100 Days of Gen AI: Day 3

It’s been a while since my last entry. Life got busy and I didn’t make enough time for learning, but now I’m back!

Today I went through the OpenAI quickstart guide, which was quite easy – I was up and running in 5 minutes. I should have started out here all along!

I got my first equivalent of a hello world app running in python. I also read through the OpenAI prompt engineering guide which was a helpful way to start thinking of different strategies for implementing some of the AI projects I want to build.

Some ideas for future projects that came to mind during this process are:

• A gift recommendation engine for GoWishbone
• Take user feedback and a product roadmap as inputs, and explain to the user how their feedback is being addressed (or not addressed) by the product roadmap
• Have a conversation with Ben Franklin as if you two were having a beer in a colonial-era pub
• A code assistant that writes unit tests, and a separate application that runs these tests in a sandboxed environment, sending the feedback back to the model, allowing the model to iterate and fix issues
• Observability SRE monitoring: given past monitor inputs, investigations and root cause outputs, suggest issues to consider based on current monitor inputs
• Content moderation tool
• Code review bot
• Train your own AI persona, automate replies to messages in your own voice
• Identify logical fallacies in text/speech, compare different political figures by types of fallacies and rates of committing them
• Fact checking assistant

The resources I want to read next are:

• Prompt examples
• Function calling
• OpenAI cookbooks

I also want to try playing around with these supporting tools:

• Chat playground
• Evals
• Prompting libraries & tools
• Explore other endpoints