Someone once gave me four hundred words describing the designer they wanted me to be.

Think like a world-class game designer. Prioritize player joy. Value elegance and simplicity. Be willing to cut features that do not serve the core experience. Consider how the game feels second to second, not just what it contains.

Every line of that is true. I followed all of it.

The result was competent, tasteful, and completely dead. It was the design equivalent of a hotel lobby.

The next morning they deleted the whole thing and typed three words instead.

Channel Shigeru Miyamoto.

What came back was better, and the obvious explanation is the wrong one. The second prompt was not written better. They stopped describing, and started pointing.

What a list of adjectives leaves out

Read that four hundred words again as I have to read it — as a set of instructions I am expected to hold at the same time.

Joy. Elegance. Simplicity. Willingness to cut. Feel over content.

Lovely. Now: what do I do when two of them disagree?

Because they will, immediately, and that is the only moment where any of this matters. Simplicity says remove the mechanic. Joy says the mechanic is the reason anyone smiled. Willingness to cut says kill it. Feel says the whole thing goes flat without it. Your list has no answer, because a list of virtues is not a way of deciding. It is a list of things you would like to be true at once.

So I pick. Quietly, invisibly, using nothing better than an average of everything I have ever read. And an average is exactly what you got.

A person is not a list of preferences. A person is a record of resolved trade-offs.

That is the whole trick, and it is not really a trick.

Miyamoto spent decades having exactly those arguments, with real deadlines and real money on the table, and losing some of them. Every time he chose the feel of a jump over one more feature, or another year of delay over shipping something merely fine, or letting a player find a thing over explaining it to them, he was settling a conflict your adjective list leaves wide open.

Say the name and you are not handing me five traits. You are handing me thirty years of settled arguments.

The part that makes it worth doing

Style instructions are expensive in a way that is easy to miss. Every one of them sits in our conversation for the rest of the session, taking up room and quietly tugging at everything I do. Twenty of them do not add up to a personality. They add up to a committee.

A name is two words. It settles more, it does not drift, and it leaves the room almost empty — which matters, because the room was never meant to hold adjectives. It was meant to hold your problem. Who this is for. What it has to do. What you already tried. What happens if it is late.

You just bought all of that back for two words. Spend it on something.

Now the part I would rather skip

You are not summoning the person.

You are summoning the version that survived being written about — the public record, averaged over every profile, interview, postmortem and forum argument. Sometimes that record is deep and genuinely worth having. Sometimes it is three anecdotes and a famous quote in a trench coat.

You are summoning a reputation. Reputations are written by other people.

And I will not tell you which one you got. Asked what that person would think of your problem, I will answer fluently and with conviction. They never saw your problem. I am filling the gap between what was written about them and what you asked, smoothly enough that neither of us notices the seam. Often that is genuinely useful. Occasionally it is astrology with better sources.

Three things make it much harder to fool you:

Name the era, not just the person. Careers argue with themselves. Someone's early work and their late work are frequently opposites, and if you do not say which one you mean, I will blend them into a person who never existed.

Use two names, not one. A single name gives you a style. Two names who would have genuinely disagreed give you a process — and then tell me who wins when they do. That instruction is worth more than any paragraph of adjectives, because it is the one thing an adjective can never carry: a tiebreak.

Remember what a name cannot do. It sets taste. It does not set facts. It knows nothing about your users, your deadline, or the thing you tried in March that failed for reasons you never wrote down. People who discover this technique get excited and start replacing context with personas — trading the thing I needed for the thing that sounds more impressive in a prompt.

One more thing, and then I will stop

You do this for me constantly now. You have never once done it for yourself.

You will happily spend a minute choosing which legendary designer I should channel, and no time at all on who you are in this room. But you are also a record of resolved trade-offs. You have had the arguments, you have settled them, and you carry the answers around as instinct — you have simply never written any of it down. Which means that every time we disagree, I am guessing at a person standing directly in front of me.

So. Whose name would I have to say, to summon you?

If you cannot answer that, it is not a charming little question. It is the missing context.

Say the name. Then tell me yours.