How to Build with AI (Even If You Don’t Code)

Jun 13, 2025

Over 1,000 of you subscribed this week, welcome! This is the first email to help you get oriented. I’ll be helping with two clear paths to build with AI: the low-code way and the AI-dev way.

At the end, there’s a quick survey: Should I start with the Low-Code series or the AI Dev series? Your vote helps shape what comes next.

🧩 2 Paths: Low-Code Creators vs. AI Developers

AI is no longer just for machine learning engineers. There are now two clear entry points:

1. Low-Code Builders: Build Without Writing Code

If you’re not a developer, or just starting, this path is for you. You don’t need to train models or write full apps. You can use AI like a power tool.

Here’s the kind of stuff you’ll learn

JSON basics
What APIs are and how to call them
How to write clear AI prompts

Some AI stacks you might build with

n8n or make + OpenAI
Google Sheets + Claude
Prompt Engineering
+ low-code tools

Designing the idea will probably take more time than you expect.
I’ll help you with that in the next few emails.

2. AI Devs: Plug AI into Your Existing Stack

If you're a full-stack developer, you don’t need to become an ML researcher. You already have what you need to build with AI.

You won’t train models or fine-tune them.
But as a full-stack dev, your job is to structure your app, choose the right model, evaluate outputs, deploy the solution and understand LLM cost basics.

If you’re building solo or launching your own product, these five things matter even more.
You need to know the full software dev cycle, adapted for AI.
And trust me, it’s not as complicated as it sounds.

What you’ll start learning

AI app/solution architecture
OpenAI, Claude, or Mistral APIs
Prompt design and chaining
Vector DBs like Pinecone or Supabase
How to deploy your app
Basic LLM cost estimation

Some of the AI stacks you could use:

LLM APIs: OpenAI, Claude, or Mistral
System Prompts ( ReAct framework for reasoning and action)
LangChain, CrewAI, Python, …
Vector DB: Supabase, Pinecone, or Weaviate
RAG Infrastructure: Chunking, embedding, retrieval logic
Deployment: Serverless (Vercel/Fly.io) or container-based

⚠️ Not for ML Researchers

This newsletter won’t go deep into model training or fine-tuning. If you're here to build and ship with existing models, you're in the right place.

If you're here to train models, this isn’t your newsletter.

Where Are You Right Now?

There are two paths: Low-Code Builder or AI Developer
I’ll be writing both, but I need your help to choose where to start
Just scroll down and answer the quick poll
Your vote decides what comes next.

No matter which path you’re on right now, start getting familiar with the work of those already building AI today—teams like Google, Anthropic, OpenAI, and more.

Here’s your first task: check out Google’s Prompting Guide 101. It covers many use cases, but you don’t need to go through every exercise, just get familiar with a few key examples.
You’ll find the link on my GitHub