The Effort

At the great London chess tournament of 1851, Elijah Williams of Bristol became notorious. His rival Howard Staunton, who suffered through some of the longest of those afternoons, accused him of averaging two and a half hours over a single move, and never forgave him. There was no rule against it. Thinking was free, so Williams bought as much of it as he liked.

The game had to invent a way to price it. Sandglasses appeared in 1861, first in a match and weeks later at the Bristol tournament. The mechanical chess clock, two dials on a seesaw, arrived at the London tournament of 1883 and has governed serious chess ever since.

Notice what the clock did not do. It did not make anyone cleverer or duller. The same grandmaster is a different player at one minute a game and at two hours, and everyone understands why. The brain is fixed. The clock is not.

Your AI now has a clock, and you hold the dial.

The control sits near the model picker and goes by names like effort or thinking time. The natural reading is a quality slider: high effort summons a smarter machine, low effort a lazier one, and turning it down means accepting a worse brain to save money. Under that reading, anyone with an important question should push it to maximum and leave it there.

That is not what happens.

Part two opened the rough book: a reasoning model writes private working before its fair copy. The effort dial sets how much of that working the model is allowed to do. The model itself does not change. Its weights, its knowledge, its habits are identical at every setting. What changes is the clock: how long it may sit with your question, drafting and checking in the scratchpad, before it must answer. High effort buys a long think, more tokens, more money, more seconds. Low effort buys a move in blitz.

The surprising part is how far time goes. Researchers at Berkeley and DeepMind showed in 2024 that a small model, given a well-spent thinking budget, can outperform a model fourteen times its size answering directly. Hugging Face reproduced the effect in the open: a three-billion-parameter model beating a seventy-billion-parameter one on a hard mathematics benchmark, purely by thinking longer. The trade goes the other way too. When Anthropic shipped an effort control in late 2025, it reported that its flagship on the middle setting matched its own smaller model’s best coding benchmark score while using seventy-six percent fewer output tokens than that model had needed. The same destination cost a fraction of the ink.

Two cautions apply. More thinking is not always better: on straightforward questions, research keeps finding a point where longer working stops helping and can start hurting, the machine equivalent of talking yourself out of the right answer. And the dial cannot buy knowledge. A model that has never learned a fact will not derive it by staring longer, any more than Williams could think his way to a move in a position he did not understand.

Chess players know this vocabulary in their bones. A blitz game is the same opponent on a faster clock, and players choose the time control to match the occasion.

Run the experiment this week. Turn the effort down one notch for your routine work, summaries, rewrites, everyday questions, and see whether you can tell the difference. Most people are paying tournament rates for blitz positions.

Then spend deliberately. When the question has real steps in it, a contract to pick apart, a plan with dependencies, a calculation that must be right, turn the dial up and let the machine fill its rough book.

And when the answer disappoints, change the clock before you change the machine. A middling reply on low effort says less about the model than about the second you gave it.

Knowledge is what the model has, and effort is what you give it. The big model knows more, but a small model with time on the clock keeps working the answer out.

Next week, the picker. You will learn the four differences that count, none of them printed on the name, and what happened the day the menu tried to choose for everyone.