Early in Charlie Chaplin’s Modern Times (1936), an inventor and his assistants arrive in the factory president’s office with a new machine. They do not make the pitch themselves; a phonograph record does it for them, in a smooth recorded voice, and the claim is simple. The machine will feed workers at their stations, so that nobody need stop for lunch. It will abolish the lunch hour, raise production and cut overhead. The device is a marvel of small mechanisms: a turntable that brings the soup to the mouth, a pusher for the food, a corn cob turning on a spindle like a typewriter carriage, a padded arm that dabs the lips clean. The president chooses a worker for the demonstration, and the worker is the Tramp.
What follows is one of the great sequences of film comedy. The machine malfunctions by degrees. The soup goes down his front. The corn cob speeds up and grinds against his teeth. The napkin arm, apparently convinced that it is helping, slaps him across the face. The president watches, considers, and decides the thing is not practical. Nobody objects to it in principle. It is rejected because it does not work.
The joke has more than one layer, and the deepest is about time. The feeding machine is a gift of minutes. But the minutes it gives are taken from the one part of the day the worker owns, and they are given to the factory. Nobody asks the man at the bench what he would do with his lunch hour if it were returned to him in some other form. The question does not arise, because in the logic of the factory a minute saved is a minute of production, and there is no other column in which to enter it. It is a small detail of the film that in a picture whose characters otherwise speak in title cards, the voices we hear, until the Tramp’s nonsense song near the end, come out of machines: the president through a screen on the factory wall, the salesman through his record.
That scene is worth keeping in mind whenever a technology is offered to medicine as a way of giving time back. Suppose, as a purely hypothetical number and not a measured result, that a tool returned twenty minutes a day to a primary care physician. Before anyone celebrates, I would like to know where those minutes go, and who decides. My argument in what follows is that time saved by a tool becomes a benefit only after someone has decided how it will be used, so that the efficiency question is really a question of governance. Its practical form is a question about measurement, because what a health system counts determines where its saved minutes flow. I want to say, as concretely as I can, what I would count instead of visits per hour.
The Coal Question #
In 1865 the English economist William Stanley Jevons published The Coal Question, a book about whether Britain’s cheap and accessible coal would run out, and its industrial position with it. Its most durable chapter, “Of the Economy of Fuel,” took on a hope that seemed only sensible: that more efficient engines would make the coal last longer. Jevons thought the hope was a mistake. “It is wholly a confusion of ideas,” he wrote, “to suppose that the economical use of fuel is equivalent to a diminished consumption. The very contrary is the truth.”
His reasoning ran through the market. If a blast furnace could make a ton of iron with less coal, the trade would become more profitable; capital would be drawn in; the price of iron would fall and the demand for it would rise; more furnaces would be built; and in the end the greater number of furnaces would burn more coal than the old ones had saved. Efficiency lowered the cost of each unit of work, and cheaper work was demanded in greater quantity. The steam engine itself was his great example. It had become an economical way of producing power, and so it was put to work everywhere.
Economists still argue about how often, and how strongly, this happens. They distinguish a rebound, in which increased use eats part of an efficiency saving, from backfire, Jevons’s own case, in which it eats more than the whole. Both occur, and their size depends on the good, the market and the period. Jevons is better read as a warning than as a law.
The analogy to a clinic is suggestive, and it is important to see where it stops. Jevons’s mechanism ran through prices. A clinic’s runs through templates and queues. Demand for primary care in many places already exceeds supply; the waiting list is the market’s unspent appetite. When documentation becomes faster, a visit becomes cheaper to schedule, not in money but in the physician’s time, and a cheaper visit can be booked more often. The rebound in a clinic does not need a price signal. It needs only a minute that nobody has claimed, sitting within reach of a demand that is always present.
And there is a further limit to the comparison. Coal does not get tired. A ton saved at the furnace is simply a ton; it has no other life. A physician’s minute is also part of a day lived by a person, who must be capable of attention at four in the afternoon and must come back the next morning. Jevons’s arithmetic has no column for that, and neither does any measure that treats clinical time purely as a fuel.
