Follow along with the project by going here: https://www.parsity.io/5-day-ai-advisory-board

In this episode, I walk you through a side project that, if you actually build it, will put you ahead of roughly 90–95% of software developers right now.

Most developers are using AI tools. Very few are actually building with them.

So instead of talking about it, we build something real.

The project is an AI advisory board — a system where you can ask questions like:

• What would Theo Browne say about Node?

• What would ThePrimeagen say about Vim?

• What would Alex Hormozi say about sales?

• What would I say about testing?

And instead of generic AI responses, you get answers grounded in real transcripts and real opinions.

We go step-by-step through:

• Calling an LLM programmatically (Gemini API)

• Setting up a backend route to handle requests

• Adding system prompts and guardrails

• Building a simple knowledge base

• Pulling in real YouTube transcripts

• Injecting data into the model at the right time (naive RAG)

This is the kind of project that forces you out of “AI user” mode and into actually building AI-powered products.

Resources mentioned 👇

• parsity.io/ai-with-rag
• The Router Pattern – A Smarter Way

I mentioned that I got a lot more requests for interviews. Here's the proof: Da Proof

3 Ways We Work with Software Developers (and noobs) 👇

• AI Dev Program – parsity.io/ai-dev
• Apply for our full stack software engineering program
• Learn JavaScript in 30 days... the hard way

I created a free project for developers who want to learn AI the hard way: https://parsity.io/ai-with-rag

I don't envy new coders. AI is being jammed down your throat and at the same time, you're being told you must understand the code it generates.

It's a confusing time.

Using AI too early comes with a cost. More importantly, we'll look at some proven ways to make sure you are learning deeply and can work efficiently with your little code bot.

You can read the study from Anthropic here: https://www.anthropic.com/research/AI-assistance-coding-skills