The Exceptional-Excellent Paradox
The early mornings. The courses paid from your own pocket. The weekends at congresses, far from the people you love. The quiet evenings when you replayed a decision and asked yourself if you could have done better. You did all of it to become excellent. And you did.
Then, almost without noticing, the world changed the rules. The knowledge you fought for is now one question away from anyone. The result you were proud of is becoming the standard. Somewhere inside, a question you do not say out loud: Was it all for this?
It was. But not for the reason you think.
At Dental Leaders Academy, we have spent years listening to dentists in clinics and lecture halls across many countries. We hear the same silent fear everywhere. We also see the same hidden truth: what makes you irreplaceable was never only what you know. It is who you have become while learning it.
This article is about that difference. It is small in words and immense in consequence.
Read it slowly. It was written for you.
Abstract
In every culture, enthusiasm speaks in pairs. Excellent, exceptional…! The second word is meant to reinforce the first. This article argues that the pairing, once harmless, has become a category error. Excellence and exception are not two intensities of the same quality. They belong to different orders.
Excellence is a position on a shared scale. It can be measured, compared, taught, reproduced, and, increasingly, generated by machines. Exception is the absence of a scale. It describes what cannot be placed on the same axis as anything else. Evidence from the cognitive and neural sciences, together with recent data on artificial intelligence and human performance, suggests that in the AI era excellence is converging toward a commodity, while exception is relocating to the only domain that resists convergence: identity.
In medicine the distinction is not semantic. It is ethical. When excellence is exceptional, the system has failed. When identity is ordinary, the clinician has become replaceable.
1. The doublet
Listen to any enthusiastic conversation, in Bucharest, Boston, Dubai or Milan. Someone describes a colleague, a lecture, a result. The words arrive together: excellent, exceptional. Linguists would call this a reinforcing doublet, two near-synonyms stacked to raise intensity. The habit is universal because it once described a real situation. For most of history, excellence was rare enough to be an exception.
That coincidence is over. And the doublet hides it.
The etymology already contains the distinction the doublet erases. Excellere means to rise above, to stand higher on the same ground. Excellence is vertical. It presupposes a common scale on which everyone is placed, and it names the upper end. Exceptio, from ex-capere, means to be taken out. Exception is lateral. It does not stand higher in the set. It stands outside it.
The excellent is the best of its kind. The exceptional has no kind.
This is not wordplay. It is a structural claim with a precise consequence. Anything that sits on a scale can be compared, and anything that can be compared can be approached. Every point on a scale is, in principle, reachable by others. Only what has no scale cannot be reached, because there is no direction in which to travel toward it.
2. The convergence of excellence
The first empirical studies of generative AI in professional work point in one direction. AI does not mainly raise the ceiling. It raises the floor, and it narrows the distance between them.
In a controlled experiment with professional writing tasks, Noy and Zhang (2023) found that access to a language model cut completion time by roughly 40% and raised output quality by roughly 18%. The less skilled writers gained the most, and inequality between participants shrank. In a field study of over 5,000 customer support agents, Brynjolfsson, Li and Raymond (2023) observed an average productivity gain of about 14%, rising to about 34% for novices, with minimal gains for the most experienced. In a study with consultants at a global firm, Dell’Acqua and colleagues (2023) reported that participants below the median improved their performance by about 43%, those above it by about 17%.
The pattern is consistent across settings. The distribution compresses. What used to separate the top quartile from the middle is now, in part, available on demand.
Medicine shows the same movement. Within a few years, large language models went from near the passing threshold of the United States Medical Licensing Examination (Kung et al., 2023) to expert-level scores on medical question benchmarks (Singhal et al., 2025). In simulated text consultations, a diagnostic dialogue system was rated by specialists as equal or superior to primary care physicians on most evaluated dimensions (Tu et al., 2025). These are controlled conditions, not clinical reality. The direction, however, is not in doubt.
The consequence for the doublet is simple. When excellent output can be produced at scale, excellence stops being an exception. It remains valuable. It remains necessary. It no longer distinguishes.
Knowledge follows the same curve. A body of knowledge that can be written down can be retrieved. What can be retrieved is no longer a possession. It is access.
