I write long-form, illustrated essays about what is actually happening inside AI: how coding agents work, how multi-agent systems are built, how AI evaluation breaks down, and what happens when models start looking conscious but aren't.
I only publish what I wish already existed.
The posts are deliberately slow. I use metaphors which are the load-bearing structures that let you carry a complicated technical idea around in your head after you finish reading.
I am Tanzim Saqib, an software engineer by education and trade, now in this strange new era. I focus on making AI actually useful. I have been building with these systems long enough to be skeptical of the hype and curious enough to keep digging anyway. The posts here come out of that mix: I read the papers, watch the talks, ship the code, and then write the thing I wish someone had written for me.
Before all this, I was a CTO/VP in the old software world, DevRel head at Microsoft, built and led large teams.
First essays are dropping soon.
