dms-class-pyai

Py in my AI - an ongoing class/meeting at the Dallas Makerspace.

Setup Python

• install uv
  • https://docs.astral.sh/uv/getting-started/installation/
• $ uv python install

Setup and use a LLM

• install llm Python package as a uv tool
• $ uv tool install llm
• $ llm install llm-gemini
• $ llm keys set gemini
• $ llm -m gemini-2.0-flash 'Tell me about the Dallas Makerspace'

Create an agent

• mkdir and cd into...
• $ uv init
• $ uv add "pydantic-ai[logfire,examples,cli,google]"

Resources:

• https://docs.astral.sh/uv/getting-started/installation/
• https://ai.pydantic.dev/agents/
• https://llm.datasette.io/en/stable/
• https://context7.com
• https://github.com/google-gemini/gemini-cli/blob/main/docs/cli/configuration.md

Extra Resources:

• https://www.tiobe.com/tiobe-index/
• https://automatetheboringstuff.com
• https://treyhunner.com

Tips & Tips

• Sending links to Perplexity to summarize (like youtube videos)

Tools:

• yt-dlp - for downloading YouTube videos.

Ideas:

• Try to use NotebookLM from outside of itself (unofficial API)
• Scrape bot for reddit - AI-related subreddits (to learn more AI stuff)
• Scrape Discord to get a summary of what was talked about

TODO:

1. beginner landing page with instructs to get started (Github Codespaces)
2. initial project

• using pocketflow
• have an agent that scrapes a subreddit (input)
• display summary of data to termianl
• send output as file to Google Drive (output)

3. use Discord -> Dallas Makerspace -> SIG -> "artificial-intelligence"
4. Figure out which PocketFlow tutorials should be done in what order, which ones build upon others, and which can be grouped for a more complex demo.
5. Prompt people to give demos of what they created (like a lightening talk)
6. Ask for ideas/examples of what things or types of things people want to see in the future.