Introduction

This essay pursues two questions that turn out, on inspection, to be one. The first is practical and forward-looking: can anything improve our perception of the life we have not yet lived — our foresight about marriages, careers, characters, and our own motives? The second sounds mystical but will be treated empirically: can the ego, which both the contemplative traditions and modern cognitive science identify as the chief obstacle to clear perception, be displaced — and if so, can it be displaced quickly, or only by the slow disciplines of a lifetime?

The questions converge because they share a diagnosis. Perception of human affairs fails, when it fails, in a characteristic direction: toward the self. We see our prospects through hope, other people through need, and our past decisions through vanity. Any technology for seeing better — whether it comes from the psychology laboratory, the philosophy seminar, the meditation hall, or the pharmacy — is in the end a technology for getting the self out of the way of the seen. The first half of this essay examines what is known about improving foresight: the conditions under which perceptual learning works, why ordinary life mostly fails to meet them, and what can be done to repair the failure. The second half turns to the deeper obstacle, the ego itself, and surveys the candidate methods for dissolving it — educational, philosophical, psychological, pharmacological — asking of each not only whether it works but how fast, how durably, and at what risk. The conclusion argues that the fast and slow routes are not competitors but stand in a relation more like that of a preview to a film: the fast methods show you, briefly and vividly, that the self is optional; only the slow ones make the lesson stick.

The Conditions of Perceptual Learning

The decisive evidence on whether foresight can be trained comes from an unusual episode in the history of psychology: an adversarial collaboration between Daniel Kahneman and Gary Klein, published in 2009 under the title “Conditions for Intuitive Expertise: A Failure to Disagree.” The two men began from opposed camps. Kahneman’s heuristics-and-biases programme had spent decades cataloguing the systematic errors of human judgement, and its natural conclusion was scepticism about intuition. Klein’s naturalistic decision-making programme had spent the same decades studying fireground commanders, neonatal nurses, and military officers whose intuitions were demonstrably excellent, often life-savingly so. Rather than continue past one another, they collaborated to find the boundary between their findings — and found it with surprising precision.

Intuitive expertise, they concluded, develops when and only when two conditions hold. First, the environment must be regular enough that valid cues exist: the situation must actually contain patterns that predict outcomes. Second, the learner must receive feedback that is fast, frequent, and unambiguous, so that wrong predictions are punished and right ones confirmed before memory and self-interest can rewrite the record. Where both conditions are met, the results are spectacular. Chess masters reproduce mid-game positions after a glance, because the configurations are lawful and every game ends in a verdict. Fireground commanders feel a building about to collapse without being able to say how; the cues are real and the feedback is immediate. Anaesthesiologists, radiologists in well-audited practices, weather forecasters — whose calibration is among the best ever measured, because they predict daily and are scored daily — all inhabit high-validity environments with tight feedback loops. So, famously, do poultry sexers, who learn to determine the sex of day-old chicks with extraordinary accuracy while remaining largely unable to articulate the cues they use. The skill is genuinely perceptual: they see something they cannot say.

Where either condition fails, something darker happens: experience continues to produce confidence, but stops producing accuracy. Stock pickers do not beat the market with seniority, because price movements are dominated by noise — the valid cues barely exist. Political pundits, as Philip Tetlock showed in his twenty-year study of expert political judgement, perform near chance, in part because their predictions are vague enough to evade scoring and their failures arrive years later, easily reinterpreted. Admissions interviewers and personnel selectors persist for whole careers in the conviction that they can read a candidate in half an hour, against decades of evidence that unstructured interviews barely predict performance; the feedback — how the rejected candidates would have done — never arrives at all. The general lesson is uncomfortable: experience is not a teacher. Experience plus valid cues plus prompt honest feedback is a teacher. Remove any element and what experience produces is not wisdom but seniority.