3. What the brain registers as exceptional
The human brain does not register quality. It registers deviation.
Contemporary neuroscience describes the brain as a prediction system (Friston, 2010; Clark, 2013). It continuously generates expectations about incoming input and processes, above all, the difference between what it expected and what arrived. This difference, the prediction error, is what captures attention, drives learning and is signalled by midbrain dopamine neurons (Schultz, Dayan and Montague, 1997). Novelty engages a loop between the hippocampus and the ventral tegmental area that marks events as worth remembering (Lisman and Grace, 2005).
The reverse mechanism is equally robust. Stimuli that repeat produce progressively weaker neural responses, a phenomenon known as repetition suppression (Grill-Spector, Henson and Martin, 2006). What becomes predictable becomes, neurally, quieter.
Apply this to excellence. The first time a level of performance appears, it generates a large prediction error. It is perceived as exceptional. The tenth time, the brain has updated its model. The same level of performance now matches expectation. It produces no error, no salience, no memory trace worth keeping. The performance has not declined. The perception of it has habituated.
This is why excellence cannot remain exceptional by becoming more excellent. A higher score is still a value on the same axis. The brain absorbs it into its prior within a few exposures and resets the baseline. Every gain on a scale is temporary as a signal, because a scale is exactly the kind of structure a predictive brain learns fastest.
Language habituates the same way. Superlatives repeated across a culture lose force, which is why enthusiasm keeps stacking them. The doublet excellent, exceptional is itself a symptom of habituation. One word is no longer enough.
What resists habituation is not a higher value. It is a pattern the observer cannot fully compress. The brain treats persons in exactly this way. Faces engage dedicated regions of the ventral visual cortex specialised for individual identity (Kanwisher, McDermott and Chun, 1997). We do not habituate to a person as we habituate to a number, because a person is not a value. A person is a model, and a model keeps generating new predictions, and new errors.
4. The neural signature of identity
If exception has left the domain of results, where has it gone? Neuroscience offers a precise answer. It has gone where it always was, and where it was hidden by the rarity of excellence: into the individual brain.
In 2015, Finn and colleagues showed that a person can be identified from the pattern of functional connectivity in their brain, measured with fMRI, with accuracy above 90% across separate scanning sessions. The connectome behaves like a fingerprint. The most distinctive networks were the frontoparietal ones, the regions most involved in higher cognition, judgment and the control of attention. In other words, the part of the brain that makes us most competent is also the part that makes us most singular.
This singularity is not decoration. It is built by experience. The brain reorganises itself according to what it does repeatedly, a principle demonstrated, among many examples, by the enlarged posterior hippocampus of experienced London taxi drivers (Maguire et al., 2000). No two professionals have the same history of cases, teachers, failures, languages, cultures and decisions. Therefore no two professionals have the same brain. Twenty-five years of practice do not produce a better copy of the textbook. They produce an original.
Three further mechanisms make this originality operative rather than merely anatomical.
1. Tacit knowledge. Polanyi (1966) observed that we know more than we can tell. Expertise is largely stored in forms that cannot be fully verbalised: perceptual discriminations, timing, the sense that something is wrong before it can be named. What cannot be told cannot be written. What cannot be written cannot be trained into a model.
2. Somatic markers. Damasio (1994) showed that decisions under uncertainty are guided by bodily signals shaped by one’s own emotional history. Two clinicians with identical knowledge, facing the same ambiguity, are guided by different internal maps. Judgment is knowledge passed through a biography.
3. The self-referential system. Cortical midline structures, particularly the medial prefrontal cortex, process information in relation to the self (Northoff et al., 2006). Information related to the self is encoded more deeply and remembered better. Identity is not a story told after the work. It is the filter through which the work is perceived.
A metaphor clarifies the distinction. A thousand tuning forks can produce a perfect A at 440 Hz. All are excellent. None is exceptional. Now let a cellist and a violinist play the same A. The pitch is identical. The sound is not. The difference is timbre, the unique structure of overtones produced by a particular body made of particular wood, aged in a particular way. Pitch is the measure of correctness against a standard. Timbre is the signature of an instrument.