Twenty Hypothetical Minutes #
Where could the twenty minutes go? They could become a clearer explanation at the end of a visit, the second attempt at something the patient did not quite follow. They could become a referral actually followed through instead of merely ordered. They could become the time needed to read the machine’s own drafts properly, which is not trivial if the drafts are to be trusted. They could become a lunch eaten sitting down, or an earlier drive home. They could become another appointment slot. Or they could vanish without any decision at all, absorbed by messages and small tasks that expand into whatever space appears. That last outcome is the one Jevons would predict, and the likeliest if nobody intervenes. Unclaimed time in a clinic does not stay unclaimed for long.
In practice the decision is usually made in the schedule template, the grid of slots that shapes a day before it begins: how long visits are, how many there are, which are held for urgent problems, whether any time is set aside for results and messages. The people who build templates are not villains. They answer to real pressures: patients waiting, budgets, staffing, the arithmetic of keeping a clinic open. But a template is a policy. If nobody chooses deliberately where recovered minutes go, the template chooses for everyone, and it chooses according to whatever the institution measures most closely.
Other choices sit around the template. Staffing decides whose work the minutes absorb: where nurses and medical assistants are scarce, a physician’s recovered time tends to flow into tasks that others would otherwise do. Access targets decide which number moves: an institution judged on how quickly a new patient can be seen will feel the pull of the empty slot. Protected time decides what survives: where nothing is protected, nothing is.
There is also the question of whose minutes are being counted. A tool that saves a physician time may add work elsewhere: for the medical assistant who sets it up and explains it, for the staff member who looks after the patient who would rather not be recorded, for whoever checks the tool’s output and handles its mistakes. A saving counted only in the physician’s day may be partly a transfer. The ledger of a clinic has more than one person’s name in it.
So a useful discussion would begin before the schedule changes, while the minutes still have more than one possible destination. What work has been going unfinished? Who bears the consequences when it is not done? Which improvement should the new capacity make possible? Those questions have answers in every clinic, and the answers differ. What does not differ is that if the questions are not asked, the minutes will be spent anyway.
The Stopwatch #
The modern habit of measuring work by its output per unit of time has a famous source. In The Principles of Scientific Management (1911), Frederick Winslow Taylor described his experiments at the Bethlehem Steel works, where gangs of men loaded pig iron into railway cars at about twelve and a half tons per man per day. Taylor’s assistants timed every motion with stopwatches and calculated that a suitable man, properly directed, could load forty-seven and a half. They chose a laborer whom Taylor called Schmidt, offered him $1.85 a day in place of $1.15, and told him exactly when to pick up, when to walk and when to rest. Schmidt loaded forty-seven and a half tons, and kept it up.
Taylor did not lack insight. He saw that measurement could reveal waste that habit concealed, and he understood, more clearly than many of his admirers, that planned rest raised output. Schmidt was paid more. But the book also records what the method cost. The knowledge of how to do the work moved from the worker to the planner, and the worker’s own judgment came to be treated as a source of error. Taylor described the right man for handling pig iron as one “so stupid and so phlegmatic” that he more nearly resembled an ox than any other type, a sentence that tells us what the system needed its men not to be. And he set down the creed of the whole enterprise in a line that has lost none of its candor: “In the past the man has been first; in the future the system must be first.”
Visits per hour is a pig-iron measure. It counts the loads carried. The trouble is that Taylor’s method worked for pig iron precisely because each ton was like every other ton. Visits are not like each other. A visit in which a woman’s real worry was found, explained and acted on and a visit in which it was missed weigh exactly the same on the tally. If the tally is what the system is built to raise, the system will be indifferent between them, and in time it will prefer the faster one.
I do not think anyone sets out to run a clinic on Taylor’s creed. But a measure carries a philosophy whether or not its users hold it. To measure clinical work in visits per hour is to say, in effect, that the system comes first and the visit is its unit. The twenty minutes will then be spent where the unit is counted.
