The Smallest Lever

After two decades of studying high performers, a group of New York University researchers reached a conclusion that should bother anyone with a vision board. Vividly imagining your success can make you less likely to achieve it. The brain hands out the reward for the fantasy before any work is done, and the motivation to do the work quietly drains away.

That is the uncomfortable starting point for a talk by Sandeep Swadia, a former monk turned MIT graduate and CEO, who argues that meaningful change almost never comes from ambition. It comes from small, boring habits that protect the three things a life actually runs on: attention, energy, and happiness. He offers nine of them. What he leaves open is the hardest part of any habit, which is keeping it alive after the motivation that started it is gone. That gap is where AI now earns its place.

Why small beats ambitious

A friend of Swadia's has climbed Mount Everest more than once. His advice for the summit was simple: do not look up. Above 8,000 meters, where the oxygen is too thin to sustain you, looking at the whole mountain breaks you. You keep your eyes down, take one step, pause, gather yourself, then decide on the next step.

Big goals do the same thing to the brain at sea level. A marathon, a book, a business, a set of fixed finances all look like the full mountain from a distance, and the brain responds by pretending it can climb the whole thing at once. Then nothing starts. The job of a habit is to stop that pretending and turn ambition into a small set of actions that get done on ordinary days, not only on inspired ones.

The real lever in a life is not a bigger goal. It is the smallest action you can repeat when you do not feel like it. AI is most useful as the patient system that keeps that small action alive, not as the thing that sets the goal.

Attention: closing the open loops

Swadia's first three habits defend attention. "Correct after the click" accepts that you will reach for the phone or the ice cream before you can stop yourself, so the useful move is to attach one small good action the instant you notice the slip. A pilot's autopilot works the same way. The plane is almost never exactly on course; it flies a series of small corrections and still arrives. "Control your input" batches email and messages into a few fixed windows and handles each item once with a simple rule: do it, delegate it, date it, or delete it. The aim is to close the open loops that, as the psychologist Bluma Zeigarnik found, keep draining attention in the background until a task is finished. "Write to think" uses handwriting to clarify a decision, a habit Jeff Bezos built into Amazon's memo culture and that note-taking research supports.

Attention is where AI helps most directly, because most of what erodes it is unfinished administrative work. An assistant can sort a full inbox into those four buckets, draft the replies you would have written, and put the snoozed items on a real date, which leaves you a short list of things only you can decide. Used as a partner you write to rather than a search box, it can take a foggy problem and force the three honest questions: what do I know, what am I assuming, what is the next step. The habit stays human. The friction around it does not have to.

Energy: working with the body in packets

The next three habits protect energy. "Quantize the action" shrinks a habit to its smallest possible unit, because the hardest part of any habit is starting it. Write one line. Do half a push-up. Walk for one minute. This is the core of both BJ Fogg's Tiny Habits and James Clear's Atomic Habits: reduce the force required to begin, and consistency follows. "Cut caffeine six hours before bed" respects a fact most people ignore. Caffeine has a half-life of about six hours, so an afternoon coffee leaves a quarter of its dose in your system near midnight, and a controlled study found that caffeine taken even six hours before bed measurably degraded sleep. "Ride the waves" works with the brain's ultradian cycles, the 80 to 120 minute rhythm of focus and fatigue first described by the sleep researcher Nathaniel Kleitman. The afternoon crash is a signal, not a character flaw, and the fix is a genuine recovery block with no new input before the next wave of work.

Here AI acts as a scheduler and a reducer. Ask it to break a daunting task into a two-minute first move and it will, which is exactly the quantizing the habit requires. Point it at your calendar and it can defend a focus block and a real recovery block instead of stacking meetings through your natural dip. It can hold the caffeine cutoff as a plain rule and remind you before the window closes, not after. None of this needs a wearable or an elaborate system. It needs a tool that remembers the rule on the days you would rather forget it.

Happiness: the habits AI cannot do for you

The last three habits are about happiness, and they expose the limits of any tool. "Look up" means going outside at night and seeing that the light from many of those stars left them thousands of years ago, that some are already gone, and that you are looking at a memory of a universe almost fourteen billion years old. "Look in the mirror" turns that same attention inward: the atoms that built those stars built you, and there is exactly one version of you. The first habit produces awe, the second gratitude. The ninth and final habit is self-forgiveness, because every system eventually breaks, and the spiral of shame after a missed habit costs far more than the missed habit itself. You do not have to be your best self on your worst day.

These resist automation, and they should. An AI can prompt the nightly gratitude entry, ask the one question that reframes a hard day, or notice from your own logs that a broken streak is not the catastrophe it feels like. It cannot feel the awe for you or grant the forgiveness. The honest position is that AI carries the mechanical weight around a habit, the reminding, the sorting, the scheduling, so that you have the attention left for the part only a person can do.

Wiring AI to a habit, not a goal

The practical move is to point AI at the action, not the ambition. A request like "build me a 90 day plan to transform my mornings" recreates the exact fantasy that drains motivation. A better instruction is narrow and repeatable:

Every weekday at 8am, give me one two-minute version
of my writing habit and nothing else.

At 2pm, ask which energy wave I am in and protect a
20-minute recovery block on my calendar.

At 9pm, ask me the three writing questions and log my answer.

If I miss a day, do not scold me. Show me last week instead.

That instruction does the quantizing, defends the recovery block, runs the write-to-think prompt, and builds in self-forgiveness. It never sets a heroic goal, because the goal was never the problem. The tool holds the structure so the person can keep taking the next step.

What leaders should do

Swadia's nine habits work because they are small enough to survive a bad week, and they fail for the same reason every good intention fails: the day you need them most is the day you have the least energy to run them. That is a systems problem, and systems are what AI is good at. The opportunity is not to outsource the discipline. It is to remove the friction that made the discipline necessary in the first place. The same logic scales past one person. A team does not change through ambitious vision statements either. It changes through small repeated actions that survive the bad quarters, held in place by systems that keep working after the motivation fades.

1. Pick the single smallest action that, repeated, would move your most important goal, and assign an AI assistant to surface only that action each day. Drop the multi-step plan.
2. Route administrative inputs through a do, delegate, date, or delete triage that your AI runs first, so your attention reaches only the decisions that need a person.
3. Put one real recovery block on every working day and let your scheduling tool defend it against meetings the way it defends a deadline.
4. Build "show me last week, do not scold me" into any tracking system you adopt, so a broken streak triggers a review instead of abandonment.