The most sobering datum in this literature concerns psychotherapy. Experienced therapists, across many studies, barely outperform novice therapists on patient outcomes. This is not because therapy fails — it helps — but because the therapist’s learning environment is broken. Feedback on a patient’s life arrives over years, confounded by everything else in that life, and filtered through the patient’s wish to please and the therapist’s wish to have helped. The domain closest of all to “perceiving human lives” turns out to be one in which decades of practice teach remarkably little. Anyone who hopes that simply living attentively will make them wise about lives should sit with that finding.

Life as a Low-Validity Environment

The Kahneman–Klein framework, applied to life as a whole, yields a bleak audit. Consider the structure of the judgements we most want to get right. Whether to marry this person: the outcome unfolds over decades, the feedback is hopelessly confounded (was the divorce caused by the choice, or by what happened after?), and we make the judgement at most a handful of times, which is no sample at all. Whether this career will suit the person we are becoming: the person doing the predicting will not be the person living the result. Whether a friend’s reassurance is honest: the truth may never surface. Whether our own motives are what we say they are: here the feedback mechanism is not merely slow but actively adversarial, since the self that would have to register the error is the same self that benefits from not registering it.

Life, in short, is a low-validity environment with corrupt feedback. This single fact explains a great deal that otherwise seems paradoxical: why intelligent people make catastrophic personal decisions; why the old are not reliably wiser than the young; why every generation must relearn the same emotional lessons despite the libraries written by the last one. It is not that humans are stupid. It is that the conditions under which human perceptual learning works — the conditions that make chess masters and weather forecasters — are precisely the conditions that the most important domains of life fail to provide. The expertise machinery is intact; the curriculum is missing.

Prosthetic Feedback

What demonstrably works is anything that artificially repairs the feedback loop, and the proof of concept is Tetlock’s second great research programme. The Good Judgement Project, run within a forecasting tournament sponsored by the American intelligence community, identified “superforecasters” — ordinary citizens, retired engineers and pharmacists among them, who consistently outperformed intelligence analysts with access to classified information. More importantly, the project showed that the skill could be taught. A brief training module in probabilistic reasoning improved forecasting accuracy by roughly ten percent, and the improvement persisted. The ingredients were unglamorous: start from the outside view (ask what usually happens in situations of this class, before asking what feels true of this one); break big questions into smaller estimable parts; update beliefs in small increments as evidence arrives; and above all, keep score — make predictions explicit, with numbers and dates, so that being wrong becomes undeniable. The active ingredient across all of these is the same: the conversion of life’s vague, self-flattering feedback into fast, scored feedback. Superforecasters are not cleverer than the rest of us so much as they are people who have arranged to be told, promptly and numerically, when they are wrong.

Once the principle is stated, its other instances become visible everywhere. The decision journal — writing down, at the moment of choice, what you expect to happen and why — exists to defeat hindsight bias, the memory’s habit of revising past beliefs to match outcomes; it is a way of trapping your past self’s predictions so your future self can score them. Gary Klein’s premortem asks a team to assume the project has already failed and explain why — a framing that licenses dissent and surfaces foreseeable failure before it occurs, generating feedback from a future that has not yet happened. Red teams institutionalise the adversarial perspective; surgical checklists, in Atul Gawande’s telling, substitute reliable procedure for the surgeon’s overconfident perception. Each of these is a prosthetic feedback channel strapped onto a domain whose natural feedback is too slow or too kind.

And so, notably, is another person. A trusted friend, an honest spouse, a good supervisor — these are, epistemically considered, real-time error signals: the only corrective most lives contain that operates at the speed of the mistakes. One strong and rarely stated argument for marriage and deep friendship is precisely this: not companionship but calibration. The person who has arranged to be honestly contradicted has repaired, in the one way available, the corrupt feedback loop of self-perception. It follows — uncomfortably — that surrounding oneself with agreeable people is not merely a social indulgence but an epistemic injury, the equivalent of a forecaster who refuses to be scored.