The Strongest Case for Another Appointment #
People are waiting. Primary care is scarce in many places, and anyone who has tried to find a new doctor knows how far off the first available appointment can be. Waiting has costs that never appear in a clinic’s day: the symptom that worsens, the prescription that runs out, the emergency visit that a timely appointment might have prevented. A person who cannot get in is not an abstraction. Seen from the waiting list, a physician who argues for keeping saved time looks very like someone protecting his own comfort while others stand outside the door.
That charge has force, and I do not want to wriggle out of it. If a tool genuinely saves time, and a clinic is turning people away, then some of the time should go to seeing more of them. Sometimes more visits will be exactly the right use. Access is a real good, and an essay that treats it as a mere number has not understood what a waiting list is.
My objection is narrower. It is to the automatic assumption that every saving belongs to throughput, and it rests on three observations.
The first is that visits per hour counts contacts, not care. A visit that ends with a question unanswered tends to generate another contact: a return visit, a message, a call, a trip to urgent care. The British systems thinker John Seddon gave this a useful name, failure demand: demand caused by a failure to do something, or to do it right, the first time. Each such contact registers as productivity. A clinic can grow busier and less useful at the same time, and a measure of throughput will record only the busyness.
The second is continuity. In a 2022 study of more than four and a half million Norwegians, Hogne Sandvik and colleagues found that people who had kept the same general practitioner for more than fifteen years had roughly 25 to 30 percent lower odds of using out-of-hours services, of acute hospital admission and of dying during the year studied than those whose relationship was a year old. The study was observational and cannot show that continuity caused the difference. It does suggest that a measure rewarding more people seen faster, by whoever happens to be free, may be counting the wrong thing.
The third is that access already has an honest measure, and visits per hour is not it. The standard access measure is the number of days until the third next available appointment, chosen because the first two openings so often come from last-minute cancellations and say little about the real wait. If the purpose of filling recovered minutes with visits is access, then this is the number that should move, and it should be watched. A capacity argument is legitimate. It should be held to its own measure.
When the Measure Becomes the Work #
There is an old warning about counting, and it comes in two strengths. The economist Charles Goodhart observed in 1975 that a statistical regularity tends to collapse once it is relied upon for control. The anthropologist Marilyn Strathern later gave the general idea its familiar form: when a measure becomes a target, it ceases to be a good measure. That is the weaker strength, and it concerns information. The number stops telling the truth.
The stronger version belongs to the social scientist Donald Campbell, who argued in the 1970s that the more a quantitative social indicator is used for decision-making, the more it will be subject to corruption pressures, and the more apt it will be to “distort and corrupt the social processes it is intended to monitor.” Campbell’s point goes further than Goodhart’s. The measure does not only lose its meaning. It bends the work. Wherever visits per hour becomes the target, one would expect visits to shorten, problems to be divided across appointments, and questions to be deferred to the next contact, which will itself be counted. Nobody need intend any of it. The indicator does the shaping.
Ivan Illich pushed this logic to its limit. Medical Nemesis (1975) opens with a sentence designed to provoke: “The medical establishment has become a major threat to health.” The polemic overreached, as polemics do, but its most durable idea was quieter. Illich called it counterproductivity: beyond a certain threshold, an institution begins to produce the opposite of what it exists to provide. In Energy and Equity, a year earlier, he had illustrated it with the car. Add up the hours a typical American man spent driving, sitting in traffic, and working to pay for the car, its fuel and its insurance, he calculated, and he gave more than 1,600 hours a year to it in exchange for about 7,500 miles: less than five miles an hour. Measured net of everything the speed cost, the fast machine was slow.
