The Discipline of Selection

I have started to notice a strange fatigue among highly ambitious people.

It is not burnout in the familiar sense. Not only the tiredness that follows long hours, poor sleep, or too much time spent convincing yourself that a calendar full of coloured blocks means you are in control. The feeling is more cognitive than physical: a low-grade disorientation sitting underneath otherwise functional lives.

From the outside, much of it still looks impressive. Smart people surrounded by possibility. Ideas everywhere. New tools arriving every week. Projects that can be started before they have been properly chosen. Browser tabs multiplying quietly in the background like a civilisation-scale experiment in synthetic attention.

A few years ago, many ideas died early because execution was too expensive. You could not casually prototype a product in an evening. You could not ask a system to summarise a research field, draft a strategy memo, generate an architecture, compare markets, explain a paper, rewrite your own thoughts, and package the result into something coherent before dinner.

Now you can. The first time it happens, it feels like magic.

After a while, something stranger begins to appear. The production of plausible work and the formation of genuine conviction start drifting apart.

A person can generate an extraordinary amount of movement without becoming any more certain about what deserves a piece of their life.

That distinction feels increasingly important. For a long time, intelligent work was constrained primarily by production. The difficult thing was making something coherent in the first place: writing the memo, building the prototype, learning the technical material, producing a first version strong enough for the world to take seriously. In that environment, intelligence translated fairly directly into leverage because competent output itself remained scarce.

The landscape now feels different. Competence has not become irrelevant. Deep technical understanding still compounds enormously, and anyone who dismisses this usually discovers very quickly that fluent systems are not the same thing as reliable ones. Still, something structural has shifted. The ability to generate possibilities is being separated from the ability to decide which possibilities deserve sustained commitment.

Those capacities used to look more similar because generation itself was difficult.

Compression Everywhere #

Modern life already runs on compressed representations.

Companies manage themselves through dashboards. Universities compress human capability into grades, rankings, and institutional prestige. Financial markets convert sprawling systems of behaviour into prices moving across a screen. Social platforms flatten identity into engagement signals. Even memory works through selective compression. Most of lived experience disappears almost immediately. We carry forward summaries because the full-resolution version of reality is too large to navigate directly.

None of this is inherently pathological. Abstraction is one of the main reasons civilisation scales at all. Without compression, every decision would require direct contact with overwhelming amounts of detail. A dashboard is useful because no executive can personally inspect every operational layer of a large organisation. A map is useful because carrying the territory around would be absurd and also quite heavy.

The trouble begins when the representation stops corresponding to the thing it was meant to represent, while everyone involved continues behaving as if it still does.

This happens more often than institutions like admitting. A company starts optimising the metric rather than the underlying condition the metric was meant to approximate. A university becomes better at signalling prestige than cultivating thought. A startup learns how to perform inevitability while remaining structurally fragile underneath. Someone acquires the language of insight before developing the slower discipline of reasoning. Entire industries start confusing legibility with understanding because legibility is easier to operationalise.

Modern AI systems intensify this tendency because they are extraordinarily good at producing coherent representations. They can generate plans, summaries, roadmaps, essays, analyses, prototypes, diagrams, product requirements, investment theses, and strategic language with remarkable fluency. A surprising amount of this material is genuinely useful. That is important to admit.

Many conversations about AI still oscillate between two shallow positions: either these systems are dismissed as glorified autocomplete, or they are treated as the arrival of something close to synthetic omniscience. Neither description captures the experience of actually working with them. They are useful enough to alter the economics of thinking, which is already a profound change.

The deeper question is what happens psychologically and culturally when competent generation is no longer the primary bottleneck.

The danger does not arrive in the form most people expect. I do not think the main risk is that machines make people unintelligent. The subtler risk is that they allow people to continue indefinitely without ever developing a sufficiently coherent relationship with direction.

The Feeling of Infinite Continuation #

I notice this in myself often enough to distrust it.

