NAEOS

We're teaching AI to write code. But we're not teaching it how we engineer software.

Over the past year, AI coding assistants have become remarkably good at generating code.

Yet I keep running into the same problem.

Every new project starts with rebuilding the engineering context from scratch.

Not because the AI isn't capable.

Because it doesn't know:

• Why the architecture is designed this way.

• Which coding standards are mandatory.

• What security policies must always be enforced.

• How services depend on each other.

• Which design decisions have already been made.

• What trade-offs the team has accepted.

We try to solve this with longer prompts.

Or by pasting documentation into every conversation.

Or by hoping the AI remembers enough context.

None of those approaches feel sustainable.

The more I think about it, the more I believe the real challenge isn't code generation anymore.

It's engineering knowledge.

What if architecture, documentation, policies, workflows, and engineering decisions became structured assets that every AI agent could understand?

That's the question that led me to start building NAEOS, an open AI Engineering Operating System.

The goal isn't to replace AI coding assistants.

It's to give them a shared engineering foundation so they can produce more consistent, production-ready software.

I'm still in the early stages, and I don't know if this is the right direction.

That's why I'm building in public.

I'd genuinely love to hear from other founders and engineers:

What's the biggest engineering challenge you've encountered when working with AI coding tools?