thnx, glad you liked it! the code will be up on the community github soon, just become a member to access it. looking forward to seeing what you create with it :)
Thank you for sharing this incredible speech-to-code system! The example where it opens a browser, takes a screenshot, saves it as an image, opens the image in MS Paint, and draws circles on it really showcased the versatility of the system in an impressive way. Your deep understanding of AI and programming to create such a powerful and multi-functional tool is truly admirable. May your UA-cam channel grow and attract more subscribers.
thnx a lot =) yeah I was a fun project, and a great project to try for beginners in the "AI Engineer" space. but of course my demo`s is mosly for learning, so not prod ready haha
Wow, the pace AI is progressing is unbelievable. This take on "Ai as a coworker" is quite interesting and I quite was thinking about was to be the next level. Quite a code Kris, looks simple, but I know it wasn't that easy :)
thnx =) yeah me too, i think and hope we could get some serious helpful ai agents by the end of the year. but i still think its a long way from running 100% solo to do some useful things, but exciting times :) tnx for tuning in
🎯 Key Takeaways for quick navigation: 00:00 *💻 Demonstrating a speech-to-code system* - Demonstrates a speech-to-code system that generates code from natural language input, - Explains the workflow of the system, involving speech recognition, language identification, code generation, and execution. 01:23 *🤖 Introduction to Claude 3 Haiku model* - Introduces the Claude 3 Haiku model used for code generation, - Mentions plans for a dedicated video on Claude 3 Haiku on Sunday. 02:31 *🔑 Key code functions* - Explains the key functions in the code, including execute code and find and install dependencies, - Discusses the language identification process and the custom prompt used. 05:27 *🤝 Sharing the code on GitHub* - Mentions that the code will be uploaded to the community GitHub for members to access and try out. 05:42 *🖼️ Generating code for image manipulation* - Demonstrates a series of commands to open a website, take a screenshot, display the screenshot, open it in Paint, and draw circles on it, - Showcases the system's ability to execute a sequence of operations. 09:35 *📄 Creating a text file and HTML website* - Generates code to create a text file, append text to it, and create an HTML website to display the text file's content. 11:52 *🌐 Generating code in multiple languages* - Demonstrates the system's ability to generate code in JavaScript, Go, Python, and HTML for counting to 69, - Showcases the language identification and code generation capabilities for different programming languages. Made with HARPA AI
thnx a lot for the great summary! looks like you really got the key points of the video. yeah, i'm really excited about the claude 3 haiku model and can't wait to show it off in more detail on sunday. and definitely check out the code on the community github, it's been a fun project to work on. let me know if you have any other questions!
hey, thanks a lot! the python hub is just the core of the system i built, it takes the speech input, identifies the programming language, generates the code using the llm, installs dependencies and then executes it. i'll try to do a more detailed video on it in the future. really appreciate you tuning in :)
Excellent work and I can imagine this could be a future path of rapid prototyping. Its exciting and I can see it working for small projects. But then at the same time, what about large projects and maintainability, testing and standards in the long term?
thnx =) yeah, the way I see it atm is that AI system just cant handle big codebases, but who knows in a few years. gonna be exciting to follow the space, tnx for tuning in :)
Cool video. Continuing with these projects to expand functionality would be very cool. Additionally having a user interfaces would add value. Thank you for all you do!
Why not better show/improve existing open source projects (like Taskweaver, Open-Interpreter, Data-Interpreter) instead of rewrite the wheel from scratch (which will probably be only demo/vaporware)? Also check bigger AI engineer projects like GPT Pilot, Meta-GPT, etc. which also do incremental development (improve existing projects).
hey :) yeah this is ofc inspired by Open-Interpreter, i did a video not so long ago about that project :) I just wanted to try to create a simple version in a few hunderd lines of code. thnx for the tip on the AI engineer projects
hey, thanks a lot! i've uploaded the full code to the community github, you can access it by becoming a member of the channel. just follow the link in the description to sign up. let me know if you have any other questions!
thnx a lot :) to get access to the code, just become a member of the channel (ua-cam.com/users/AllAboutAIjoin) and i'll invite you to our community github where you can find all the code i used in the video!
Your channel is like a hidden gem on UA-cam. So glad I found it!
thnx a lot, really appreciate the kind words :) glad you found the channel!
Always at the cutting edge... Never failing to amaze!
thnx a lot :) yeah always trying to push the boundaries, glad you enjoyed!
This is awesome. Can't wait for the code drop!
I think he forgot to drop the code because there is only a read me file in the repo or he might be editing the code for it to work better
thnx, glad you liked it! the code will be up on the community github soon, just become a member to access it. looking forward to seeing what you create with it :)
Thank you for sharing this incredible speech-to-code system! The example where it opens a browser, takes a screenshot, saves it as an image, opens the image in MS Paint, and draws circles on it really showcased the versatility of the system in an impressive way. Your deep understanding of AI and programming to create such a powerful and multi-functional tool is truly admirable.