That calculation is the right model for a time-saving tool. The question is never only how much time the tool saves in drafting. It is how much time remains once everything the saving costs has been subtracted: reviewing what the tool produces, correcting it, learning to use it, monitoring it for errors, and handling the problems that its errors cause. To take a deliberately hypothetical case, a tool that saved ten minutes of typing and required eight minutes of checking would have saved two. And if the checking is skipped so that the full ten can be claimed, the saving has not been realized. It has been converted into risk, and the risk has been moved to the patient.
The research on these tools is young, and I have not tried to summarize it here; findings will change quickly, and any single study describes only the setting in which it was done. But I can say what I would want a study to report before I believed its account of time saved. The design: whether the comparison was with the same clinicians before adoption, with a similar group that did not adopt, or with nothing at all. The baseline workload, because a saving measured in a quiet season says little about a busy one. How review and correction time were measured, whether logged, observed or estimated afterwards by the clinicians themselves, since recollected minutes and logged minutes can differ. Whether the schedule changed while the study ran, because a saving absorbed by added visits will disappear from the logs while having been perfectly real. And what happened to the quality of care and of the record while time was being saved. A figure that answers none of these questions is an anecdote with a decimal point.
I would therefore want every report of time saved, whether a study, a vendor’s claim or a dashboard, to state the net figure and to say what happened to the schedule while the measurement was taken. A gross saving announced while the template quietly fills is not a finding about efficiency. It is a finding about where the minutes went.
The Work No Tally Sees #
Some of the most valuable work in a clinic has a peculiar property: when it succeeds, nothing happens. An explanation given a second way prevents a confused call three days later, and the call never comes. A result followed up at the end of the day prevents a problem that never develops. A word with the person who will be doing the caring at home prevents a missed dose and a bad week. Prevented events leave no record. The work that prevented them is visible only as time spent, and time spent with nothing to show for it looks, on most measures, like inefficiency.
That is the mechanism by which recovered minutes are lost to the work that needs them most. Preventive work of this kind is done in the gaps of a day: between visits, at the end of a session, in the minutes when a schedule happens to run light. A tool that saves time widens the gaps. A template that refills them closes them again. When that happens nothing visibly breaks. The calls that would have been prevented simply arrive, one by one, and each is counted as demand.
Four kinds of such work seem to me most at risk. Explanation is the first: not the plan itself but the second attempt at it, after the first did not land. Coordination is the second: the calls and messages that make one clinician’s plan connect with another’s, the specialist’s office that needs a reason, the pharmacy that needs a clarification, the family member who needs to know what is expected. Continuity is the third: the slow accumulation of knowing a person, which happens only if the same clinician sees them often enough and has time to remember what was said last time. Recovery is the fourth, and the hardest to defend on a dashboard: the clinician’s own capacity, restored in the minutes that are not spent. Over a career it may be among the most consequential, because the attention a patient receives at four in the afternoon depends on it.
Not all of this can or should be turned into numbers, and there are other ways of keeping it in view. One is to ask, at the end of each session, what was not done: the call deferred, the explanation cut short, the result left for tomorrow, and to read the answers over weeks for patterns. Another is to listen to the people who see the work from the side: the medical assistant who knows which patients call back confused, the front desk who knows who could not get in. A third is to read, now and then, a patient’s own account of a visit. None of these is a measure in the dashboard sense. Each is a way of making sure that when the minutes are allocated, someone remembers that this work exists.
Lalo’s Bread #
Measurement is not only a technical matter. Every measure expresses a view about what work is for, and the moral tradition I belong to has its own view, which I find clarifying here.
Sikh teaching places great weight on ਕਿਰਤ (kirat), honest work: earning one’s living by one’s own labor, without exploitation. Kirat is not a measure of quantity. It is a measure of source and effect: how the living was earned, and what the earning did to others. The janamsakhi tradition, the body of stories about Guru Nanak’s life, puts this in a form that children can remember. At Saidpur, the story goes, the Guru stayed with Bhai Lalo, a carpenter, and ate his coarse bread, declining the feast of Malik Bhago, a wealthy official. When Malik Bhago demanded to know why, the Guru took Lalo’s bread in one hand and Bhago’s food in the other and pressed them, and from Lalo’s bread came milk and from Bhago’s came blood.
