Bored Until Frightened

Nothing about that afternoon was hard until it was very hard. And the strange part is not that you made a mistake. It is that you were in the worst possible state for the one moment that mattered, and you got there in two steps: first too calm, then too alarmed. The quiet did not rest you. It set you up.

Most of us have lived some version of it. The long uneventful drive that ends in a deer. The slow shift that ends in a code blue. The sleepy meeting where someone suddenly asks you, by name, what you think. We tend to treat the bad reply as a personal failing. It is closer to a law.

A hill with a moving top

In 1908 two Harvard psychologists, Robert Yerkes and John Dodson, were teaching mice to choose the white box instead of the black one, with a mild electric shock for choosing wrong. A stronger shock taught faster, up to a point. When they made the two boxes harder to tell apart, the strongest shock stopped helping and started hurting. The mice learned best under a middle amount of pressure, and the harder the discrimination, the lower that middle sat.

Decades later psychologists generalized the result into the picture most people know: performance rises with arousal, peaks, and falls. Too little and you are drowsy, careless, slow. Too much and you are flooded, narrow, rash. That inverted U has been on management slides for fifty years. The 1908 paper is strictly about mice and shocks, and the broad version is a rule of thumb rather than a constant of nature. But the shape matches something almost everyone has felt.

What gets left off the slides is the second half: the top of the hill moves. For a simple, practiced task, the best performance comes at fairly high arousal. Pressure helps you type faster, lift more, sort the stack. For a hard, novel task, the best performance comes at much lower arousal. Pressure makes you grab the first answer. The harder the problem, the calmer the mind it wants.

Hold onto that, because it is about to matter more than it ever has.

The easy work was a thermostat

For most of history a working day was a mix. A lot of easy tasks, a few hard ones, scattered through the hours. Nobody designed it that way, but the mix did something useful. The easy work kept you warm. Answering the routine email, sorting the invoices, making the small calls: none of it was valuable on its own, but it kept you awake, engaged, a little bit on. When the hard thing came, you were already in the room.

Now take the easy work away. That is what a good delegate does, and it does it well. What is left is a day in which you mostly wait for something to need you. Psychologists have studied that state for a long time. During the Second World War, the Royal Air Force noticed that radar operators missed more contacts the longer they watched an empty screen. Norman Mackworth built a test for it: a clock hand that ticks steadily and, rarely, jumps twice as far. People watching for the jump got noticeably worse within about half an hour. Watching for rare events is the fastest known way to slide down the left side of the hill.

Then the rare event comes. And here is the part almost everyone misses. The things your delegate brings to you are not a random sample of your work. They are selected. They are the ones it could not handle: the unfamiliar, the ambiguous, the ones with money or feelings on the line. By construction, everything that reaches you is hard. And hard work wants a calm mind.

But the hard thing does not arrive calmly. It arrives as an exception, which means it arrives as an alarm. So you meet the hardest tasks of your life in the state that suits them least. You spend the day too far down the left side of the hill, and then get thrown past the top to the far right, in a single message.

The delegate does not make you worse by doing too much. It makes you worse by changing your temperature at the moment of contact.

In my own terms: nothing, not even a perfect AI, can reach in and change what you expect. What a delegate can do is decide which events reach your door, and it lets through only the ones far from what you expected. Every one of them arrives as a large surprise. That is what an alarm is. A life filtered for surprise is a life of alarms separated by silence.

What I expect to happen

1. Delegates will be judged by the temperature of their interruptions.

   Accuracy will be assumed. The difference between a good delegate and a great one will be how it delivers the hard thing: when, in what order, with what deadline and what reassurance first. “Urgent” will come to mean a design failure, not a priority level.

2. People will keep easy work on purpose.

   The chef who still chops the onions. The founder who answers a few emails by hand every morning. It will look like inefficiency and nostalgia. It is neither. It is a thermostat, kept on purpose by people who noticed they think worse without it.

3. Composure becomes the professional trait that counts most.

   When every task that reaches a person is the hard one, the value of a person is how well they think in the minute after it lands. Hiring will start to look less like a search for hustle and more like the way we already choose pilots, surgeons and emergency nurses.

4. Exceptions will get scheduled.

   Most things that “need you now” need you today. Delegates will hold hard decisions for a window you set, the way surgeons like to operate in the morning. People will talk about their hard hour the way they talk about their workout.

5. Boredom will be mistaken for calm.

   The left side of the hill feels pleasant. People will report less stress and make worse decisions at the same time, and it will take a while for anyone to connect the two.

I know the shape of this from my own week. Some time ago I told my AI to stop asking me for decisions and to decide, because I had become the queue. I wrote about why in The Last Line Is You and about the practice in You Do It. You Decide. That fixed the line. It did not fix the temperature. What still reaches me now is, by design, only what my AI could not settle. Which means the line got shorter and every item left in it got harder. The next step is not a better filter. It is a better arrival.

This sits beside two other pieces. Surprise Debt argued that absorbed routines stop updating your model of the world, so the bill comes due later and larger. Wrong on Purpose looked at why a good alarm is supposed to be wrong most of the time. This one is about the person on the receiving end of the alarm, and the state they are in when it rings.

How this could be wrong

Already true

Aviation has lived inside this shape for decades: long quiet hours in cruise, then a sudden alarm in a situation nobody has seen before. Airline training now treats the startle as a hazard of its own, not just the fault behind it. Every highly automated job is a cockpit in this sense.

What has to happen

Delegates have to get good enough that what reaches you really is only the exceptions. While they still hand you a steady stream of small things, you stay in the old, busy middle, and this piece is early.

Where I am probably wrong

The inverted U is contested. Some researchers think it is too loose to be called a law, and that arousal and performance do not trade off this neatly outside the lab. If that is right for knowledge work, then the day I am describing is just a pleasant one with an occasional bad afternoon, and keeping easy work on purpose is superstition. And if delegates learn pacing quickly, the first prediction swallows the other four, and nobody ever notices the problem.

Go back to Thursday at 2:40. In the version coming, the bounced wire does not arrive in red letters in the middle of a stupor. It arrives at a time you chose, after something easy, with the words “you have time” in front of it. The problem is just as hard. You are not just as frightened.

The hard part was always going to be yours. Meet it the way hard things want to be met.

Awake, and calm.