Over the past few months, AI has become a huge part of how I build software.
Claude Code, ChatGPT, Cursor—they're all incredibly capable.
But after the excitement wore off, I noticed something frustrating.
I wasn't building products.
I was managing AI.
Every session looked the same.
Write a prompt.
Wait.
Review.
Fix.
Prompt again.
Leave for lunch.
Come back.
Start over.
The AI wasn't the engineer.
I was the project manager.
And that felt backwards.
The Problem
Today's AI coding tools are amazing at writing code.
But they don't really own a project.
They don't remember momentum the way a teammate would.
They don't naturally continue working while you're away.
If you're not actively supervising, progress usually stops.
For small scripts, that's fine.
For real software, it becomes exhausting.
What If AI Worked More Like an Engineering Team?
That question led me to start building Forge.
Instead of another chat window, I wanted something that felt more like a persistent software team.
The idea is simple:
Connect a GitHub repository.
Give Forge a goal.
Let it keep moving the project forward.
Come back to see what changed.
Not because I want AI to replace developers.
Because I want developers spending less time managing tools and more time making decisions.
Where Forge Is Today
Forge is still in development.
Right now I'm focused on building the core experience before adding anything flashy.
My goal isn't to make another AI chatbot.
It's to build something that helps software actually move forward.
Looking for Early Builders
If this sounds interesting, I'd love to hear from other developers and founders.
What frustrates you most about current AI coding workflows?
I've also opened an early-access waitlist while Forge is under development:
👉 https://forge-ai-waitlist.vercel.app/
I'd genuinely appreciate feedback from people who build software every day.
Top comments (0)