Why & When You Should be Using Claude over ChatGPT

[@kennyphan9612]

to save you up some time: Based on the transcript, here are some key reasons to use Claude over ChatGPT: 1. Better tone and style: Claude's responses are generally considered more human-like and fluid in writing style. 2. Superior coding abilities: Claude (especially the Claude 3.5 Sonnet model) is widely regarded as better at coding tasks compared to ChatGPT. 3. Longer context window: Claude projects can handle more context in the knowledge base, allowing for more comprehensive understanding of larger documents or multiple files. 4. Multiple chats in one interface: Claude projects allow you to have multiple related conversations within a single project, which is useful for complex or multi-faceted tasks. 5. Better for project-based work: Claude projects are well-suited for tasks that involve multiple steps or evolving requirements, especially writing and coding projects. 6. More exploratory: Claude projects are better for open-ended tasks where you're not sure exactly what direction the work will take. 7. No need for prompt presets: Claude doesn't rely on conversation starters, making it more flexible for various types of interactions. 8. Better for one-off projects: Claude projects are ideal for tasks that you'll complete and then archive, rather than recurring use cases. While ChatGPT has some advantages in terms of additional tools (like image generation and web browsing) and shareability, Claude is often preferred for its core language and coding capabilities, especially in project-based scenarios.

[@PuppetMasterdaath144]

Lol chat gpt is like a regular smart human, and claude is like me

[@ieatdogs666]

Yeah Claude fans are talking only about coding and never other stuff. Pretty funny.

[@Ignoranous]

You wrote this with AI huh

[@kennyphan9612]

@@Ignoranous yeah? is there a problem?

[@kennyphan9612]

@@Ignoranous I mean it's so obvious that I wouldn't write all this

[@joshuahaleio]

I signed up to Claude pro version yesterday never thought I would cheat on chat GPT but I did

[@FilmFactry]

My Comparisons of all three LLMs, regarding legal questions. Claude 3.5 was superior to OpenAI and Gemini.

[@pigeon_official]

I pay for both because ChatGPT is still better at many things

[@babyqueen2030]

Same. 🤦🏽‍♀️

[@drlordbasil]

It's worth it! xD

[@rupertllavore1731]

So how was it?

[@nickvilliers5216]

I agree. I think the features of claude are a bit better. However, Claude's limits are just ridiculous. If you use a lot of tokens, you get only 10 or 15 chats before you have to wait for the 5-hour cooldown. Also, when programming, there is no way to continue the response after the message limits triggers. I code stuff and this isn't very pleasant when working with substantial code snippets.

[@IdPreferNot1]

This is my concern. At this point in the paid OAI, i never run into limiting on my test, paste error, regenerate coding workflow as a GPT dependent coder... and it doesn't burn my API credits either. Could probably use the longer context for when gpt loses its mind well into a coding session, but assume those long context prompts burn your access quickly. Do you have a perspective on the relative difference as i'm probably looking to drop the paid, keep my funded API account, but move to an Anthropic plan.

[@paulyflynn]

I am creating Claude projects with smaller, focused set of files. Only asking code changes. Then I use OpenGPT for any other side questions or tangents.

[@hendrikbonthuys9190]

@@nickvilliers5216 agreed we do need better limits, especially on the paid plans I do find myself hitting limits faster than with paid ChatGPT. What I do find is very useful is after doing quite a bit of coding and before hitting the limits, I’ll add the artifacts to a project. That way I can keep going in a new session with a project where I loaded the artifacts into once the limits go away but Claude still retains context of what we were working on.

[@play365alltime]

That’s the main reason why I switch back to ChatGPT. Very disappointing. Most KZbinrs never mention this. Thankfullt it’s being paid month by month. Next month I am going back to ChatGPT 4o

[@saleesh0089]

Very true

[@karhadd]

Summary: Optimal Use Cases: ChatGPT: Best for specific, repetitive tasks and sharing customizable AI tools. Claude: Ideal for extensive, project-based tasks requiring a broader context or more human-like interaction.

[@aiadvantage]

2 updates since the recording: - Claude added a limited data analysis tool - You can now share artifacts (not projects (except with team on teams plan)

[@ben_kendall]

Thanks Igor, I will be coming back to this video for at least a few days, it is packed with good information.

