light you can trust

Can you trust what your AI systems do? And can you prove it?

NourFide is a newsletter about keeping autonomous AI under control: the AI agents that now act on their own, learn on their own, and don't wait for permission. Written for the leaders accountable for overseeing them. Light you can trust.

What you get

The pitch most AI advice gives you is "move fast." That's the easy part. The hard part is what happens after you hand an AI agent the keys: when it makes a call you didn't see coming, can you tell why, catch it, and answer for it?

That's what I write about. Every issue takes one real problem in AI agent trust, reliability, governance, or oversight and works it through to something you can use. How to verify what an autonomous system did. How to keep control when you can't supervise every action. No hype. No doom. Just the questions a builder asks before putting a system into production, applied to AI that runs itself.

Who it's for

You're responsible for AI you didn't fully build and can't fully see inside. Maybe an AI agent is making decisions in your product. Maybe it's touching customers, money, or risk, and learning as it goes. You want the upside of agentic AI without handing over control you can't get back.

If that's the seat you're in, this is written for you.

Why listen to me

I'm Thaer. I've spent 25 years building custom software and automation for businesses, which means I was making automated systems trustworthy in production long before anyone called it AI governance. That's the lens I write from: builder and operator, not commentator.

I'm finishing a master's in AI, I've written a book on overseeing autonomous AI, and I hold patent-pending research in AI governance. But the credential that matters here is the one you'll judge for yourself, issue by issue.

Read the full story

Sample the writing

  1. 001 They sound exactly as sure when they're wrong

    Stop asking whether the answer sounds right. Start asking how you'd know if it weren't.

  2. 002 You can't verify everything, so stop pretending you will

    The verification is the work now.

  3. 003 I spent two days proving my own thesis on an AI

    Confident is not correct. Complete-looking is not complete. Make it prove it.

The archive

One issue. One question worth your time. Straight to the point, the way I'd want it if I were the one reading.