Autonomous AI Agent in Python from Scratch (ChatGPT Function Calling)

preview_player
Показать описание
In today's video I create a simple AI Agent that can write to files (and therefore create multi-file code projects) in Python using the new ChatGPT Function Calling API

00:00 Intro
00:43 Start coding
01:17 Setting OpenAI API-key
02:25 Sending a message to ChatGPT
03:43 Getting response from ChatGPT
04:11 Defining messages
04:39 Defining functions
06:41 First test
07:24 Parsing function call arguments
12:18 Second test
13:18 How function calling works
14:12 Implementing chat
20:53 Testing chat
22:47 Reversing the example
25:05 Debugging
25:47 Testing "tell 3 jokes"
27:16 Letting it write to files
30:38 Adding a system message
32:02 Calling sequential functions
34:14 Creating a web app with it
36:29 Outro
Рекомендации по теме
Комментарии
Автор

Man Your content is legit. Keep it up, great things await you.

housamkak
Автор

Clear, simple, straight forward. GREAT video: Thanks!

jaoltr
Автор

Wonderful explanation 👌 😃 .... Just a suggestion, Next step could be to read existing code files so that we can ask questions based on that existing codes.

shailesh.kharche
Автор

Very impressive contribution 😈😎💥💪 This opens many doors. Thank you.

gregorymagnusson
Автор

Thank for accepting my request, i really appreciable. I have requested many YouTuber but your are only one helped me. I want see you 1 M subscribe in near future. Have great day . Thank you very much. We want see many videos like this.

oxytic
Автор

Can you make it take pdf files, change them to text, and extract data from them?

housamkak
Автор

Wondering if you can create a video on how to make an autonomous agent in php similar to langchain ? Thanks

BuyCarsTVPakistan
Автор

I love this video thanks a lot indeed!!!!

sealandland
Автор

If you get the time, can you work on a dashboard that allows the user to select from multiple LLMs to participate in a group style project with the user as a project manager that helps to guide the project. Essentially the same thing as your gpt autopilot but with the ability for the LLMS to communicate with each other to achieve a maximal result. The dashboard should include a space for the user to be able to upload his/her own files or information for the AI group to be able to view as a way to pre load memories for them to work from.

aidrivendesigners
join shbcf.ru