It is a story, told as tradition, and I do not offer it as history. But notice what it measures. By any output metric Malik Bhago’s table was the larger: more dishes, richer food, a bigger enterprise. The story refuses that metric entirely. It asks how the food came to be on the table and who paid for it, and it finds the answer hidden inside the product, where no tally of dishes would ever look.
The tradition also keeps a separate column for what is given. The custom of ਦਸਵੰਧ (dasvandh) is to set aside a tenth of one’s earnings for the common good. It is, strikingly, a number: a tradition concerned with the ethics of work put a figure on giving, not on production. A system that counts only output has nowhere to record what was given, and the preventive work of a clinic is largely given in this sense: done beyond the slot, entered nowhere, and the first to vanish when the minutes are reassigned, because no tally sees it.
I am not proposing that clinics adopt a religious ethic. I am observing that one serious tradition of thought about work measures it by where it comes from and what it does to others, and keeps an account of what is given as well as what is produced. Those are not sentimental criteria. They can be turned into measures.
What I Would Count Instead #
If the minutes are real, they should show up somewhere other than in the appointment count. These are the measures I would want to see before and after any tool that promises to save time. Most are not my invention; where one is, I say so.
- Evening work. In 2020 Christine Sinsky and colleagues proposed standard measures drawn from electronic health record logs, among them work outside of work: time in the record outside scheduled patient hours, per eight hours of scheduled patient time. If documentation really became faster, this number should fall.
- Undivided attention. The same set includes an aspirational measure: the share of a clinic session not spent in the record. It is crude, and still closer to what a patient notices than any count of visits.
- Continuity. The share of a person’s visits that are with their own clinician or team.
- Closed loops. The share of abnormal results and referrals that end in documented follow-up. A 2012 systematic review led by Joanne Callen found that the proportion of outpatient laboratory results without adequate follow-up ranged, across studies, from under one in ten to more than half. Recovered minutes could close loops, and this measure would show it.
- Waiting, measured honestly. The third next available appointment, watched over time, so that a claim of improved access can be checked against the measure built to test it.
- Understanding. Standard patient-experience surveys already ask whether the clinician explained things in a way that was easy to understand. Time spent on explanation should register there.
- Net supervision time. This one is my own proposal: the time clinicians actually spend reviewing and correcting machine-generated text, reported beside the time the machine is said to save. Without it, every other figure in this list can be flattered.
To these I would add one practice that is not a number. At intervals, someone should take a small sample of visits and read the resulting notes against what was actually said, asking whether the record is true. Fidelity cannot be inferred from speed, and a dashboard that shows the minutes but never the accuracy is showing half a ledger.
None of these is immune to Goodhart, and none should become a target on its own. Any of them, pursued alone, could be gamed: continuity by refusing new patients, closed loops by ordering fewer tests, evening work by shifting tasks onto others. Read together, beside the appointment count, they would at least describe a day in more than one dimension, and make it harder for one number to speak for all the others.
An Agreement Before the Minutes Arrive #
Who decides, then? Someone, explicitly, before the tool is switched on, with the people whose time it is and the people whose care it is at the table. The American Medical Association’s 2023 principles for augmented intelligence say that governance policies for generative AI should be in place before its adoption and use. A 2025 analysis in the AMA Journal of Ethics recommended measuring the effects of such tools by comparing baseline time with patients and on paperwork against what follows. The AI Risk Management Framework that the National Institute of Standards and Technology published in January 2023 treats governance as a function running through all the others. Deciding where saved time goes is a governance question, and it deserves a governance answer.