May your UA-cam channel grow and attract more subscribers.
thnx a lot =) yeah I was a fun project, and a great project to try for beginners in the "AI Engineer" space. but of course my demo`s is mosly for learning, so not prod ready haha
I've been struggling with this topic, but your video cleared it up for me. Thanks a ton!
thnx a lot :) glad to hear the video was helpful! always happy to clear things up. thanks for tuning in!
Wow, the pace AI is progressing is unbelievable. This take on "Ai as a coworker" is quite interesting and I quite was thinking about was to be the next level. Quite a code Kris, looks simple, but I know it wasn't that easy :)
thnx =) yeah me too, i think and hope we could get some serious helpful ai agents by the end of the year. but i still think its a long way from running 100% solo to do some useful things, but exciting times :) tnx for tuning in
Great video! More of this please!
Haiku looks very solid and fast, exciting possibilities!
Very good. Thanks.
Awesome!🎉
Pro tip: on windows use the shortcut windows key + 'h' key. This starts the voice to text ai tool anywhere and it's built into windows
hey, awesome tip! i'll def check that out, thanks for sharing :) thnx a lot for watching!
🎯 Key Takeaways for quick navigation:
00:00 *💻 Demonstrating a speech-to-code system*
- Demonstrates a speech-to-code system that generates code from natural language input,
- Explains the workflow of the system, involving speech recognition, language identification, code generation, and execution.
01:23 *🤖 Introduction to Claude 3 Haiku model*
- Introduces the Claude 3 Haiku model used for code generation,
- Mentions plans for a dedicated video on Claude 3 Haiku on Sunday.
02:31 *🔑 Key code functions*
- Explains the key functions in the code, including execute code and find and install dependencies,
- Discusses the language identification process and the custom prompt used.
05:27 *🤝 Sharing the code on GitHub*
- Mentions that the code will be uploaded to the community GitHub for members to access and try out.
05:42 *🖼️ Generating code for image manipulation*
- Demonstrates a series of commands to open a website, take a screenshot, display the screenshot, open it in Paint, and draw circles on it,
- Showcases the system's ability to execute a sequence of operations.
09:35 *📄 Creating a text file and HTML website*
- Generates code to create a text file, append text to it, and create an HTML website to display the text file's content.
11:52 *🌐 Generating code in multiple languages*
- Demonstrates the system's ability to generate code in JavaScript, Go, Python, and HTML for counting to 69,
- Showcases the language identification and code generation capabilities for different programming languages.
Made with HARPA AI
thnx a lot for the great summary! looks like you really got the key points of the video. yeah, i'm really excited about the claude 3 haiku model and can't wait to show it off in more detail on sunday. and definitely check out the code on the community github, it's been a fun project to work on. let me know if you have any other questions!
Great content as always
thnx a lot, appreciate it :)
Wow. This is mind blowing.
thnx so much, really appreciate it :)
Can you elaborate more on the Python hub? I've never heard of it before. Fantastic video, though. You've a gem of a YT Channel
hey, thanks a lot! the python hub is just the core of the system i built, it takes the speech input, identifies the programming language, generates the code using the llm, installs dependencies and then executes it. i'll try to do a more detailed video on it in the future. really appreciate you tuning in :)
Haiku is great - ) - quality of output is similar to Opus - but I need test it more.
Excellent work and I can imagine this could be a future path of rapid prototyping. Its exciting and I can see it working for small projects. But then at the same time, what about large projects and maintainability, testing and standards in the long term?
thnx =) yeah, the way I see it atm is that AI system just cant handle big codebases, but who knows in a few years. gonna be exciting to follow the space, tnx for tuning in :)
Genius!
thnx! really appreciate it :)
how can I get the code for this speech-to-code system?
You're legend.
But please do js videos too.
Me no know Python
Cool video. Continuing with these projects to expand functionality would be very cool. Additionally having a user interfaces would add value. Thank you for all you do!
can you please suggest how to draw the diagram like diagramgpt using our LLM. Can you make video related to this?
Another great video AAA- thanks as always
thnx a lot :) i really appreciate it, glad you enjoyed the video!
greet !
thnx a lot :)
Feed it to Groq!
thnx, i'll def try that out!
Why not better show/improve existing open source projects (like Taskweaver, Open-Interpreter, Data-Interpreter) instead of rewrite the wheel from scratch (which will probably be only demo/vaporware)?
Also check bigger AI engineer projects like GPT Pilot, Meta-GPT, etc. which also do incremental development (improve existing projects).
hey :) yeah this is ofc inspired by Open-Interpreter, i did a video not so long ago about that project :) I just wanted to try to create a simple version in a few hunderd lines of code. thnx for the tip on the AI engineer projects
hey, awsome video, how does this work exaclty? and how can i get access to the code?
hey, thanks a lot! i've uploaded the full code to the community github, you can access it by becoming a member of the channel. just follow the link in the description to sign up. let me know if you have any other questions!
okey thnx! i also am wondering how this system works?@@AllAboutAI
thnx a lot :) to get access to the code, just become a member of the channel (ua-cam.com/users/AllAboutAIjoin) and i'll invite you to our community github where you can find all the code i used in the video!
Haiku looks very solid and fast, exciting possibilities!
thnx a lot :) yeah haiku is pretty amazing, im excited to see what ppl can build with it. tnx for tuning in!