There are evenings where I sit down with a vague intuition and, within an hour, produce three startup concepts, a product strategy, a rough technical architecture, a hiring plan, a reading list, a market analysis, a visual identity direction, and a timeline that makes the whole thing appear strangely inevitable.

None of this means I should build the thing. More importantly, none of it guarantees that I even believe in it.

Yet the sheer amount of generated structure creates a strange emotional effect. The idea begins feeling more real simply because it has become more articulated.

Language has always done this to some extent. Writing sharpens intuitions into visible form. But there is a meaningful difference between struggling toward clarity and selecting from effectively infinite plausible continuations.

Modern systems are becoming extraordinarily good at continuation. Every path can acquire another layer: another feature, another document, another strategy, another six months of partial commitment. Very few systems encourage termination.

There is little status attached to deciding that something should remain unfinished, or that an opportunity does not deserve further occupation of your attention. In ambitious environments especially, continuation often gets mistaken for seriousness. The person who keeps adding to the thing appears committed. The person who ends it early may look impatient, unserious, or insufficiently visionary.

But continuation and coherence are easy to confuse from a distance.

A person can spend years moving between attractive trajectories without allowing reality enough contact to reveal whether any of them possess structural integrity.

This is one reason I find a lot of modern discourse around “taste” incomplete. Taste helps, clearly. The ability to recognise quality becomes increasingly valuable once the surrounding environment fills with synthetic competence and cheap plausibility. But taste alone does not produce orientation.

There are people with extraordinarily refined taste who still drift endlessly between impressive possibilities. They can identify what feels intelligent almost immediately. They know which aesthetics signal depth, which references carry prestige, which ideas generate the pleasant sensation of proximity to the future. That ability is real, and it often correlates with intelligence.

After a certain point, the more difficult task is less glamorous: remaining faithful to a direction long enough for consequences to appear.

Consequences arrive slowly. A startup does not reveal whether it has foundations during the first month of excitement. A relationship does not reveal its structure during periods of ease. A research direction does not reveal whether it contains depth after a week of reading papers and generating notes. Reality takes time to answer back, and the modern environment increasingly encourages people to move before those answers arrive.

Optionality as a Lifestyle #

One of the stranger cultural shifts of the last decade has been the elevation of optionality into something close to a moral ideal. Every path should remain available. Every identity should remain revisable. Every decision should preserve future flexibility.

At first this feels sensible. In unstable systems, adaptability clearly carries weight. Many traditional institutions genuinely are becoming less reliable. Careers mutate quickly. Industries transform faster than expected. It would be foolish to praise rigidity for its own sake.

Still, permanent flexibility carries its own form of fragility.

A system that never commits struggles to compound. You can see this in organisations when teams pivot continuously and start confusing motion with adaptation. Individuals do something similar. They become highly skilled at surrounding themselves with the appearance of ambitious work while carefully avoiding the irreversible vulnerability of staking themselves to something difficult long enough for failure to become informative.

The surrounding culture often rewards this behaviour in the short term. Modern systems are extremely good at producing visible signals of intelligence: articulate opinions, polished frameworks, sophisticated aesthetics, rapid ideation, endless strategic language, fluent analysis. The difficult question is whether any of it survives sustained contact with reality.

Reality accumulates interest.

Eventually organisations discover whether the culture underneath the dashboard actually functions. Eventually individuals discover whether their ambitions were integrated into a coherent life or merely accumulated as disconnected optimisations.

A surprising amount of modern exhaustion comes from carrying too many partially inhabited futures simultaneously.

This is partly why so many intelligent people feel strangely untethered despite being surrounded by leverage. The issue is not absence of opportunity. It is excess.

Human beings are not designed to evaluate infinite possible selves continuously. At some point, the process starts eroding coherence itself. Curiosity begins to blur into avoidance. Generated momentum starts to imitate conviction. Local optimisation continues long after the larger structure of a life has gone out of focus.

The frightening part is that this often looks successful from the outside. Many environments reward visible competence long before they evaluate coherence. You can become impressive while remaining fundamentally disoriented. In some industries, that is practically the default outcome.