[@GoodBaleadaMusic]

One way to cut down on claude usage is to use it to give suggestions rather than change things. And then you bring that back to chat GTP. And then soon your chat GTP comes pretty close to Claude. But having them compete against each other like that is the best way to create anything great.

[@Ben_D.]

'GPT-4o works *for* you and Claude Sonnet works *with* you' So well spoken and really helps people pick. Like a Sorting Hat at Hogwarts, you know where your affiliation lies.

[@razorbackroar]

Man no one could have said this better my friend. Pin this man’s comment.

[@ekot0419]

I got Claude 3.5 Subscription. Thats because I really pushed Chatgpt. And i finally sees its flaws. It SUCKS processing math. Basically it is like a genious with both ADD and dementia. I did not use Claude to code yet. I got Claude to process a large amount of Data. Still have errors. But it catches it when I remind it once. Chatgpt... I know that it is not alive. okay. But I yelled and screamed at it like a Boss yelling at their staff. lol Now I have to keep both. Since I cannot remove Chatgpt as it is the swiss army knife to everything else.

[@abbdiego]

GREAT VIDEO , i watched it and didnt realize how fast time went

[@gabrieldvorak7670]

“Absolutely fascinating breakdown of why and when to choose Claude over ChatGPT! Your insights into the differences between the two models really clarify their strengths and use cases. It’s great to see such a detailed comparison that can help users make informed decisions based on their specific needs. Keep up the amazing work in demystifying AI for everyone!” This is an ai commentary response.

[@sangiliboopathi8203]

What is good for RAG Claude or GPT?

[@MichaelBoyens]

Great video and sums up the differences very well. I find Claude Projects on a Pro plan are great for exploring and brainstorming, but limits get in the way most days if the project is complex or has a lot of context. I found GPTs more accurate with large context (40,000 words) than Claude, IF the document is well structured and the prompt well formulated, and more importantly had zero issues with limits (I'm on a GPT Teams account)

[@marcosantiago6818]

I'm a big fan of Claud Projects. I use it for my creative as well as my technical endeavors. It all depends on the problems I'm trying to solve. As an IT Project Manager and Business Analysts I used to do a lot of coding, but not as much as I use to. I regularly use it to generate very sophisticated database code database queries and stored procedures. I do this by exploring SQL Sever database tables and schema definitions and saving them as text in a very specific project. This allows me to use natural language to create very complex queries and or scripts or stored procedures. As a screenwriter, I use Claude projects to maintain persistent and evolving information on story outlines, character bios, story development documents, and my screenplay as I develop it. In this context, I provide Claude with as much structure information as possible to build a very robust context space. Unfortunately, Chat GPT cannot do this as well, even if you use GPTs. But I do like to use GPT for other sophisticated tasks that are not as context-driven.

[@merveilmeok2416]

👍🔝

[@Sekhmmett]

Claude refuses simple things and initiates lecturing about safety for anything it misunderstood

[@namikazedevj46]

mine refuses to answer test question and said if that was academic dishonesty

[@blackgptinfo]

I use both but for business writing and research claude is so much a head it is my preferred tool

[@PaulVoorberg]

8:46 - Writing with Claude using projects

[@hendrikbonthuys9190]

Nice comparison but doesn’t Claude now do data analysis and allow sharing of projects and artifacts? If I load csv files and PDFs etc it does create charts and graphs and do analysis. And I can see an option to ‘Publish’ artifacts from Claude as well, which allows other users to see them and ‘remix’ them.

[@cait7822]

Thank you! Can we have a video on custom instructions in projects? Like a use case deep dive for making something come to life in a claude project using the key features.

[@kjierstenbartlett4866]

Which of these are better for searching papers/articles/journals to assist in research, please? TY

[@tratkotratkov126]

You can add the ChatGPT ability to understand and speak different languages as well the multimodal capabilities where you can just talk to the model and it can respond directly with voice

[@carpye]

What do I do to make my GPT chat look like yours?

[@ericgibson847]

Hi, can anyone tell me about the GPT database that is at 12:13? Is there a website that lets you browse the a bunch of different community created GPTs?

[@drlordbasil]

I think sonnet 3.5 is showcasing that they know how to tune keywords so to speak in the models mind to give it further tuning to coding, thats my guess at least.

[@MojaveHigh]

I write GPTs that use my custom APIs. I can only do that with ChatGPT. I do agree Claude's code generation is better and Artifacts is a great UI tool, so I tend to use Claude for some basic function coding, until I run out of chats.

