How to get promoted

My goal is not to tell you: "You need to do this and that to get promoted". Goal is to actually help you to get promoted.

How? is less important half of the story. I need to tell you Why?.

I saw this again and again: people need why? to actually execute on how.

Telling you why and how is good, but not good enough for me. I want also to enable you to change the way you think about the corporation. So the how becomes easy and natural for you.

Or at least: not painful.

There are exactly 6 emotions: fear, anger, sadness, happiness, disgust, surprise.

What we call “color” is not a property of the world — it’s a concept your brain applies to wavelengths of light. Different cultures draw the boundaries differently. And that changes what people literally see.

People with higher emotional intelligence have better life outcomes — better health, stronger relationships, higher performance at work. This holds across dozens of studies and multiple meta-analyses.

Barrett showed that emotions are concepts. And that concepts influence the way you feel about the world. This is science, not self-help mumbo jumbo.

So what I’m gonna do in this presentation is show you a new concept that will (hopefully) allow you to think about “what I need to get promoted” not as: “this is hopeless busywork” — but rather: “this is how large organisations work by necessity, this work makes sense in context of a large organisation.”

Current process at all cybercorporations:

Have the candidate do N programming tasks, each under an hour. Each task requires solving a moderately complex algorithmic puzzle.

This process is fair and equitable: everybody knows what to expect. Everybody gets the same tasks (modulo organisational chaos and luck). Everybody is graded the same way (modulo inherent biases).

For the sake of argument: experienced SWEs can just take people in, to their teams. If we trust our senior engineers to do the right thing, we end up with a process that will hire people who deserve the job.

Let’s call this process property: justice.

There is a real tension between fairness and justice.

Corporate processes are strongly biased towards fairness, and by necessity: away from justice.

Let that sink in.

(It is possible to overestimate one’s abilities though.)

But mostly — failure is a result of the system being fair and not just.

…it means that it is not enough to do work that is good enough to deserve a good outcome. You need also to ensure that you explicitly match the process criteria.

It is not enough to be a good engineer to get hired at Google.

It is not enough to be a good researcher to thrive in Academia.

I made the specific meanings up (or hyper-localised their meaning). This doesn’t matter — Barrett’s research shows that infants can easily form new mental categories when exemplars of the category are paired with “word-like” sounds.

Large organisations tend to optimise for it over anything else: legibility.

Not only this. I want you also to understand that:

Legibility is not some failure of the system, but rather a pervasive property of all large organisations — if it is a failure, it is a failure of the human condition.

That if you squint a little, you can see sense in all of this.

And oh boy, is it easier to live in a world that makes sense.

1. Legal and compliance.

2. Human cognitive limits.

3. Every large organisation does it — not just corporations.

4. Wittgenstein’s beetle in a box.

If someone sues the company for unfair dismissal, the company needs to show what happened. Not explain. Not argue. Show. Paper trail. Documented decisions. Legible process.

Compliance is the same principle applied preventively. Regulators don’t trust your word. They trust your records. If you can’t show it, it didn’t happen.

These two alone make legibility non-negotiable. But they’re the boring reasons. Let’s look at the interesting ones.

Dunbar’s number: you can maintain meaningful relationships with roughly 150 people. That’s it. Beyond that, your brain can’t track who knows what, who did what, who is good at what.

In a 20-person startup, the CTO knows who the best engineer is. In a 2,000-person company, nobody does. Not because they’re lazy. Because it’s cognitively impossible.

So decisions about people must be based on artifacts — things that can be read, compared, and evaluated by someone who has never met you.

Chang et al. (2024) ran 21 experiments with over 23,000 participants. The setup: participants choose between two options that trade off on two dimensions. One dimension is presented as a number, the other qualitatively. Result: people systematically favour whichever option dominates on the quantified dimension — regardless of which dimension that is.

Two experiments stand out. In Experiment 4 (public works project choice, N = 2,000), making the numbers harder to compare (e.g. 51/68 vs. 23/92 instead of 75/100 vs. 25/100) nearly eliminated the bias. In Experiment 5 (charity donation, nationally representative sample, N = 602), people who felt comfortable with numbers showed stronger bias — but actual maths ability did not predict it at all.

Human cognition is biased towards easily measurable numbers. This is not a property of “stupid” American corporations. It is a part of the human condition.

You cannot grow a large organisation based on people doing the right thing, like, just that. To let an organisation grow you need to codify the rules. This is not a perfect system — show me a perfect system — but it is a system that scales.

Militaries run on legibility. Every order documented. Every decision traceable. Not because generals are bureaucrats — because lives depend on someone being able to reconstruct what happened and why.

Universities: publish or perish. Your research exists only if it’s written down in a legible form. Brilliant thoughts that stay in your head count for nothing.

Governments: censuses, cadastral maps, standardised surnames. James C. Scott wrote an entire book about this — Seeing Like a State. States that can’t read their population can’t govern it.

Any large organisation. All of them. Every single one.

Imagine everyone has a box with something inside. Everyone calls the thing in their box a “beetle.” But nobody can look into anyone else’s box.

Maybe everyone’s beetle is different. Maybe some boxes are empty. It doesn’t matter — the word “beetle” can’t mean the private thing in your box. It can only mean whatever is shared and visible.

☑ I explained to you that you can influence your emotions and how you perceive the world by changing the concepts you use.

☑ I gave you the concept of legibility, which you can apply to your work, to influence your emotions towards “corporate busywork.”

☐ I will now go into explaining why legibility actually can be useful.

Yes, they are artifacts for promo and graduation. That’s the obvious part.

Write design docs!

If you work on something, have the work tracked in the usual tracking system. Have the issue you work on have a description, progress reports, etc.

Have some sensible metrics proving your work brought value.

Make sure your Tech Lead and Manager know what you’re doing.

Yes: this helps with promo. Produces artifacts.

Your local leadership knows what you’re doing. I will not be able to understand what 20 people do without their help.

By documenting your work, you learn to be a better communicator. And guess what — there is actual research proving that being a good communicator improves your life outcomes.

It is faster to learn by reading documentation than by reading code. Especially now, when you can just feed 5 design docs to an AI and ask for the gist and what is most interesting.

Shared understanding leads to better decisions. More standardisation. You know how you get 10 experiment frameworks inside a single service? By not writing a clear design for the first one!

Code explains how. Documents explain why.

Imagine two teams doing the same work on similar projects. One does maybe 10% less output but spends 10% of their time making their work visible. The second goes 100% coding. Which team will be axed when the layoffs come?

If there are no layoffs — which team will get the headcount?

I cannot make the team’s work visible to leadership, if it is not visible to me.

Now you have a new concept of legibility.

I have also shown you how to think about legibility as something positive.

I have also shown you why the bias towards legibility is not some aberration of American corporations, but rather a systematic property of large human groups. Complaining about this fact is as useful as complaining about rain while not putting a rain jacket on.

Please don’t be the drenched guy screaming at the clouds. Clouds don’t really care.

Now you can think: “design docs are useful”, and stop: “design docs are a necessary evil.”

I have shown you how, but the work you need to do yourself. Barrett proved you can!