Simple does not mean easy is the most important lesson I learned, or recalled, in recent days. Not that it’s a new concept (what is, really?), but I am glad it was fiercely brought to my attention by renowned author Aagaard. And while his point is well taken — chess moves are simple but by no means easy to play1 — I have asked myself in what other contexts there exists a similar pattern.

I believe that, in engineering, there wasn’t a similar pattern until very recently. The illusion of ease in chess comes from the presence of computer programs that are very, very strong, much better than any human player in fact. And because in most positions there are approximately only five sensible moves, a skilled-enough player will spot all of them and then, by looking at one such computer program, she will see what the best move is, recognize that it was among her candidate moves, and think ‘Of course, that’s easy’. Alas, if that was a real game of hers, she would probably struggle to make a choice among those five candidates and, in four cases out of five, pick a sub-optimal move.

Thus we understand the elements that create the illusion: a rather limited pool of sensible options; a skilled-enough, but not outstanding, human player; the quick and effortless feedback provided by the machine.

Now, back to engineering: do those elements ring a bell? They probably wouldn’t, until a year, or so, ago. But what about today? The chatbots, so-called AI tools, play the role of the chess program rather well. The engineer, if she’s skilled-enough, which these days means basically everyone with half a brain, can read what the chatbot says in response to some prompt and think ‘Ah, of course that’s the way to do it.'

There’s one major difference, though. In engineering, and especially so in enterprise engineering, most things are really easy. It’s not just an illusion. I’ve written at length about this, so here’s the summary of it: enterprises almost never invent, they only copy each other, and every new thing is not really new, but just a tiny variation of something that’s already out there, made to please some mid-level manager who couldn’t even make tomato sauce with his own hands, let alone to build software, nor hardware, systems. The direct consequence is that things are simple and easy. And the practical aspect is that the skilled-enough engineer is mostly expecting, from the chatbot, a confirmation of her own thoughts that she knows from experience to be correct, not the discovery of something new and exciting.

To be sure, those bots type much faster than me. Nearly at infinite speed, in fact. That alone is a great reason to utilize them, like I do.

Is this as bad as it sounds? No, I think it isn’t. The reason, in one short sentence, is that it’s just work and who cares. If enterprises want brainless, repetitive work done even by the brightest minds, then engineers should be bright enough to let them have it. We cannot expect to find intellectual pleasure in work, except in a minority of fortunate instances. Do other things. Run, play chess, typeset a book, or anything that tickles your mind.

But if one really wanted to keep an active mind, while still leveraging the tools, would that be possible? I think it would, with a lot of dedicated practice. One would need to avoid the illusion that comes from having the answers to all questions effortlessly. There are benefits in making an active, conscious effort in engineering design, even when the devised solution is not optimal and must be revisited. Just like there are benefits in making an effort at the boundary of our capabilities in sport, music, chess, etc. I don’t know yet what shape, or form, this type of practice should have for engineers. I suppose it must consist of poking the bot and forcing it to give hints without giving away the solution. I haven’t developed anything substantial, yet.


I did, however, develop something along the same line of thoughts and in the context of a chess game. Back to my first point, the key is to avoid the mindless activity that comes from looking at what moves and variations a chess engine recommends when a player analyzes a position. Instead of that, the player should want to actively engage his brain (or, at least, the part of his brain that thinks about chess), and ask questions about the position, calculate variations, and make conclusions, and then check the soundness of his thoughts by using the engine as an oracle.

I gave it the simplest form I could think of: a CLI command that takes in a game annotated by the player and checks the annotations with an engine. The output looks like this:

Meta-analysis: White Player - Black Player
Engine: Stockfish 19 · depth 35 · hash 2048 MB · 2 threads
Scope: all 
──────────────────────────────────────────────

Main line
  1W (Nc2)    Evaluation swing not reflected in the annotation.

Variation 1W.1 (1. Nb3)
  4B (d5)     Evaluation slightly off.

Variation 1B.1 (1... a5)
  5W (Qb3)    Incorrect evaluation: check which side is better.

Thus the player, if he really wants to work and improve his chess skills, looks at the output and understands that he missed something at move 1W (first White’s move), because there was a swing in the engine’s evaluation. Crucially, the output doesn’t say anything else. It’s up to the player to put in the work and figure things out, and not fall in the laziness trap. I am still testing this feature out, and so far I am quite pleased with it. It goes nicely with my goals of enjoying the process before the results, and to “simply” become better at the game.


I got my first two copies of Chalkdust. I commend all of the authors for the efforts, and for making a very enjoyable read. I think I will get more copies in the future.

It brought back to my mind memories of the personas who populate the academic world; I was, after all, in there for well over three years. What really jumped to my eyes, while reading and enjoying the magazine, was the unmistakable, immediately recognizable nerdy humor, (more often than not) totally unnecessary and yet always present, in speech as well as in writing. I think the magazine could do just as well without.


I read The Elements of Typographic style, and the first thing I wanted to do after reading the last page was to go back to the first. It’s one of those books that I know, even while I am reading it, I will need to read and re-read again and again.

The last note I took from it was “writing is an instrument of freedom, not an instrument of power”, written in the context of giving space to the readers, room to breathe for their minds. Perhaps I am just looking for confirmation of things I believe and have said in the past; perhaps I can’t help it. But if you are reading this, pause for a minute and try to think how many times in just the past few days you’ve noticed people talking (or writing) way, way too much in order to express a simple concept, just with the hope to steer your own thoughts and interpretation towards their own. Managers at work do it all the time, don’t they? Talking and talking and talking, essentially to say a very short idea and then adding in their own, unnecessary, interpretation. I don’t blame them, because talking is basically their job, but it is annoying. What about books, though? In how many have you spotted the same sin?


I got on my desk a copy of Grid systems in graphic design, which I thought would be the logical continuation after Bringhurst’s book. This one is surprisingly scarce on words, it mostly consists of drawings. All considered, it’s a promising start.


  1. These days, every time I type an em dash, or even an en dash, I stop for a second and think: Will anyone believe this was written by a chatbot? It’s just one second though, because, frankly, I don’t care. ↩︎