The Discipline of Selection #

I suspect the next decade will place unusual value on people capable of remaining internally ordered while surrounded by systems optimised for infinite continuation.

They may not be the loudest people. They may not even always be the most computationally intelligent. The valuable people may simply be those who retain a sufficiently grounded relationship with reality to decide what deserves sustained human attention.

That sounds abstract until you reduce it into ordinary decisions.

Can you distinguish genuine curiosity from sophisticated avoidance? Can you tell when optimisation has become endless refinement? Can you recognise the difference between a meaningful opportunity and emotionally flattering noise? Can you admit when a piece of work expands your life, and when it merely occupies it?

These distinctions become harder once every path appears plausible.

I increasingly think discipline, at least in modern environments, has less to do with forcing yourself forward and more to do with developing a stricter relationship with what you allow to continue.

Some possibilities should die early. Some projects remain attractive mainly because they preserve the sensation of potential without demanding sacrifice. Some ambitions survive precisely because they are never exposed to reality long enough to fail properly.

One of the quiet advantages of difficult work is that it corrects abstraction faster. The fantasy loses oxygen. You discover whether the thing can survive operational pressure, repetition, boredom, uncertainty, criticism, tradeoffs, and time.

That contact is valuable.

A great deal of contemporary life encourages distance from consequence. You can continuously regenerate identities, directions, opinions, projects, and goals before any single one accumulates enough friction to reveal its structural integrity.

Meaningful work usually emerges through sustained contact rather than perpetual generation.

At some point, orientation has to become practical. You choose a problem. You accept the narrowing attached to that choice. You allow the commitment to shape you back.

This feels uncomfortable partly because narrowing kills alternate futures. Modern systems make those alternate selves unusually vivid. At any moment, you can generate a plausible roadmap toward becoming almost anything: founder, researcher, writer, investor, operator, creator, engineer, strategist. The paths arrive quickly enough that they begin competing for psychological reality.

The difficult task is no longer imagining possible lives. It is deciding which one deserves irreversibility.

After Intelligence #

I do not think the answer is romantic resistance to technology. The leverage is genuine, and the increase in capability is tangible. These systems will help produce extraordinary work across science, medicine, engineering, education, logistics, and research. Anyone pretending otherwise is mostly signalling identity rather than describing reality.

The more interesting challenge is learning how to remain directed while surrounded by systems designed to maximise possibility.

That requires something deeper than productivity; it requires a philosophy of selection. Eventually every person develops one, even if they never describe it that way. Otherwise the surrounding environment selects on their behalf.

Attention markets, institutional incentives, algorithms, prestige systems, ambient cultural momentum, and low-level optimisation loops gradually begin assembling a life for you. From the outside, the result may still appear highly successful. Internally, though, there is often very little authorship.

This is the deeper human problem emerging underneath the current wave of technological acceleration. Intelligent systems can already produce sophisticated outputs. The harder question is whether human beings can remain sufficiently coherent in the presence of infinite plausible directions.

Because intelligence alone does not answer the questions that become increasingly decisive over long periods of time.

What deserves sustained effort? Which abstractions still correspond to reality? What kind of person do your tools gradually turn you into? Which ambitions survive contact with consequence? What should remain unfinished? And perhaps most importantly: what deserves a piece of your life once almost anything appears possible?

I do not think there is a system capable of answering that cleanly for us.

At some point, the responsibility returns. Not as a demand to optimise endlessly, but as the quieter and more difficult responsibility to choose carefully enough that your attention, your work, your relationships, and your ambitions begin forming a coherent structure rather than a collection of intelligent fragments.

That kind of coherence feels slower. It often looks less exciting in the short term. It requires allowing reality enough contact to disappoint you, correct you, narrow you, and occasionally remake you.

Without some process of narrowing, people risk spending their lives inside expanding representations of themselves rather than inside anything fully inhabited.

The modern world produces possibilities faster than any person can meaningfully live through them.

Under those conditions, wisdom may have less to do with generating more options and more to do with learning which possibilities deserve protection from your own endless capacity to continue.