The Limits of Calibration

Two objections mark the boundary of this whole approach, and they deserve full weight. The first is a version of Goodhart’s law — the observation that when a measure becomes a target, it ceases to be a good measure — applied to wisdom itself. The calibration toolkit works on questions that can be scored: outcomes with dates, events that either happen or do not. It may therefore quietly redirect attention toward the scoreable. One becomes well-calibrated about whether the project will ship and the market will turn, while the unquantifiable things — what this job is doing to your character, what this relationship is doing to the person you love — remain exactly as invisible as before, now with the added camouflage of a justified confidence in one’s own judgement. Optimising the measurable is itself a perceptual distortion, and the most dangerous kind, because it feels like rigour.

The second objection is deeper and comes from the philosopher Iris Murdoch. Perhaps “forward-looking perception” misdescribes the goal. Prediction is about controlling the future; attention is about being adequate to the present. In The Sovereignty of Good, Murdoch gives a famous example: a mother-in-law, M, who sees her son’s wife, D, as vulgar, juvenile, tiresomely undignified — and who behaves impeccably toward her all the while. Nothing in M’s conduct needs correcting; what needs correcting is her seeing. Over time, by deliberate moral effort — “looking again,” justly and lovingly — M comes to see D as refreshingly simple, spontaneous, youthful. Nothing observable has changed; everything has changed. Murdoch’s point is that the moral life is not primarily a series of choices but a continuous fabric of perception, and that the quality of our attention between decisions determines what we are able to see when decisions come. On this view the person who perceives well is not the one who saw the betrayal coming but the one who sees the person in front of them without fantasy. And the obstacle to that seeing is not a missing feedback loop. It is, in Murdoch’s blunt diagnosis, “the fat relentless ego” — the self’s continuous, anxious, self-serving dream, which she regarded as the natural condition of the human mind. Which raises the question the rest of this essay pursues: if the ego is the obstacle, what actually removes it — and how fast?

The Self as a Hypothesis

Modern cognitive science gives Murdoch’s diagnosis an unexpected mechanical underpinning. On predictive-processing accounts of the brain — now among the dominant frameworks in computational neuroscience — perception is not passive reception but controlled hallucination: the brain continuously generates a model of the world, predicts its sensory input, and updates the model when prediction and input disagree. Within this architecture, the self is not an observer standing behind the model. It is part of the model — plausibly its highest-level component, the master hypothesis that organises prediction across every domain: this body is mine, these thoughts are mine, this life is happening to me. Thomas Metzinger’s philosophical version of the claim, developed in Being No One, is that what we call the ego is a phenomenal self-model, a representation so transparent that we look through it rather than at it, mistaking the model for a thing.

The framework makes a striking prediction: anything that relaxes the brain’s high-level priors should loosen the sense of self. And this appears to be exactly what happens. Robin Carhart-Harris’s REBUS model — “relaxed beliefs under psychedelics” — proposes that psychedelic drugs act precisely by reducing the weight of high-level priors, allowing bottom-up sensory information and ordinarily suppressed patterns of activity to flow upward unconstrained. The self-prior, as the topmost belief, is the most conspicuous casualty: hence ego dissolution, the signature experience of high-dose psychedelics. The neuroimaging is suggestive in the same direction. The default mode network — a set of midline brain structures most active during self-referential thought, rumination, and mental time-travel — shows reduced activity and disrupted integrity both in people on psilocybin and in long-term meditators. Whatever their other differences, the pharmacological and contemplative routes appear to converge on the same neural territory, and on the same computational operation: a slackening of the self-model’s grip on perception. The traditions that spoke of dissolving the ego were not, it turns out, speaking metaphorically. They were describing an adjustable parameter.