Excellence is pitch. Exception is timbre. Machines have learned to hit the pitch. Timbre is made of history, and history cannot be downloaded.
5. Why medicine raises the stakes
In most fields, confusing excellence with exception is a strategic mistake. In medicine it is a conceptual inversion with ethical weight.
Modern medicine was built to make excellence ordinary. Evidence-based practice, clinical guidelines, protocols and the standard of care all serve one purpose: to reduce unwarranted variation, so that the quality of care does not depend on which door the patient happens to open (Wennberg, 2010). In medicine, excellence is not a summit to be admired. It is a floor to be guaranteed.
This produces the sharpest form of the paradox. In medicine, excellence that is exceptional is a symptom of system failure. If a clinician is praised as excellent, exceptional for doing what the evidence requires, the praise contains an unintended diagnosis: that the expected level of care is something less. The doublet, so natural in a congress corridor, quietly lowers the norm it means to celebrate.
Where, then, is the legitimate exception in medicine? The evidence points to the encounter itself.
The clinician is part of the treatment. In a randomised trial with patients suffering from irritable bowel syndrome, a warm, attentive and confident clinical relationship significantly increased symptom relief compared with the same intervention delivered with limited interaction (Kaptchuk et al., 2008). In an experimental study, the perceived warmth and competence of the provider changed how quickly a physical allergic reaction subsided (Howe, Goyer and Crum, 2017). Using simultaneous brain imaging of clinicians and patients, Ellingsen and colleagues (2020) found that brain-to-brain concordance, particularly in the right temporoparietal junction, increased after the clinician built rapport and was associated with greater pain relief.
These findings describe something no protocol can standardise and no dataset can contain: two specific nervous systems coupling in real time. The effect is not produced by excellent information. It is produced by a particular person, present, accountable, and trusted.
There is a finer point. Language models already produce written replies that evaluators rate as more empathetic than those of physicians (Ayers et al., 2023). This does not weaken the argument. It sharpens it. Even empathetic language, when it can be generated, joins the convergent domain. What remains singular is not the sentence. It is the one who stands behind it, carries responsibility for it, and is changed by the patient in return.
The patient does not choose a protocol. The patient chooses a person. Increasingly, the protocol will be the same everywhere. The person never will be.
6. The fineness of the paradox
The distinction, set side by side:
A paradox earns its name only if the obvious reading is wrong. Here there are three obvious readings, and all three are wrong.
First misreading: excellence no longer matters. The opposite is true. Exception without excellence is eccentricity. A cello with a rich timbre that cannot hold pitch is not an exceptional instrument; it is a broken one. Excellence has not lost value. It has changed function. It used to be the distinction. It is now the admission ticket. Necessary, and no longer sufficient.
Second misreading: become more excellent and you will become exceptional again. This is the trap of optimisation. Biology offers a precise illustration. Sharks and dolphins, separated by hundreds of millions of years of evolution, arrived at nearly the same shape. The water imposed it. When the medium defines a single optimum, everything that optimises converges. A profession that optimises only toward the shared standard produces professionals who resemble one another and, increasingly, resemble the machine. The more perfectly you meet the standard, the more interchangeable you become. The exception lies in what the medium does not impose.
Third misreading: identity is a matter of personal branding. It is not. Branding is a claim, and a claim is placed beside other claims, which reintroduces the scale. Identity, in the sense used here, is a structure: the particular way one brain integrates its knowledge, its history and its judgment into decisions. It cannot be declared, only recognised. And it cannot be copied, because imitation always takes a model, and a model is a scale.
Here lies the finest point. Exception cannot be pursued directly. Every attempt to become exceptional by outperforming others returns you to comparison, and comparison is the native territory of excellence. Exception appears only as a by-product of becoming more precisely oneself while remaining excellent. It is not won against others. It is uncovered within.
7. Identity as exception in the age of AI
For centuries, the scarcity of excellence allowed two different things to share one word. Artificial intelligence has separated them. It has done to excellence what the printing press did to the copying of manuscripts. It has not made excellence worthless. It has made it abundant.
