4. Open AI vs Anthropic (GPT vs Claude) - SCAI017
Sloup
1 follower
10 Video Views·Sep 15, 2025
Nate Herk - AI Automation Society Plus\2. Intro to AI Automation\1. Start Here\4. Open AI vs Anthropic (GPT vs Claude) - SCAI017
Okay, so I've definitely got a lot of questions about, you know, in my videos I always pretty
much use OpenAI 4.0. I've got a lot of questions about different models and when to use which ones
and then also specifically within OpenAI GPTs like 4.0 versus 4.0 mini, stuff like that. So
I wanted to start off by just talking about, I guess, the difference in my mind about OpenAI's
GPTs and CLOD and sort of when to use each one. So, you know, it's no doubt these are kind of
two most standout models in today's AI landscape, pretty much OpenAI's Chad GPT or Anthropic's
CLOD. And actually, I don't know why I didn't put this to Anthropic, but so we'll do that real
quick. But anyways, Anthropic's CLOD and OpenAI's Chad GPT. So when to use which one, for what
projects, that sort of thing. So these are the main players. Let's talk about first Chad GPT's
advantages. So it's got very powerful tools. It comes with obviously specialized tools within it
like data analysis and image generation. And then also it has like browsing capabilities too.
So super powerful. And then also just in general with, you know, their custom GPTs,
you have other custom GPTs that you can share with others even if, you know, like they're free.
There's kind of like a resource library of exploring these that are super cool. You've got
prompt presets. So also with the GPTs, you can add preset things. So just really fine tuned to what a
user is specifically looking for. Also, there are a ton of external integrations. So OpenAI
supports, you know, connections from with APIs to trigger different actions to save data to external
platforms. And, you know, obviously integrations are super powerful. And then mobile access too.
It's easy to, you know, pick up OpenAI on your mobile devices. So it's great for on the go use.
And the best case is when you need something repeatable with an AI assistant or a solution
that others can easily use and share like GPTs. So some of the disadvantages now I'd say
are that with specific tone and style, some people definitely prefer Claude's sort of writing style,
which can feel more natural and less AI generated, less robotic. And then there are some
context limitations with GPT. Their window may limit how much information you can upload.
And, you know, for documents over, you know, 20 something pages or something like that,
you may lose information, even though you might not realize you did. So that's not good.
And then let's just move on to Claude. So you have multiple chat sessions. So you can,
you know, keep multiple projects organized, multiple chats within a single project. So
it's good for like ongoing, larger work, ongoing projects. They've got I think they're
Claude's definitely better with code. They have their model excels in code generation and sort of
the problem solving aspect of writing code, especially for, like I said, larger projects
that are ongoing. And then it has a larger context window length. So it can process larger documents,
it can just larger, larger characters as far as text. And I think this is all these things have
really just kind of added up to the fact that Claude is going to be better for extensive,
you know, analysis, larger projects, stuff like that. So multiple stages, when you need consistent
handling of your data and your documents, and very organized and stuff like that.
And then with its disadvantages, you could talk about limited tooling, which is more like web
browsing, or sorry, lacks tools like web browsing, and sort of the, like image generation aspects
that the GPT is offer really well. And then their integrations aren't as aren't as easy to connect to
or you can't even connect to with external APIs, and stuff like that. But um, so when would you
use GPT is, you know, if you want to share solutions, if you want repetitive tasks, if you
want assistance and tutors, and obviously, that's why, you know, this video isn't super tailored
towards which model to choose for your agents exactly. But that is usually why I go towards
the OpenAI 4.0 and my agents with their ability to, you know, be personalized for, you know,
set them up, set them up for what kind of role they're taking on their background, good for,
you know, personalized assistance, because they're very flexible and their data integration options.
So that's why I pretty much always use 4.0. I can definitely get into more of the details about why
or when I would choose 4.0 mini, 3.5, all that kind of stuff. But I really like 4.0. And then
when you'd want to use Cloud in certain projects, obviously, larger documents, like we said,
code projects, it's really good at coding. So if you had some sort of agent where you wanted to
help, you wanted it to help you write specific code for certain tools or whatever, that would
be a case where you maybe want to use Cloud in there. And yeah, also multiple chat threads,
of course. So the good thing is obviously both models are super powerful, and they definitely
do have their specific use cases. And, you know, the best way really is just to get in there and
play around. Because like I said, with the whole writing style thing, some people just prefer the
way that Cloud comes out better than the GPTs. But, you know, you can do, you can still personalize
both of them, even though I think the GPT is better at that. Cloud lets you add custom instructions as
well. You know, you can help it tailor responses, and it can be a more personalized experience,
which is obviously helpful for AI in general, because you need it to fit a specific use case,
a specific role, because every project is super different. And then, yeah, when it comes to
choosing between, first of all, I would say just play around with both, see what you see what you
think about, you know, run it through the exact same situation and see the two outputs you get.
But obviously, consider using GPTs for something more versatile and sharing and structuring your
So just wanted to throw down some super quick thoughts. I hope that this was at least helpful.
Hopefully, it was at least a good introduction into exploring how different models are structured
and how different models can be utilized in different ways. But yeah, like I said,
it's one of those things similar to the embeddings where I feel like I don't completely understand
every aspect of every model of every single different type of large language model.
But what you do need to understand is the high level use case of here, use case of this one.
And in general, if you understand them enough to be able to use them and get your agents to
work the way you want them to, over time, you're going to start to just understand them better.
But the most important thing is getting enough understanding to have your agents run the way
you want them to because at the end of the day, that's more what I care about. But yeah, thanks,
guys.