The Fast Routes

If the self is an adjustable parameter, the question of speed becomes empirical. The fastest known adjustment is pharmacological. Psilocybin and LSD produce dose-dependent ego dissolution within an hour or two — an effect reliable enough that psychometricians have built a validated instrument, the Ego-Dissolution Inventory, to measure it. At the extreme, 5-MeO-DMT, smoked, produces in many users a total dissolution of the subject–object boundary within seconds, lasting perhaps twenty minutes. The clinical research programme that grew from the Johns Hopkins psilocybin studies — beginning with Roland Griffiths’s landmark 2006 paper — found that a majority of volunteers rated a single high-dose session among the most personally meaningful experiences of their lives, comparable to the birth of a child, and that measured increases in the personality trait of openness persisted at fourteen-month follow-up. For a trait psychologists had regarded as essentially fixed in adulthood to move after a single session is, whatever else one says, remarkable.

The non-pharmacological fast routes form a descending series. Intensive breathwork of the holotropic type — sustained voluntary hyperventilation — can approximate psychedelic phenomenology, probably through CO₂ depletion and the resulting changes in blood pH and cerebral circulation; it is legal everywhere and dose-controlled by the simple expedient of breathing normally again. Flotation tanks remove the sensory input against which the body-model is continuously checked, and reliably soften body boundaries within an hour. Ten-day silent retreats in the Goenka tradition — slow by pharmacological standards, blisteringly fast by contemplative ones — compress years of ordinary practice into a single immersion and routinely produce at least transient self-transcendence in beginners. Dacher Keltner’s research programme on awe documents a milder but repeatable “small self” effect: vast landscapes, great music, the night sky, and collective ritual all measurably shrink self-focus and increase prosocial behaviour, at a cost of nothing but attention. Near-death experiences produce perhaps the most durable ego-restructuring on record — survivors show lasting drops in self-concern and death anxiety — but are, mercifully, not schedulable.

One further route deserves separate notice because it is the only one that works through argument: philosophy. David Hume reported that when he looked inward he could never catch a self, only particular perceptions — heat, cold, love, hatred — and concluded that the self is a bundle, not a thing. The Buddhist doctrine of anattā reaches the same conclusion by meditative rather than introspective-sceptical means. The modern landmark is Derek Parfit’s Reasons and Persons, which argues with great technical care that personal identity is not what matters — that the deep further fact we imagine connecting us to our future selves does not exist, only overlapping chains of memory and character. Parfit reported that internalising his own argument changed his life: where he had previously felt he was moving through a glass tunnel, walled off from others and rushing toward death, the walls of the tunnel disappeared; he came to care less about his own death and more about others, since the boundary between his life and theirs now seemed a matter of degree. The philosophical route is slow and works for few — most readers assent to the argument and feel nothing — but it is unique in one respect: when it works, the change arrives already integrated with one’s beliefs, requiring no reconciliation between an overwhelming experience and a worldview that cannot accommodate it.

State Is Not Trait

Now the central complication, on which the whole question of speed turns. Everything the fast methods produce is a state: an acute, temporary condition. What Murdoch wanted — what the question of ego dissolution is really about — is a trait: a durable reduction in the self-centeredness of ordinary perception, operative on a Tuesday afternoon, unaided. And the relation between the two is much weaker than either the psychedelic evangelists or the retreat brochures suggest.

The evidence is consistent on this point. In the clinical psilocybin trials, lasting benefit is predicted not by the raw intensity of the ego dissolution but by the qualities of the experience (the “mystical” factor structure: unity, sacredness, ineffability) and, crucially, by the integration work that follows — the slow, ordinary, effortful translation of a six-hour state into changed habits of attention and behaviour. Strip away the integration and the state fades like any other memory. Daniel Goleman and Richard Davidson’s survey of the meditation research, Altered Traits, draws the same line from the other side: most measurable effects of casual meditation practice are state effects that decay within hours, and genuine trait change — the kind visible in a practitioner’s baseline brain function, off the cushion, unwarned — appears only after thousands of hours, with the most dramatic findings reserved for yogis with tens of thousands. The contemplative traditions encoded the point in a slogan the data now support: states can be induced; traits must be grown. A six-hour preview of selflessness no more makes a selfless person than a flight simulator makes a pilot — what it makes is someone who knows what the destination looks like.