[@aiadvantage]

Thanks for sharing and yeah agreed the rate limits are super limiting once you get into the flow of it.

[@KK-bx1qm]

Do you mean custom GPT (not API)?

[@MojaveHigh]

@@KK-bx1qm Custom GPTs that use my custom APIs. I write both.

[@brianjanssens8020]

One thing i do not like about claude is that it out right refuses to do some things. Call me crazy but so far Chatgpt is the most based out of any AI model.

[@cbnewham5633]

With clever prompts you can get around most of this.

[@igor_in_theusa]

@@cbnewham5633 How?

[@jurgisvalancauskas4006]

Yes. Claude is designed with more ethical bias than chatgpt but I work around it

[@patrickdijkman8974]

I wanted to signup for your newsletter, but where is the CTA on you signup page? (using Firefox).

[@aiadvantage]

oh wow! thabks for bringing this up will look into it asap

[@razorbackroar]

Claude is so much better in so many ways. It’s clearer. Less prompt bull crap. I wrote code all day long all day long & i always have issues with custom gtps on the gtp store & with chatgtp 4o Claude does miracles

[@IstvanM10]

hi, can you tell me which ap you use to put yourself in different scenes on your thumbnails? thanks

[@aiadvantage]

Combo of stable diffusion dreambooth, photoshop and chatgpt. We have a detailed step by step guide in the community

[@IstvanM10]

@@aiadvantage thank you :)

[@reoroman]

what about this message in Claude when you try to upload a pptx file which you can't in CHATGPT: Conversation is 16206% over the length limit. Try replacing the attached file with smaller excerpts.

[@strategictechnologist]

Which one censors the training set you more?

[@Yipper64]

1:14 which Claude has in its artifacts feature. 1:31 which you can get via bing with less censorship, of course at the cost of not being able to adjust the aspect ratio. 1:40 I think this is like the one thing chat GPT has over Claude at the moment, but not by much given you can copy paste. 2:40 oh yeah that's a second thing it gets. Shame that chat GPT seems to have too much of a narrow context window to actually USE said API calls. 6:30 I would argue that artifacts can analyse data much better than chat GPT. and is it not the same thing? Taking the data and visualizing it with code? One thing I love about Claude is that its so good at it, it'll do it without prompting. As in, "Hey claude I have this thing I want to do, how do I go about it?" it starts programming a flow chart to VISUALIZE how you do it.

[@malikmoeeznawaz504]

Dear can you honestly recommend which one i have to buy. I want coding

[@aiadvantage]

Claude. Sonnet 3.5 is best in class for code generation.

[@HeyyyyMacarena]

Cant share GPT convos if there are any images in thread 🧐

[@djmpath971]

What do you guys think of Poe? It has them all...? Is it worth it? Is it missing any features?

[@happentobehere]

@2:07 Gpt v Claude ≈ Linux v Mac

[@mohamedkarim-p7j]

Thank for share

[@michaelcbond]

I consume a lot of content. I really appreciate the depth of your content. Thank you!

[@MohimmollaMohim]

The adaptability of Vematum to market changes is a key strength.

[@dylan_curious]

The folder of chats and better at code is the best for building an app. Otherwise I still like ChatGPT

[@aiadvantage]

Both have a place in my heart 🫶

[@the_micks]

I am developing one app, and I was using GPT to help me out, i tested Claude, and it's a lot better, like a lot.

[@trash_the_place]

I love your videos @aiadvantage but I really wish you limited the unnecessary cuts to stock footage, it's too distracting and takes away from your perfect delivery of the information. If you need to cut, cut to the UI of the services you are talking about.

[@aiadvantage]

Feedback received and I do agree with this. Will cut back a bit.

[@FilmFactry]

When will we be able to do a simple task like : Call 5 Honda Dealers. Ask for parts Department. As for the price for ATF Transmission Fluid. Is it in stock, Then report to you which has the best price.

[@Reese_PLAITE_Fitness_App]

Build it

[@sakibsk4623]

Vematum's consistent progress updates show dedication.

