On 25 June 2018 I built a slide for a training session in Tokyo. It carries a title and nothing else. Two words: Skill Spectrum. Seven years and two hundred and sixteen revisions later, I opened that file again while writing this, and the slide still says exactly the same nothing.
Everything on it lived in my mouth. I stood in front of engineers from a global clothing brand and talked over an empty rectangle for ten minutes, because the model was so obvious to me that writing it down felt unnecessary. I have taught that idea for years. I have never once written it down properly. That is the honest reason this essay exists.
- Three old words, Apprentice, Journeyman, Master, that almost everyone reads as a ranking and almost nobody uses correctly.
- The tier you should worry about is not the bottom one, and the reason is uncomfortable.
- What changes when you stop asking where you are, and start asking where you are in each thing.
Let me be precise about ownership, because it matters. Apprentice, Journeyman and Master are not mine. They come from the guild tradition, and the reference sitting in my own deck is The Pragmatic Programmer, whose original subtitle was From Journeyman to Master. Anyone can borrow those three words, and most people do.
What I did with them is the part I will defend. Almost everyone treats the three as a ladder: you were an apprentice, you became a journeyman, one day you will be a master, and the goal is to climb. Read that way, the model is a ranking system, which makes it useless. It tells you one thing about a person, and the one thing it tells you is usually wrong.
I read them as orientations instead. Not where you have arrived, but how you are currently standing toward a thing.
You have been wrong about this enough times to know the shape of your own blind spots. You can explain it to a beginner without lying, and to an expert without bluffing.
Do I know what this costs?
You are genuinely competent. You ship. The work is good. And the questions you stopped asking six months ago are the ones you most need.
What have I stopped asking?
You know you do not know. Everything is a question, nothing is automatic, and you are paying full attention. This is the most alive place to stand.
Who can show me?
Written as orientations, they stop being a hierarchy and start being a diagnosis. And the diagnosis has a sharp edge.
The Journeyman Is the Risky Tier
The instinct is to worry about apprentices. They know least, so surely they are the risk. In practice they are the safest people in the room, because an apprentice asks. They interrupt. They say the thing everyone else is too senior to admit they did not follow. Nothing rots quietly around a person who is still asking.
The journeyman is different, and the difference is what makes it dangerous. A journeyman is competent enough that the work looks fine and nobody checks. They have crossed the line where things mostly succeed, which is exactly the line where the questions stop. The same slide deck from 2018 carries the Dunning-Kruger effect two slides earlier, and this is why: the trouble is not ignorance, it is illusory competence, the state where you have enough skill to produce a result and not enough to evaluate it.
The dangerous stage is not the one where you know least. It is the one where you stopped asking.
That is also why the model must not be a ladder. A ladder tells the journeyman they are doing well and should keep going. A diagnosis tells them to go and find the thing they have quietly stopped examining.
You Are Not One Level
Here is the thing I should have written on that slide in 2018. You do not have a tier. You have a spectrum. On any given day you are master in a few things, journeyman in many, and apprentice in others, and all three are true at once.
I keep my own spectrum published, and it is deliberately unflattering. Master in Java, cloud and payments, because two decades bought that. Journeyman in Kubernetes and observability, where I am competent and still practising. Apprentice in agentic AI and everything around it, where I am openly a beginner in public at forty-plus with twenty years behind me. That last row is not an embarrassment to be minimised. It is the row where I am most alive.
Two failure modes come from reading yourself as a single level. The first is the CV read, where you claim the highest tier you have ever touched and quietly stop learning underneath it. The second is worse and more common among good engineers: you take the lowest tier as a verdict on your worth. Neither is honest. An apprentice row is not a deficiency report, it is a map of where your attention should go next, and a person with no apprentice row at all has simply stopped moving.
Run it without self-flagellation. The point of naming a gap is to aim at it, not to carry it around.
A Team Has a Spectrum Too
The same instrument works on a group, and it is the version I have used most as a lead. I have written before about a team's capability as a distributed range rather than a fixed set of skills, where the goal is a spectrum with no critical gaps and a floor that keeps rising. That is the foundation. What I want to add here is what the spectrum shows you that a skills matrix does not.
A skills matrix asks who can do what. A team spectrum asks something better: where is each capability standing, and who is standing there alone? Three patterns fall out of it immediately, and I have watched every one of them cause damage.
The single master is the obvious one. One person is master, nobody else is even journeyman, and that capability is now a person rather than a team property. Everyone knows this is a risk and almost nobody schedules the fix, because the fix is slow and the risk is invisible until that person is on leave.
The all-journeyman band is the one nobody sees. A whole team competent in the same thing at the same middling depth, all of them past the point of asking, none of them deep enough to catch the flaw. It looks like a strong team. It reviews its own work and approves it. This is the team-scale version of illusory competence, and it is the reason a room full of experienced people can still walk into a wall together.
And the missing apprentice row. If nobody on your team is openly a beginner at anything, your team is not learning, it is maintaining. That is fine for a quarter. Over two years it is how a group becomes expert at a world that no longer exists.
Map it once, honestly, with the team in the room and their own names against the tiers. It takes an hour and it is uncomfortable in a useful way, because it converts a vague feeling that we should upskill into a specific and arguable picture of where the depth actually sits.
One Question, Asked Often
On the slide after the empty one, I asked the room a question I still ask myself: when did you last learn something new? People hear it as motivational, the kind of line that belongs on an office wall. It is not. It is diagnostic.
If the answer arrives instantly and specifically, your apprentice row is populated and you are still moving. If you have to reach for it, or if the honest answer is a while ago, that is not a character failing. It is a reading, and it tells you the spectrum has quietly narrowed while you were busy being competent. The value of the question is entirely in how quickly you can answer it.
This is why the assessment is not an annual exercise. Annually is how you produce a document. Regularly is how you produce a habit, and the habit is the point, because a spectrum is not a snapshot. It moves, and it only moves in the direction you keep looking.