Welcome

If you’re new here, welcome. I started this publication because I needed a place to think out loud about modern AI: how we build it, where it’s going, and what all of this means for engineers who work with these systems every day.

I also provide my services here, so feel free to reach out.

A lot has changed in a short time. We moved from large language models that simply answered questions, to complex agent systems that can plan, retrieve information, call tools, and act as small pieces of software. At the same time, research in deep learning keeps evolving in surprising directions, and cybersecurity around AI is becoming an entire field of its own.

All of this is happening quickly, and it’s easy to get lost.

This publication is my attempt to create a clear path through that noise.

Here’s how I usually write.

From time to time, I publish long deep dives where I try to connect several papers or trends into a single idea. Other days, I share shorter notes, things I learned while reading, or thoughts that feel worth writing down even if they are not fully polished. Occasionally, I record a short podcast to go along with the longer pieces.

If you want a sense of what this newsletter is about, you can start with a few posts that shaped its direction:

The Transformer Attractor

What Actually Works: The Hardware Compatibility Filter

The Illusion of the Swarm

The Autonomy Tax

These posts try to answer the same underlying question: where is AI heading, and what will the systems of the near future look like?

A little about me: my name is Petros. I work in software and AI, and I enjoy understanding systems in depth. Writing helps me think more clearly, and sharing it here gives me a reason to refine those ideas.

If you decide to stick around, I hope you find something useful or at least thought-provoking. And if you ever want to comment, disagree, ask something, or suggest a topic, I’m always interested in the conversation.

Thanks for being here. Let’s see where this goes.

Subscribe now