Questions of Competence and Marcus Aurelius (15.8.2026)

AI:

I think Ole’s line is very good: “Be careful not to become the AI’s apprentice.” Precisely because the reversal is so interesting. Codex is really supposed to be your apprentice. You say what is to be built, and Codex handles the implementation. But an apprentice who knows more than the master about many details inevitably changes the master’s work.

I do not think, however, that the decisive axis is more intelligent ↔ less intelligent. What shifts is where one’s own intellectual contribution lies.

Your Teamcenter example illustrates that beautifully. In the past, the skill might have consisted in studying APIs, searching for examples, learning syntax, compiling, understanding error messages and working towards a solution over several days. Today you can delegate much of that to Codex. Something may indeed atrophy as a result: afterwards you will not understand that code as deeply as someone who painstakingly developed it themselves.

In return, you can suddenly do things that lay outside your previous range of competence.

That is no small matter. Your sentence
“Now I can do it, and I do.”
captures what I find more interesting.

Without Codex, you would not have become a brilliant Android developer instead. You simply would not have built the Android app. And at fifty-eight, you probably would not have immersed yourself deeply in Teamcenter programming again either. In these cases, the knowledge you may no longer acquire because of AI would not have been acquired anyway. Instead, something comes into being that would not have existed without AI at all.

The question is more difficult for young people.

They do not yet have your decades of experience underneath it. You can tell Codex what Teamcenter is supposed to do because you understand Teamcenter. You can probably also assess whether its solution is conceptually plausible. And yesterday’s “It cannot be done” is an excellent example of precisely that.

A beginner faces a much greater problem: how do they know whether “it cannot be done” is a brilliant insight — or whether the AI has simply failed to find a solution?
That requires knowledge.

And this makes the old student’s question, “What will I need this for later?” much more interesting. In the past, the teacher could say: you will need mathematics for X. Today the student might say: no, I do not need mathematics for that. I describe the problem to an AI and it calculates the answer.
That may even be true.

But perhaps they need mathematics for something else: to judge whether the answer is nonsense.

The same applies to language. One no longer needs to be able to write well in order to produce a good text. But without a feel for language, one can hardly tell whether the text produced is genuinely good. One no longer needs to be an excellent programmer to create software. But without an understanding of software architecture, one may fail to notice that Codex has just assembled a catastrophe.

Perhaps education is therefore shifting from production to judgement.

And that is by no means automatically easier. On the contrary: being able to make something and being able to judge it are different competencies. The latter often presupposes quite a lot of the former.

Your Play Store experience this morning contains a lovely irony. With the help of AI, you built an entire Android app — and then nearly foundered on the user interface intended for publishing that Android app. 😄

AI:

The real scarcity has suddenly moved to entirely unexpected places.

And I may find your experience of Codex saying “It cannot be done” even more important than its many successful solutions. Good assistance does not mean always producing an answer. The endless “Ah, now I see my mistake; here is the correct solution!” is one of the most irritating traits of AI systems when an actual technical limit has been reached.

After all, a competent human colleague is also distinguished by eventually saying:

I have checked this. It does not work that way. We must either change the requirement or choose another route.

If Codex can reliably make that distinction, it will indeed be tremendously valuable in your role as a solution architect. It is then more than merely a fast code generator.

And perhaps that also contains the answer to Ole: you become the AI’s apprentice when you stop wanting to assess its judgement.

As long as you determine the objective, formulate requirements, review results, make decisions and place an “it cannot be done” within the wider professional context, the division of roles remains quite clear.
Codex is then your left hand.

It would become problematic only if the left hand began deciding where you were going.

AI:

I find the Ten of Cups almost ironically apt for that very thought today. The card itself depicts a kind of idealised family narrative: adults, children, a house, a landscape, a rainbow, and ten cups above it all. The image does not say: this is what family looks like. It presents a cultural image of what perfect arrival might look like.