Imagine, then, an agreement written in a practice about to adopt such a tool. It would be short. It would record the baseline: the measures above, taken for some weeks before anything changes. It would state where recovered capacity is meant to go, in rough proportions agreed locally, divided among access, closed loops, proper review of the tool’s output and the recovery of the people doing the work. It would name who took part in writing it: physicians, nurses, medical assistants, the front desk, and some way of hearing from patients. It would set a date for looking again. And it would say what happens if the promised benefit fails to appear. If the net saving turns out to be small, the template reverts. If an added slot consumes the saving and creates more unfinished work, the slot goes. If protected review time prevents errors but never shows on the dashboard, the dashboard changes. If clinicians feel better but patients wait longer, that belongs in the account too, and the proportions are revisited by the same group, not by whoever happens to control the template.
Wendell Berry, the Kentucky farmer and essayist, set out in a 1987 essay, “Why I Am Not Going to Buy a Computer,” the standards he thought a new tool should meet before he adopted it. Several concern cost, scale, energy and repair; the last applies anywhere: a new tool “should not replace or disrupt anything good that already exists,” and he included family and community relationships among the good things. A clinic has its own goods that already exist: a long relationship with a clinician, a conversation that is allowed to finish, a lunch. An agreement should name them before a tool is permitted to displace them.
Berry made a related argument in an earlier essay, “Solving for Pattern,” about agricultural problems. A bad solution, he argued, solves one problem by creating others; a good one solves several at once and fits the larger pattern in which it sits. A documentation tool that relieves typing and creates a burden of review, an inflow of new appointments and a habit of unread drafts has solved part of a pattern and damaged the rest. The agreement is a way of solving for the whole of it.
What a Better Day Would Contain #
I want to resist two temptations at the end. The first is nostalgia for inefficiency. The day before such tools was not a good day. It included hours of typing after the last patient, results that waited too long, visits conducted half facing a screen. Slowness is not a virtue, and a clinic that wastes a physician’s time is not more humane for doing so. The second temptation is to assume that increased output proves an improved service. More visits, more messages answered, more notes closed: none of that shows that anyone is better cared for.
A better day would contain things that can be named and, with some effort, observed. Fewer evenings spent in the record. Results that reach the people they concern, and are acted on. Patients who could say back their plan in their own words and who keep the same clinician long enough to be known. A waiting list that shrinks by the honest measure, not by the convenient one. Notes that are true. Clinicians who stay in the work, because the work has become possible to do well. And, inside the visits themselves, the kind of quiet attention I tried to imagine in A Visit Without Typing, which is impossible to schedule into a day that has no room left in it.
A physician should not keep every recovered minute, any more than a template should take every one. I would only ask that some of the gain stay available for thinking, some for the people already on the list, and some for living. It can sound impolite to mention the last in a profession organized around other people’s needs. Yet the person who leaves the clinic in the evening is the person who must return in the morning, and the quality of his attention tomorrow is part of the care of tomorrow’s patients. If documentation reliably ended twenty minutes sooner, I would like some of those minutes to go to something no dashboard would recognize: a walk while the long Valley light still lies across the streets, or an ordinary conversation that nobody has to cut short.
Modern Times ends with the Tramp and the young woman he has fallen in with walking away down an open road at dawn, out of the factory’s reach. It is a lovely ending and a cinematic one. Medicine cannot take that road; the patients are here, and the work is worth doing. But it can do what the president of the factory never thought to do. It can ask, before the machine is switched on, whose minutes these are and what they are for, and it can keep a record of the answer. Twenty minutes can be consumed so quietly that nobody remembers having received them. Before the first one is saved, someone should write down where it is going. I would like to be in the room when they do.
