My Complete AI Developer Setup (and How to Copy It)
If you want to build software with AI, your default laptop setup is slowing you down. Most dev-channel content reviews one tool. Almost nobody shows the integrated workflow — Windows → WSL → Warp → AI in one chain. That integration is the whole point.
This is my exact setup, and exactly how to copy it to your own machine today.
1. WSL — a real Linux environment inside Windows
I'm on Windows. But the tools I use — and the servers my code runs on — are Linux. WSL gives me a real Linux environment inside Windows, so my terminal, my file paths, and my build tools match production. No more "works on my machine" surprises.
One command, then restart:
wsl --install
That's it. Reboot, and you've got Ubuntu running inside Windows. If you already have WSL, run wsl --update to get the latest.
Pro tip: you can cd into your Windows files from inside WSL at /mnt/c/Users/..., so you don't have to move anything.
2. Warp — the terminal that treats you like an editor
The default terminal hasn't changed in 30 years. Warp treats the terminal like a modern editor — it blocks every command, autocompletes, and you can click to edit. Once you use it, the normal terminal feels broken.
Setup:
- Download from warp.dev (free).
- In Warp settings, set the default shell to WSL (
\\wsl$\Ubuntuor pick Ubuntu from the profile dropdown). - Type
gitand let Warp suggest the rest — that's the autocomplete demo.
Make the font huge. Phone viewers will thank you, and it matters more than you think.
3. The AI part — Claude + GLM
The setup is just table stakes. The part that actually changes how I work is letting AI write the boring code while I stay in flow.
Pick a small, self-contained task so it's believable in 60 seconds. Example: write a Node script that renames every .jpeg in a folder to .jpg. The prompt:
Write a Node.js script that renames every .jpeg file in the
current folder to .jpg. Use fs.promises. Include error handling.
The flow: paste the prompt → get the code → save it as rename.js → run node rename.js → show the before/after.
Which AI to start with? Claude for reasoning, GLM if you want a free option that still writes solid code. Both work. Don't get paralyzed choosing — pick one and actually use it.
FAQ
Is Warp free?
Yes — download it from warp.dev, then set the default shell to WSL (pick Ubuntu from the profile dropdown) in Warp settings. It treats the terminal like a modern editor: blocks every command, autocompletes, click-to-edit.
How do I install WSL on Windows?
One command: wsl --install, then restart. You get Ubuntu running inside Windows. If you already have WSL, run wsl --update for the latest version.
Can I access my Windows files from inside WSL?
Yes — your Windows drives are mounted at /mnt/c/Users/..., so you can cd into your existing files without moving anything.
Should I start with Claude or GLM?
Claude if you want the stronger reasoning; GLM if you want a free option that still writes solid code. Both work — don't get paralyzed choosing. Pick one and actually use it.
The lead magnet
The exact commands and links — the WSL install, the Warp download, and that rename prompt — go in a free text file. Pin it as the first comment. That's the whole growth play.
What's next
This is episode 1 of a 3-part series:
- ✅ The Ultimate AI Dev Laptop Setup (this one)
- How to master WSL & Warp Terminal for coding
- Using Claude Code with GLM (deep dive)
Subscribe to devkingov for the next two.
Prefer watching?
Every post here is a lesson in a free video course — follow along on YouTube and track your progress on the portal.
Keep reading
The Composable AI Stack — My 2026 Playbook
Cursor for edits, Claude Code as host, Codex for greenfield, Antigravity for free orchestration. The build order, the cost rule, the local escape, and the safety rails — one playbook.
AI Pair Programming for Beginners — How to Review AI-Generated Code
The skill that decides whether AI makes you faster or buries you in bugs: a beginner's guide to reviewing AI-generated code, prompting for reviewable output, and knowing when to take over.
Build a Document Q&A App — Your Second AI App (RAG in Plain JavaScript)
Let users upload documents and ask questions about them. This guide builds retrieval-augmented generation from first principles in plain JavaScript: chunking, embeddings, vector search, and cited answers — no ML background needed.
Want help applying this? Book a 1-on-1 with a consultant.
Find a Consultant