What abundance cannot reach is the singular. Every human brain is a unique configuration of connections, sculpted by a history no one else has lived, carrying knowledge that exceeds what it can say, and judging through a body that remembers. This is not a romantic claim. It is the measurable conclusion of connectome research, plasticity research and decision neuroscience. The singular was always there. It was simply invisible while excellence was rare enough to take the credit.
The practical consequence for experienced practitioners is a shift of attention. The question How can I become more excellent? remains legitimate, and its answers are increasingly available to everyone. The question that now distinguishes is a different one: What, in the way I practise, could not have been produced by anyone else?
The paradox can be condensed into a few axioms.
• Excellence rises above. Exception stands apart.
• Excellence is measured. Exception is recognised.
• Excellence can be taught. Exception can only be grown.
• What can be compared can be reached. What can be reached will be reached.
• In medicine, excellence is owed. Identity is offered.
• Excellence is the entry ticket. Identity is the destination.
So the next time enthusiasm reaches for the familiar pair,excellent, exceptional, it may be worth pausing between the two words. They no longer point to the same place. The first describes how well you did the work. The second describes who did it. In the age of AI, only the second is yours alone.
References
• Ayers, J. W., et al. (2023). Comparing physician and artificial intelligence chatbot responses to patient questions posted to a public social media forum. JAMA Internal Medicine, 183(6), 589-596.
• Brynjolfsson, E., Li, D., & Raymond, L. R. (2023). Generative AI at work. NBER Working Paper 31161.
• Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36(3), 181-204.
• Damasio, A. R. (1994). Descartes’ Error: Emotion, Reason, and the Human Brain. Putnam.
• Dell’Acqua, F., et al. (2023). Navigating the jagged technological frontier. Harvard Business School Working Paper 24-013.
• Ellingsen, D. M., et al. (2020). Dynamic brain-to-brain concordance and behavioral mirroring as a mechanism of the patient-clinician interaction. Science Advances, 6(43), eabc1304.
• Finn, E. S., et al. (2015). Functional connectome fingerprinting. Nature Neuroscience, 18(11), 1664-1671.
• Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience, 11(2), 127-138.
• Grill-Spector, K., Henson, R., & Martin, A. (2006). Repetition and the brain. Trends in Cognitive Sciences, 10(1), 14-23.
• Howe, L. C., Goyer, J. P., & Crum, A. J. (2017). Harnessing the placebo effect: exploring the influence of physician characteristics on placebo response. Health Psychology, 36(11), 1074-1082.
• Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The fusiform face area. Journal of Neuroscience, 17(11), 4302-4311.
• Kaptchuk, T. J., et al. (2008). Components of placebo effect: randomised controlled trial in patients with irritable bowel syndrome. BMJ, 336, 999-1003.
• Kung, T. H., et al. (2023). Performance of ChatGPT on USMLE. PLOS Digital Health, 2(2), e0000198.
• Lisman, J. E., & Grace, A. A. (2005). The hippocampal-VTA loop. Neuron, 46(5), 703-713.
• Maguire, E. A., et al. (2000). Navigation-related structural change in the hippocampi of taxi drivers. PNAS, 97(8), 4398-4403.
• Northoff, G., et al. (2006). Self-referential processing in our brain: a meta-analysis of imaging studies on the self. NeuroImage, 31(1), 440-457.
• Noy, S., & Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Science, 381(6654), 187-192.
• Polanyi, M. (1966). The Tacit Dimension. Doubleday.
• Schultz, W., Dayan, P., & Montague, P. R. (1997). A neural substrate of prediction and reward. Science, 275(5306), 1593-1599.
• Singhal, K., et al. (2025). Toward expert-level medical question answering with large language models. Nature Medicine, 31, 943-950.
• Tu, T., et al. (2025). Towards conversational diagnostic artificial intelligence. Nature.
• Wennberg, J. E. (2010). Tracking Medicine: A Researcher’s Quest to Understand Health Care. Oxford University Press.
#ExceptionalExcellentParadox #IdentityIsTheException #ExcellenceIsTheFloor #DentalLeadersAcademy #HumanInTheAgeOfAI