[@djpuplex]

I asked claude to educate me about the new prompting features yesterday. Claude had no idea. I asked perplexity and got a bullet rundown of the new features. I than asked claude how some of the features, Claude asked me wear I heard of these features I prompted perplexity and claude was asking me questions about the new features. 😂

[@tar-yy3ub]

People make this mistake all the time. How can a large language model know about its own features when it is trained only on the past? The only reason large language models know anything about themselves is because that has been injected in later by the developers

[@RyanDoserAI]

Claude 3.5 Sonnet > GPT-4o Claude Projects > Custom GPTs Anthropic > OpenAI

[@Urban_Surfaces_Adam]

Sharing a custom GPT with a free user seems to be severely limited. They get errors like "can't do more advanced data analysis right now".

[@DigitalBrooke]

I hear if you type “continue process” it will resume. But I may be incorrect.

[@michael.pedersen]

How can we prevent Claude shutting us down for three hours by saying we are exceeding their limit? Often times I’m not doing that much. It seems and now I’m waiting three hours. FYI, I am paying for their service.

[@GambillDataEngineering]

Great video! I've found that using sonnet that I get better synthetic dataset generation than with open ai.

[@CliffordJFisher]

What do you mean custom instructions in ChatGPT? Thank you very much what a great video!!👍

[@gankimagoliguerfiehanoi7146]

Claude also has a Mobile App

[@arxmechanica-robotics]

I like Claude much better. I've used both.

[@MyScienceWorld825]

Vematum's partnerships are a testament to its credibility.

[@TheDetour789]

I use multiple chats in ChatGPT simply by opening separate tabs.

[@DaisyZeki]

totally agree with u!

[@rashmirathore1707]

Vematum's leadership team is stacked with industry experts.

[@gabrieldvorak7670]

Here's mine... As crazy as it sounds, neuralink; will eliminate the "work with" vs "work for", analysis. I'm thinking a futuristic mind set, probably not plausible for our "Time" pshhh, jk, but; we will see this year. We need a Claude or somewhat similar idea "screen record" that records progress, learns from it, and bases our every move, without question. It creates the idea, we adjust or improve the idea, and chat or Claude improvises, based on mindset. This could be any software, like photoshop, blender, visual studio, etc... Basically a brand new computer you log into, which knows you; based on simple questionnaire. Any software you install, it knows. It creates based on your mindset. It creates with you, but for you. We are really only 5 years away. If we see it, dream it, or think it; it creates, based on that core imagination. I feel like that's the next stage.

[@cbnewham5633]

OpenAI Dall-e is no longer anywhere near the best. Flux Dev is superior and you can run it locally without censorship.

[@madanjyothimj2229]

Vematum's community-driven approach is what crypto needs.

[@elixier33]

I pay for Microsoft, co-pilot chat, GPT and Claude and so far, chat GPT is the only one that I'm going to be doing with. It's not perfect. But frankly, I think it's miles ahead of its competition.

[@shareclipsave]

If you are watching this just now, ChatGPT just released a Projects feature

[@redred333]

All of these tools are essentially useless except as curiosities until privacy can be ensured. This is critical for industries that value the exclusivity of information (which is essentially everything) for example healthcare, legal, science, military/law enforcement.

[@mikgol81]

Using Anthropic for coding is like having a genius sidekick-less chat, more code!

[@yhwhlungs]

Bro woke up and started recording

[@aiadvantage]

ahaha. very true!

[@stephendunning1510]

Been using GPT for two years I think, Just started using Claude a few days ago. Claude is WAY better in my opinion.

[@reallll2778]

Can you ask Claude random questions and have it look up studies for you?

[@RakefetBenAryeCohen]

So informative and helpful, thanks a lot!

[@anthonysmith1494]

I’m a long time RPG-er, been creating my own campaigns & GM-ing for 13 years now and “ It’s Killing me Smalls, it’s Killing me!!!” So I am looking into building an AI GameMaster. Gathering info on which model would be more geared towards my needs. And there’s a ton of info….. Watched a few of ur vids, u present info in basic straight forward manner. So, if u happen to read this…. I would like to get ur opinion on this idea. Ty.

[@AllisonRoadWest]

Whoa 😳

[@samblacksolvess]

Fire video

[@akust]

0:46 the fingers show that you are artificial 😄

[@randomdreamer1619]

If only gpt memory can remember pdf....

[@gudisevamsi]

This AI agent is a must-have for anyone looking to streamline their workflow. #SmythsOS AI delivers.

