added my ideal client profile which I just got it to make a demonstration one for me. And this focuses on an accounting business in the pet industry. And then I also have the pet business Owner's Guide to Financial which is something that it created for me, which when I was making an ebook when I was playing around with it earlier. So I also have an ebook in here as well, a lead magnet. So it has that information pre stored in there and I can now have conversations with it with that knowledge in the background. So if I go to ask it to create me some marketing Material. And this is how prompt engineering is really important to know. And the better you can write your prompts, that is a skill set in itself. And you will get much better responses. When you can write really great prompts, however, you can sidestep that when you do things like this and you upload all that prior knowledge, your prompts can be so much more simpler because it already has all of that context behind the scenes. So whereas normally if I was Getting Claude or ChatGPT to write marketing material, I'd be following that structure of I'd give it a Persona, I'd give it the task, I'd give it some context, I'd tell it the audience I'm trying to reach, I would tell what kind of format I want, that you know, the, the tone and the language. I would give it all of that information in my prompt because I've already uploaded that behind the scenes and you know, I've uploaded an ideal client profile. Maybe you also write your writing style guide so that it knows how you like things to be written. Examples as well. The prompt that I put in here can be very simplistic and it will still get me a great answer because it's got all that context behind the scenes. So all I've said with that context in mind, can you create me some marketing material to promote our services? And it's gone away and it's opened an artifact here as well. And it's kind of made me almost like a long form email, I guess, or wording that you could put on a website or a long Facebook post here. Because I didn't describe, I didn't say, you know, whether I wanted it Facebook or Instagram or anything like that. So it's just made me marketing copy which is really good. Transform your pet business with poor perfect accounting. So this is the, the ideal client profile that, that we've uploaded. So it knows that it's based in the pet industry. So all of the stuff that it's talking about is pet industry specific. Perfect for independent groomers and single location retailers, perfect for veterinary practices, perfect for. This one's designed for veterinary hospitals and pet product manufacturers. So it's told a little bit about different packages. It's created for me and told me who that's going to be suited for based on my pet industry, you know, niche that I have. And it's given me all that information which is, which is really good. It's put a client success story in there for me, which is fake. But you know, maybe you can take that as A template and say, well, there's this client success story section here, I'll just put my own in. And it's made me up some, some information here. But of course this is going to be specific to your business. So when you upload your ideal client profile, it's going to have all of your information there and in your niche in your industry and that sort of stuff.

So that's an example that you might use.

A project for. That main thing being that you can upload the information here behind the scenes and then get it to do that. So with Claude, I mentioned earlier how you have to choose the. The project that it's going in before you start. So either you come into your project. So if I click here, claude project, I then choose which one I want to want this to be a part of and I start my conversation with from here. But alternatively what I can do is I can start a new chat. It's got something in there, but I can start a new chat. And then just below that it says here, use a project. So I can click which project I want it to use and it's that one that I'm doing today, the live stream on February 6th. So that's the project that I want it to be added to when I have this conversation. So if you want it to be part of your project, you need to make sure that when you start that conversation, you have chosen that project, either by going into the project and starting a conversation from there, or starting it on this main page and selecting which projects you want it to go into. Because there is no retroactive ability to add that conversation into the project at a later date. So you need to make sure you choose that first. Now with ChatGPT, let's look at the projects on here.

So also on Claude, as you see, can see only the ones that I've got starred, because when I go into projects, I have loads of them here, but I only have the starred ones on my.

In my sidebar. Whereas in Canvas it, it puts a lot of them in here. Now it does have the limit where it shows you five and then I have to click the button here to see the rest of them. So it does have that limitation. But this is the only area where you can view all of your projects in one place. So where Claude has this page where they're all listed, there's. There is no page like that on ChatGPT as of right now. They might improve that. So all my projects are listed here in the sidebar. So if you ended up having Lots and lots and lots of different projects, then you might start getting a bit overwhelmed. So let's see how that works on here. Just visually how it's slightly different. You can see that on this one I've got color coordinated folders which you might find quite a nice little feature. When I go into my project here and I click this folder at the top, that's where I can change the color for it. So I can have them themed based on different things that I'm working on in my business. You know, maybe I have marketing themed ones, maybe I have ones that are around creating service packages, ones that are around client profiles that I upload information about my client's personality and such so that I can then ask questions about my clients within that project. That's an interesting way of using them. But this is what it looks like. So I can upload my project files here and again I just go add file and I add it from there I can also connect again my Google account or my Microsoft account and I can add instructions again in that similar way where it's just going to be what is my response format, what is my.

