🪞 My MBA notes turned out to be a mirror

🧭 A year after finishing an Executive MBA, I reread the notes cold, looking for the concepts. I found a thinking pattern of mine instead - and what it quietly hides.

📧 Found an error or have a question? write to me or leave a comment

I finished an Executive MBA in March 2025: part-time, weekend classes, done while working. Twenty-two lessons over a year and a half, then nothing - the notes sat parked in a corner of Notion, untouched. A few weeks ago I finally felt like taking stock: rereading all of them and pulling out a recap of what was worth keeping.

I went in to recover the concepts. I came out with something I wasn’t looking for.

Rereading those notes cold, what jumped out wasn’t the cost of capital or the production thresholds. It was me: the same move, lesson after lesson, the way my head insisted on chewing on every single topic. Stubborn, identical, automatic. But I’ll get there - I want to build up to it the way it crept up on me. First, let me start with an embarrassing one.

🧮 First, the embarrassing part Link to heading

There was an exercise, early on, about fixed and variable costs. Textbook stuff, nothing esoteric. Like a good engineer, I tried to answer before looking at the professor’s solutions. I got several wrong. Not “oh, a slip”, properly wrong, with that slightly awkward feeling of having applied the right logic in the wrong place. I mention it because it’s the honest context all of this should be read in: I didn’t show up there to teach anyone anything. I showed up to get a fixed-cost exercise wrong like everyone else.

Then something happened that stuck with me more than the wrong exercise did.

🏭 The question I kept asking Link to heading

We were in Operations, pure manufacturing module. We were talking about production planning: how you size it, how you “plan” it. And I, coming from software consulting, ask the obvious question: ok, but how does all this apply to my field? To software, to consulting? The answer, in class, was: “we don’t know.”

I don’t say that with any bitterness. If anything, reading it back now, it almost makes me laugh: I’d asked a genuinely curious question and gotten the most disarming “dunno” possible. But the point isn’t the answer. The point is the question. Because I’d already asked that question. I’d asked it about warehousing, I’d asked it about variable costs, I’d asked it about craftsmanship. Except the other times I asked it inside my own head, in the notes, in parentheses. There, for the first time, I’d said it out loud.

And rereading it cold, that question is everywhere.

🔁 The same move, three times Link to heading

There’s the day we talk about industrial scale (production thresholds, plants, the point beyond which it’s cheaper to pay overtime than to hire) and in my notes, in the margin, I wonder whether the same logic applies to sizing an in-house dev team. Build vs buy, how many people, when the margin breaks. No answer in the slides. White space, filled only by my question.

There’s the day about banks, and there the pattern wears a different face but it’s the same reflex: a note of mine, verbatim, that banks feel “treated as privileged” to me, the more damage you do to society if you fail, the more you’re protected. I wasn’t studying the lesson. I was looking at it from the outside, hunting for where the system doesn’t add up to me.

And there’s the day, near the end, when the professor talks about mixed mini-teams: an engineer, a marketing person, a finance person, each acting as a guard-rail for the others. And I, instead of taking notes on team building, write down a product idea: what if it were specialized agents doing that? Each one reads the company documents from its own angle and flags the inconsistencies. A concrete example I’d actually jotted down: you tell clients 90-day payment terms, but in a signed contract one shows up at 120. Maybe it’s intentional, but someone should notice. This was March 2025. I’m not saying I invented anything. I’m saying that faced with an example about team building, my head went there.

Three different lessons, three different concepts, and every time the exact same move:

And in software, how does this play out? Where’s the hole in the system? And what if I put an agent on it?

🌍 You don’t learn it, you translate it Link to heading

This is the pattern. And the interesting thing isn’t that it’s particularly brilliant, it’s that it’s completely involuntary. I didn’t decide it. It’s what happens when you spend your days building software: faced with a warehouse, a WBS, a chart on the cost of capital, your head translates. It can’t not translate. It’s a professional bias, and it shows almost comically precisely because these were private notes, written for me, with no one to impress.

And here’s the thing I actually took home, more than any framework:

🎲 A number that hides a choice Link to heading

With one caveat, though. Because there’s another note, on Net Present Value: that thing where you pick a rate, discount the future cash flows, and get a number that’s supposed to tell you whether an investment is worth it or not. It looks like the objective output of a calculation. But that number rests on two things that aren’t objective at all: the rate you pick and the projections of future flows you feed into it. Change one reasonable assumption (a couple of points on the rate, a slightly more conservative revenue estimate) and the “worth it / not worth it” flips. NPV doesn’t lie: it presents as a result what is largely an input of yours, dressed up as math.

And it holds for my pattern too. “And in software?” is a lens, not the truth. It’s incredibly useful, it lets me see things nobody in that room was seeing, but it stays a lens: it shows me the world already translated into my language, and every translation loses something.

I thought about it again the other day while orchestrating yet another agent. There too you believe you’re in control because you see the logs, the runs, the steps, but it’s the same illusion as NPV: a number that looks like certainty, a pipeline that looks like determinism. In reality I’ve never had total control over anything: not over a junior I delegate a task to, not over a supplier, not over the banks from three paragraphs ago. Only degrees of trust, growing over time. The agent doesn’t change the nature of the thing, it just raises the stakes.

Noticing that wasn’t on the syllabus, and it’s maybe one of the most useful things the MBA left me. Which is also why going off to study something far from your own world genuinely pays off: not for what you expect to learn, but for the things you take home without having gone looking for them.

🤖 The loop closes Link to heading

One last thing, because it closes a loop. This whole recap, rereading twenty-two lessons, condensing them, finding a thread, I did it with AI alongside me, exactly like I described when I went back to writing “with a co-pilot”. With one small comedy inside the comedy: every time a concept got tricky I’d pull out a different assistant, ChatGPT then Gemini then DeepSeek, never the same one twice - a pattern there too. Which is almost too perfect: the pattern is about translating the world into software and agents, and I found it using those very tools to read myself.