[@manitou__]

It is still "50 First Dates"

[@TheChromePoet]

The problem with claude projects is that when I have multiple conversations inside of a project, no conversation has any knowledge or awareness of another conversation. I feel like they should be aware of eachothers existence.

[@aiadvantage]

That might get messy super fast. Copy pasting anything relevant into a pdf that you upload to the knowledge base is the way to go. (Ideally there would be a button yo save a conversation to it)

[@Corteum]

There's no one-size-fits-all answer to this question. It's just something each person can find out for themselves, empirically, experimentally.

[@TapTapUrDone]

But I thought GPT's are still using GPT-4 not GPT-4o?

[@dustinsuburbia]

Mine use 4o

[@TapTapUrDone]

@@dustinsuburbia How can I check please?

[@aiadvantage]

they officially announced all gpts use 4o a few weeks back

[@sania1500]

The integration possibilities with Vematum are endless.

[@chapmansbg]

gpts can uplaod 20 docs claude 200 so 10x

[@aiadvantage]

True but it's mostly about its ability to retrieve it too. If you max out those 200 docs you will be loosing pieces o info left and right.

[@bronzemoontr]

claude seems so limited, after a few chat, limit ends... chatgtp seems more promising an better.

[@RomeTWguy]

Always

[@prachi-k2y]

The way SmythOS allows multiple AI agents to work together is a game changer! 🤖💡 It really showcases how collaborative AI can innovate so many processes! 🚀✨ What do you think?

[@saleesh0089]

Claude is limited and costly ⛔

[@30kk]

Please explain further

[@RealKalluaVlogs]

The more I dive into Vematum, the more bullish I become.

[@POVShotgun]

Idk it kept fucking up code for me. It just referenced so much no existent stuff

[@mikezio]

Bruh can u at least fix those finders holding the microphone 😂

[@randomstuff5520]

Chatgpt so worse. 20 redo and nothing feel right. Claude? Max 3 redo and got what I want. 0/10 Chatgpt for me

[@anubisai]

Oooh, bad timing.

[@ayanansari3370]

Vematum's vision for decentralization is inspiring.

[@Eldorado66]

Please buy yourself a mic stand

[@aiadvantage]

this the mobile setup

[@Eldorado66]

@@aiadvantage ok fair enough 😅👍

[@AayushJhagames]

Vematum has the potential to outperform many in the current market.

[@razorbackroar]

Stupid LOL

[@ryanclemons1]

Yet none of them can write a well made story lol.

For me GPT is winner, don't see value in Claude.

[@SorifulMolla-me4mz]

Vematum's responsiveness to community feedback is impressive.

[@ZcqS]

1st

[@scottd1342]

I have used both GPT4, 4o and Claude Sonnet for months. Most of my work requires solid writing with a human tone and style. Claude crushes chatGPT on this front. I ended my sub with chatGPT.

[@thoughb410]

I rely heavily on writing for my job. Do you have any tips to prevent Claude from repeating ideas or any tips to avoid its deficiencies?

[@scottd1342]

@@thoughb410 I haven't noticed it repeating ideas myself. Unfortunately, it does not always remember the instructions I give it in Custom Instructions such as words it should not use or to use Australian and not US spelling. But overall, I feel it writes in "normal" relatable English unlike chatGPT. Thoughts?

[@thoughb410]

@@scottd1342 I agree, Claude’s writing quality is significantly better than the more generic and robotic style of ChatGPT. I often write lengthy pieces to propose plans and ideas, which can lead to repetitive content as it reiterates points made earlier. I try to instruct it and brainstorm ideas to reduce redundancy. I believe we have different use cases. While it excels in certain features, it also has many deficiencies. I’ve experimented with various techniques but end up with similar results. The best approach is not to give it full control; always provide it with ideas and revise until we achieve the perfect version.

[@gabrieldvorak7670]

Can you imagine typing, and getting an answer before hitting "enter" #gamechanger

[@tushybushy]

Do a video on Perplexity...

[@kawaiihikari0]

I recommend co pilot gpt much better at coding than vanilla 4o I’ve found also I have over 18 google accounts not including 3 MS accounts so I never wait for the 4o cooldown period loll 😂

[@ddfyn2393]

ChatGPT is way better

[@AmitKumar__123]

The adaptability of Vematum to market changes is a key strength.

[@White_555pro9]

Vematum's partnerships are a testament to its credibility.

[@B3FX]

Vematum's community-driven approach is what crypto needs.