Keep, Keep my answers short and sweet, use bullet points where you can, that kind of thing. I can give it that, that kind of information there as well. And then all my chats are listed just below here as well. These are all the chats that I have within this project. Now the great thing about ChatGPT is that you can retrospectively add conversations into those folders just by clicking and dragging. So this one here, this pet business, this text tax strategy, I just want to add that in. I just click and drag and it's now in my project. So if you have used a custom GPT in your chat, you can't do it because currently you can't use custom GPTs within a project. So there's that. But if you, if you just had a regular conversation or it was a conversation that used Canvas, then you can click and drag and organize your conversations afterwards. And I think similarly you can, oops, you can view those conversations from this area just by clicking on the folder. It expands it and you can see at least some of those conversations. You know, if you had 20 conversations in there, you won't see them all. You'd have to come to here, but you can just at a quick glance see which one you were working on. So that's the one that I just organized away into this, into this section here.

So that's how, that's how projects work. Again, lots of different Things you can do with that?

Oh, just to show you on Canvas as well. You can make tables in Canvas so you can ask it. I asked it here for a content creation calendar and it gave me the. The tables lined out like this. Now again, with ChatGPT you can edit this so I can, I can just erase that and change it how I want it on that table.

Whereas in Claude let me find the conversation that has that in. So Claude does do the table and I actually kind of like that it can do a lot, a lot. A longer table width wise because it can put all these different columns into it. So hopefully you can see that I'm sliding across there. But you can't immediately edit it by clicking on it and editing it. So you do have to tell it in the left hand side. Can you change this section, this column to this, you know, in the cash flow management column on March the first week, can you change that too? So it is a bit more of a faff.

I just think it looks a little bit nicer. I'm a very visual person.

So that's another thing that you can do is you can create tables and stuff in your Canvas Canvas account as well and then yeah, just organize all of that into your projects in terms of which one's better. With projects, again, it's very much a preferential thing. I like that you can color coordinate these. I like the layout of them. I do also like that Claude has them all neat. So I can just have my starred ones. But I also have a page here where I have them all. It's also kind of annoying that I can't retrospectively effectively add conversations into the project. I have to make sure I've chosen that beforehand.

But you do have the, the really handy thing where it's just there and you can click it straight from from where you're doing it. So but there you go, there you have it. So that is.

Canvas, artifacts and projects for ChatGPT and Claude, which you prefer is kind of up to you. I would say play around with both of them if you can.

Artifact Canvas is available on all three plans of ChatGPT and artifacts are not sure. I think artifacts you do have to be on the paid plan to be able to access that. And projects you have to be on the paid plan as well. I did have some just information about the size and the amount of files that you can add.

So let me just show you that as well. So chatgpt, with your projects you can upload around 10 files. Again, it's. It's a bit fuzzy because it depends on how many, how much information is in each file. But it's 512 megabytes per file. The input is 128000 tokens, which is around 96000 words, which is still a lot. It's like 320 pages of text. Not as much as Claude though. Claude is has always been elite with the amount of context window that it has. It uses GPT4O or 4O mini.

For the projects and then for. It's only available on plus plan or higher. This is for projects and then Claude. The There is no apparent limit on the amount of. On the amount of files that you can have. Rather what it means it goes based on is its context window and its context window is 200,000 tokens. So.

You know that's around. It's like it's times 0.75. So it's around 150,000 words that you can input behind the scenes for it to have that documentation behind the scenes. But then the input when you're having a conversation with Claude is massive. You can put 1 million, 1 million tokens, which is around 750,000 words, which is, you know, a couple of novels you can put in there when you, when you're talking back and forth with it. So you've got a lot of information that you can add to Claude. It's got a much bigger context window for that. It does use Sonnet 3.5 which has been ranked lower than the models in ChatGPT. So 01 which you can use in Canvas is definitely the highest. Although ChatGPT has just released O3 so if you want to go and have a play around with that again, if you're on the paid version, it's uploaded O3 which is its latest, you know, flagship model that you can use. And again Claude is Claude Projects is only available on the Pro plan which currently is 18 per month and ChatGPT is 20 per month if you want to use projects. Okay, just have a quick check, see there's no questions. I don't see any questions. So that's everything that I have for you guys today. Hopefully that's been good. If you want to attend the next live stream, Mark is doing one in two weeks time. Not sure what his content is. It will be, I'm sure chatgpt or large language model related. So if that's something that is is of interest to you, then just put that date in your diary. Remember the Profitable Bookkeeping conference is on February 25th. To 27th. So if you're interested in learning stuff about that, then you're welcome to join that as well. The tickets are free. You can go and check that out. Well, that's everything I have for you guys today, so I will play some music to see you out. Thanks for all the newbies that came. I hope you enjoyed it, and I'll see you on another one soon. Cheers, guys. Have a great week.