Haha this is really cool, you should do a whole series on different projects. Some beginners might want you to break down the video more but I understood as I’ve been doing this interchangeably with gpt and Claude. I feel like Claude 3.5 goes crazy on frontend design compared to other llms. A big tip for future designs ask Claude to use the shadcn library and you can achieve some really modern frontend designs
Great videos! Do you think including an image makes any difference? I tried asking Claude to solve a Sudoku game by taking an image of it. It didn't manage to read the numbers properly, so I had to manually corrected it. But then I asked it to tell me the next place to go and why, and it just couldn'tt give me the right answer. I'm trying to do it all on my mobile phone as a single page react application that works on a web browser. I use the project feature. I dont think it's good enough yet or i haven't worked out how to correctly prompt it. Copying the results into VSCode on a pc may be the only way it's useful. I'm a software developer by trade with an interest in AI just to see if the traditional way i work can be made quicker and easier by AI tools.
Is t there a way to do this with local LLM's and online webpages? I know the short answer is API's but I have no idea hot to achieve using the models on my server to send responses to my online service.
i'm still battling with claude when i do a small project like this, the more i want to modify it and add simple features or css layouts, thats when it starts breaking down. and forgetting things, but I guess for a first run simple MVP then it can create something
@@AllAboutAI yes, i use project, and i always have to give them by attaching the latest code, they say modify this code, and give me back a complete rewrite, otherwise they want to give me snipplets of code, i have to figure out what to replace, and that isn't fun each iteration
Haha this is really cool, you should do a whole series on different projects. Some beginners might want you to break down the video more but I understood as I’ve been doing this interchangeably with gpt and Claude. I feel like Claude 3.5 goes crazy on frontend design compared to other llms. A big tip for future designs ask Claude to use the shadcn library and you can achieve some really modern frontend designs
thanks a lot man :) good ideas. will do for sure if ppl like this one
That is a great idea, I understand the tools but without the knowledge to put everything together in only using 10% of the potential.
@@AllAboutAI Exlcillent job but how you do it that fast !
Very cool. I'd love to see how well this would have worked using Aider with Claude 3 sonnet.
Great idea!! tnx:)
Awesome video, getting lots of inspiration watching this. One question have you used cursor or Aider?
Noting that you did this with Claude… how much did it cost to create this app…. Including debugging and what not.
Thank you for your video
I used the $20 pro version of Claude. No not much more than that:) you run out of request tho, so you need a few days
Great videos!
Do you think including an image makes any difference? I tried asking Claude to solve a Sudoku game by taking an image of it. It didn't manage to read the numbers properly, so I had to manually corrected it. But then I asked it to tell me the next place to go and why, and it just couldn'tt give me the right answer.
I'm trying to do it all on my mobile phone as a single page react application that works on a web browser. I use the project feature. I dont think it's good enough yet or i haven't worked out how to correctly prompt it. Copying the results into VSCode on a pc may be the only way it's useful.
I'm a software developer by trade with an interest in AI just to see if the traditional way i work can be made quicker and easier by AI tools.
Excellent, but the prompt does not match the picture.
Yeah I could do more work on the prompting for sure haha
Is t there a way to do this with local LLM's and online webpages? I know the short answer is API's but I have no idea hot to achieve using the models on my server to send responses to my online service.
Very very cool. you glossed over a bunch of stuff (i get it) that i think many of us would love to see, but great video
i create a full stack e-commerce website for tech products, categories, everything
ai become super powerful i can do anything i want literally 😂
👏😅
Which one did you use?
Did you reupload a file fin the project rom the app everytime Claude changed the code for that specific file?
Cool stuff, but please make another video of costing of api and all that things. Also please share how much it is able to earn in few weeks or months
Yeah thats a good idea! Tnx John:)
did you switch to gpt 4o everytime you were out of claude prompts?
I can't think of any other model he would switch to that is as good and free.
Very cool!
thnx Duane:)
Try Tailwing CSS when you want a better UI
awesome saas project
This is from your website "Failed to generate meme. Please try again later." Seems like you need to go back to the drawing board.
Wen Bounty Hunt Prize?
Tomorrw!
Very cool
thnx a lot :) its was very enjoyable to do
yea
take a screenshot of this YT video, add a meme:
Not too funny = it suck
Please don’t make clickbait thumbnails 👎
there are 3 different thumbnails, one is more clickbaity i guess
i'm still battling with claude when i do a small project like this, the more i want to modify it and add simple features or css layouts, thats when it starts breaking down. and forgetting things, but I guess for a first run simple MVP then it can create something
do you use the project feature? Ill try to always keep an up to date context and refer to the files I want to change
@@AllAboutAI yes, i use project, and i always have to give them by attaching the latest code, they say modify this code, and give me back a complete rewrite, otherwise they want to give me snipplets of code, i have to figure out what to replace, and that isn't fun each iteration