The Dangers of Speed

Three further considerations darken the case for the fast routes, and they are not minor. The first is destabilisation. The predictive-processing frame itself explains why: the self-model is not an ornament but the load-bearing structure of a mind, and dissolving it in a person whose structure is fragile is not enlightenment but something adjacent to psychosis. Psychedelics carry real risks for those predisposed to psychotic illness; and lest meditation seem the safe alternative, Willoughby Britton’s research at Brown has documented a substantial population of practitioners — many with no prior psychiatric history — who suffered serious, sometimes years-long destabilisation from intensive practice: depersonalisation, fear, the loss of a functional sense of self that the traditions themselves know as the dark night. The traditional path’s slowness, this suggests, was never inefficiency. The decades of ethical training, community, and graduated practice that surround meditation in its home cultures are the scaffolding that lets a self be dismantled without collapsing — a container built before the explosive is handled. The modern consumer of ego dissolution, ordering the explosive without the container, is running an experiment the traditions deliberately declined.

The second danger is subtler and more corrosive: the ego’s capacity to annex its own dissolution. Chögyam Trungpa named the pattern spiritual materialism — the conversion of spiritual attainment into ego-property, the self-collecting experiences of selflessness as trophies. The phenomenon is familiar to every meditation teacher and increasingly to every psychedelic clinician: the practitioner who returns from ego death more self-absorbed than before, now with a credential; the dinner-party mystic; the man who has seen through the illusion of self and would like to tell you about it at length. The structure of the trap is worth stating precisely, because it is Goodhart’s law again, applied one level deeper than before. When ego dissolution becomes a target, the pursuit of it becomes an ego project — the most refined available. The self cannot will its own absence, because the willing reinstates what it would remove; this is why the traditions speak of grace, surrender, and gift rather than achievement, and why deliberate, scheduled, optimised ego-dissolution has a paradox at its core that no dosage can resolve.

The third consideration is the gap between peak experience and ordinary kindness. The point of displacing the ego, on Murdoch’s account, is to see other people justly and treat them well. The empirical correlation between dramatic dissolution experiences and that everyday capacity is positive but far weaker than expected — and weaker than the correlation with the unglamorous variables: integration practice, community, the daily disciplines of attention. Meta-analyses of meditation’s effects on prosocial behaviour find real but modest effects. Whatever produces saints, it does not appear to be peak experiences alone.

Conclusion

The two halves of this essay end in the same place, and the symmetry is the conclusion. Foresight about life cannot be conjured, but it can be engineered — by repairing feedback: kept scores, outside views, premortems, decision journals, and honest companions who contradict us in time for it to matter. Ego cannot be willed away, but it can be displaced — by long attentional training, by the sustained love of particular people, and, in Simone Weil’s grim candidacy, by affliction; the fast methods, from psilocybin to the breath, are best understood not as shortcuts but as previews, demonstrations that the self-prior is optional, valuable precisely insofar as they motivate the slow work and dangerous insofar as they substitute for it. In both domains the pattern is identical: the genuine article is produced by humble, repetitive, scored, accountable practice, and the counterfeit is produced by intensity — the vivid forecast confidently announced, the vivid experience proudly possessed. In both domains the principal enemy is the same: a self that would rather be impressive than accurate.

And in both domains there is a deeper unity still. A feedback loop, honestly kept, is an ego-displacement device: it forces the self to submit its dearest beliefs to correction by the world. A practice of attention, honestly kept, is a calibration device: it replaces the self’s anxious dream with what is actually there. The forecaster keeping score and the contemplative on the cushion are running the same essential operation at different scales — arranging, deliberately and against the grain of their own minds, to be corrected. Perhaps that is the most compact definition of wisdom this material affords: the organised willingness to be wrong. It can be glimpsed in an afternoon. It can only be acquired slowly.