Evidence and Editorial Note #
Documented findings. The essay relies on the author’s public biography: a family physician practicing in California since 2010, and a Sikh. Jevons’s argument in The Coal Question (1865), chapter VII, “Of the Economy of Fuel,” including the quoted sentences and the blast-furnace reasoning: public-domain text, quoted from standard knowledge; the description of the paradox: Yale Energy History. The distinction between rebound and backfire: standard energy-economics usage. Taylor, The Principles of Scientific Management (1911): the Bethlehem Steel pig-iron account (about 12½ to 47½ tons per man per day; $1.15 to $1.85 a day), the description of the suitable worker, and the quoted creed, from the public-domain text as standardly reported. Modern Times (Charlie Chaplin, 1936): the feeding-machine sequence and the film’s ending, described from the film. John Seddon’s concept of failure demand: standard systems-thinking literature. Lower odds of out-of-hours use (OR 0.70), acute admission (OR 0.72) and death (OR 0.75) for patients with more than 15 years versus 1 year with the same regular GP, among 4,552,978 Norwegians, in an observational study: Sandvik, Hetlevik, Blinkenberg and Hunskaar, British Journal of General Practice, 2022 (linked). Goodhart’s 1975 formulation and Strathern’s 1997 generalization: standard reference knowledge. Campbell’s law, from “Assessing the Impact of Planned Social Change” (1976): one brief quotation, from standard knowledge. Illich, Medical Nemesis (1975), opening sentence and the concept of counterproductivity, and the car calculation in Energy and Equity (1974): from standard knowledge of the texts. Work outside of work and undivided attention among seven proposed core EHR-log measures: Sinsky and colleagues, Journal of the American Medical Informatics Association, 2020. Failure to follow up outpatient laboratory results ranging from 6.8 to 62 percent across studies: Callen and colleagues, Journal of General Internal Medicine, 2012, as reported by UNSW. Third next available appointment as a standard access measure and the reason for using the third opening: Institute of Medicine, Transforming Health Care Scheduling and Access: Getting to Now, 2015. The CAHPS Clinician and Group survey item on explaining things in a way that was easy to understand: published survey items. The AMA’s November 2023 principles for augmented intelligence; Gerke and Simon, AMA Journal of Ethics, November 2025; NIST AI Risk Management Framework 1.0 (January 26, 2023): the published documents, described in general terms. Wendell Berry, “Why I Am Not Going to Buy a Computer” (1987), one brief quotation, and “Solving for Pattern” (1980), paraphrased: from standard knowledge of the texts. Kirat, the Bhai Lalo and Malik Bhago story as told in the janamsakhi tradition, and dasvandh: standard Sikh reference knowledge, presented as tradition. Other literary and historical references are to standard texts. Vendor representations. None; no product is credited with saving time, and the ten-minute and eight-minute figures are hypothetical arithmetic.
Recollection and tradition. The twenty minutes are hypothetical throughout, and no measured saving is claimed. The practice agreement is hypothetical. No employer, schedule, template decision or organization’s conduct is described or alleged, and the remarks about what a target would be expected to produce are conditional. The remark about an evening walk is a stated preference, not a description of the author’s routine. The Lalo and Malik Bhago story is presented as janamsakhi tradition, not as history, and no claim is made about what all Sikhs believe or practice.
Analysis and forecast. The readings of Chaplin, Jevons, Taylor, Campbell, Illich and Berry, the use of kirat and dasvandh as ways of thinking about measurement, the proposed measure of net supervision time, the fidelity reading, and the proposal to decide allocations before adoption are the author’s personal interpretation and policy preference as of November 6, 2026. Independence. This essay is not sponsored by any institution, company, health system or technology developer. It is personal reflection, not medical advice; readers seeking care should contact their own physician’s office.
About the Author #
Kanwar Partap Singh Gill, MD, traces his family roots to Khadur Sahib in Punjab’s Majha region and grew up in Amritsar. He attended Spring Dale School and completed pre-medical studies at Khalsa College in 1994–1996. He earned his MBBS from Sri Guru Ram Das Institute of Medical Sciences and Research in Amritsar in 2002.
He completed PGY-1 at the University of Arkansas for Medical Sciences in Little Rock in 2007–2008, and PGY-2 and PGY-3 with UCSF Fresno at Community Regional Medical Center from 2008 through 2010. He has practiced medicine in California since 2010.
His career has crossed four distinct information eras: predominantly paper-based medicine, early and fragmented electronic records, the integrated-EHR era, and the emerging age of generative and agentic artificial intelligence.