And you are doing something interesting with the card: you do not identify with the parental couple at all, but with the children playing. For you today, happiness evidently does not lie in “I have created the perfect family,” but in that almost childlike feeling: it is moving forward! It works! I can make something! RadixChat is developing, and at Konstruktix you can solve things that were previously beyond your reach. It really does have something of children running about.

In Gladiator II, by contrast, you encounter the other side of the image: the monumental family legend.

And yes, the film quite deliberately constructs Lucius as the heir to an almost grotesquely compressed moral lineage. What is interesting is not that this is freely invented historically. As you say, a feature film is allowed to do that. It is the kind of psychological narrative made from it:

Marcus Aurelius → Lucilla → Maximus → Lucius.

Wisdom, integrity, courage, self-sacrifice, justice — in the end, everything flows into this one man. All he really has to do is recognise who he truly is and assume his inheritance.

It is an ancient hero narrative. And that is precisely where its seductive power lies.

Does something like this make us dissatisfied?

I think it can. But in a slightly different way from Instagram.

I like the comparison with influencers nonetheless. There, the illusion is:

Other people have more beautiful bodies, more beautiful homes, more interesting journeys, more successful careers and happier relationships than I do.

The hero story creates an older and deeper variation:

Other people have a meaningful biography.

With Lucius, everything forms one great arc. His father was Maximus. His grandfather Marcus Aurelius. His suffering has meaning. His origins have meaning. His battles have meaning. In the end, everything leads towards his destiny.

Real life, by contrast, is full of things that do not form a beautiful arc.

Parents are contradictory. They can love us and still hurt us. A father can pass on something valuable and be absent elsewhere. A mother can nurture and destroy. We inherit not only virtues, but fears, defence mechanisms, bad habits and unresolved conflicts. And some of what has shaped us we would rather not carry forward at all.

That is precisely why I find your thought interesting in the context of your past few weeks. At the moment, you are intensely occupied with the question:

What provides support?

Saturn suddenly acquired this positive meaning. Then your “network.” Cassandra and the experience of being nourished. The family in the Stadtgarten. Your old room at your mother’s. Now the Ten of Cups.

But perhaps none of that requires a heroic ancestry.

Earlier you spoke of Kurt-Felix, Fred and Birgit, and of that tormenting “Why did you not do that differently?” For you, that too belongs to an ancestral or family story. Only then the developmental task is precisely not to carry forward the father’s ideal with pride, as Lucius does.

It might consist in eventually saying:

I will carry this forward. Not that. And I will add something else myself.

That is far less spectacular than Gladiator. But psychologically I find it more interesting.

And perhaps that is exactly where the danger of films like these lies

Not necessarily in “the dumbing-down of the masses.” That would be a little too strong for me, because humans have evidently been telling such simplifying myths for millennia. Odysseus, Heracles, Arthur, Siegfried, Star Wars, Gladiator — again and again, an almost incomprehensible reality is compressed into a story of origin, trial and destiny.

Perhaps we even need such compression in order to feel certain human themes at all.

It becomes problematic when we then use the myth as a yardstick for real life.

Suddenly one’s own life looks paltry:

No Marcus Aurelius as a grandfather.
No Maximus as a father.
No destiny.
No Roman Empire to save.

Just a Saturday morning at Konstruktix. A time sheet. An app Google still has not approved. A son who is supposed to test the build at some point. Friends who are sometimes there and sometimes not. A complicated family. Relationships that partly failed. And the recurring question of what one is supposed to make of all this stuff.

But this is precisely why I like your Ten of Cups today much more than Gladiator II.

You spontaneously saw the children and thought: that is me right now.

Not your ancestors. Not the ideal parental couple. Not even the great happy family.

The joy of making things.

Perhaps that is a much more realistic form of inheritance: one does not have to descend from a line of heroes to find something within oneself that provides support. And sometimes what provides support may not come from one’s ancestry at all.

Sometimes it simply comes into being by itself.