Skip to content

xaelophone/how-i-code

Repository files navigation

Learning to Speak Code

A practical guide for non-technical people who want to ship code with AI.

Who This Is For

  • Product Managers with strong technical literacy and can articulate technical systems in natural language
  • Sales professionals who want to build the feature requests they get from customers
  • Marketers who want to build the demo & landing page for the product they're promoting

What's Inside

The core 4-step workflow: Plan → Issues → Implement → Review. Start here.

My go-to technologies for shipping apps: Next.js, React, Supabase, Tailwind, TypeScript, Vercel.

A log of how my coding practices are evolving over time.

Frameworks

Deep dives into the mental models that make AI coding work:

Framework What You'll Learn
Comprehensible Code Why you must understand every line the AI writes
Context Engineering Managing the AI's memory for better output
Atomic Features Ship small, ship often, stay reversible
Emergent Abstractions Why you shouldn't over-architect early
Agent Phalanx Orchestrating multiple AI models (advanced)

Cheatsheets

Quick reference guides for daily work:

Cheatsheet What It Covers
Prompting Patterns How to "talk" to AI effectively
Claude Code Tips Tool-specific tips and workflows
Git Survival Guide The "social protocols" of version control

Examples

Sample templates you can copy and adapt:

Real-world notes on AI models for coding—updated as new models drop:

Model Best For
Claude Opus 4.5 Everything—my daily driver
GPT 5.X Deep context gathering, strict code review
Gemini 3.0 Front-end design, asset generation

The Core Insight

"The hardest part wasn't learning to review AI-written code... despite what everyone told me. The hill was learning how to work with other engineers on the same repo, where we all rely on the engine running smoothly."

Learning to code with AI is surprisingly tractable. The "hard" part is mastering the ancillary practices that developers take for granted: git, version control, branch management, PR etiquette, and CI/CD pipelines. These are social protocols encoded in technical systems.

This repo is designed to help you learn both.


How to Use This Repo

If you're brand new: Start with QUICKSTART.md, then read Comprehensible Code.

If you're already coding with AI: Skim the frameworks for mental models you might be missing, then keep the cheatsheets open while you work.

If you want to go deep: Read everything in order. The frameworks build on each other.


About

These frameworks have emerged from my experience evolving from product manager, to startup founder, to building as a "collapsed" talent–owning product, design, code, and adoption as a single contributor.

Your mileage may vary. But the frameworks are battle-tested.


Go Deeper

This repo is free. If you want more hands-on help:

Work with me — Workshops, 1:1 coaching, and enterprise training.

Or just say hi


Some links in this repo are referral links.

About

A practical guide for non-technical people who want to ship code with AI

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •