More reps will not take you to the next level. Being watched, closely and honestly, will.
Month 09 gave you a rule for AI use so it doesn't quietly outsource your thinking. What was the core distinction it asked you to draw?
Newport's case for deep work wasn't "focus harder." What was the actual claim?
You have done, by now, several hundred skull base and spine cases. Your hands are fast. Your indication judgment is sound. And somewhere in the last year or two you have felt it: the plateau. Not incompetence — the opposite. Competence that has quietly stopped climbing. You can still get better, technically, at the EEA corridor, at a difficult clival dissection, at closing a CSF leak cleanly — but the improvement curve has flattened, because the thing that used to drive it (volume, novelty, fear) has run its course.
This is not a surgery-specific problem. It happens to concert violinists, chess grandmasters, Olympic athletes — anyone who crosses from "learning the basics" into "expert plateau." And the uncomfortable finding, replicated across all of those domains, is the same one Atul Gawande put in front of surgeons in 2011: past a certain level, experience alone stops producing improvement. What produces it is deliberate practice under an outside eye — a coach who watches you do the actual thing, in the actual setting, and tells you what they see that you cannot see about yourself.
You are also, right now, building something that will outlast any individual case: a skull base program in Porto Alegre, a fellowship pipeline, a research group, a family of trainees. The plateau problem stops being personal the moment you're responsible for other people's growth curves too. If you never build a feedback structure for yourself, you will not know how to build one for them — and "I figured it out myself" is not a transferable curriculum.
There is also a quieter home version of this same problem. Samilly sees things about your presence, your stress carry-over, your patience under fatigue, that you do not see from the inside. The question this month asks is not only about the OR.
Gawande's essay tells one very specific story: a decorated, high-volume surgeon — Gawande himself — noticed his complication rates had stopped improving. So he did something almost nobody in surgery does: he hired a coach. Not a mentor for career advice, not a conference talk, not a journal club — a person who scrubbed in, stood behind him, watched his hands, and afterward told him things like: your elbow drifts up when you're tense, and it costs you exposure. Small, mechanical, invisible-from-the-inside things. The kind of correction a violin teacher gives, or a tennis coach gives, routinely, at every level of the sport — and that medicine, for reasons of ego and hierarchy and the myth of the finished expert, mostly refuses to build in.
The essay's real subject is not coaching technique. It's the psychological cost of being coached once you're already senior. Being watched closely enough to be corrected requires tolerating a kind of exposure that feels, at first, like a demotion — I am the attending, the person who is supposed to already know. Gawande's answer: separate the identity of "expert" from the identity of "person who has stopped needing correction." They were never the same thing. Even elite performers in every other high-stakes domain keep a coach for exactly this reason — coaching is not remedial, it is what continued excellence looks like at the top.
Note the trap in the reveal above: "outcomes are fine" is exactly the reasoning that keeps expert plateaus in place. Good outcomes measure whether you've cleared the bar of safety — they say nothing about whether you're still climbing past it.
There is a second, quieter idea in Gawande's essay that matters more for you specifically than for most readers: a coach watches action, not narrative. You are extremely good at narrating your own performance afterward — precise, self-critical, articulate. That narration is not the same signal as what a coach sees live, because you cannot narrate your own blind spot; by definition, you don't know it's there to describe it. Gawande's coach didn't ask him to reflect harder. He watched the hands and reported what the hands actually did. The lesson generalizes past the OR: your own account of a decision (a triage call, a marketing move, a family moment) is not a substitute for someone else watching the actual event.
Coaching corrects what you do. The second text this month — the ESRT literature (Enhanced Stress Resilience Training, developed at UCSF specifically for surgeons, adapting MBSR/mindfulness-based stress reduction to the OR) — addresses the state you're in while you do it. This is qualitative, evolving evidence, not a settled body of proof: reported associations with reduced burnout and stress markers, improved executive function under pressure, and better team communication in high-stakes moments. Treat it as a promising practice framework worth testing on yourself, not a guaranteed result.
The core insight, independent of any specific study, is simple and testable in your own hands: technical skill degrades under an unregulated nervous system before it degrades under fatigue or difficulty. A surgeon whose jaw is clenched, whose breath is shallow and held, whose attention has narrowed to tunnel vision during a bleed — that surgeon's actual dissection quality is already worse than it would be if the same hands were calm. Mindfulness training, in this frame, is not a wellness add-on. It's a way of keeping the instrument (you) tuned during the case, the same way you'd keep a scope calibrated.
Three pillars carry the whole practice: interoception (awareness of your own internal state as it happens), emotional regulation (observing a stress spike without being run by it), and metacognition (catching your own thinking pattern — "I'm rushing" — and being able to shift it mid-case).
One more distinction worth carrying, because it changes when the tool is useful: interoception has a threshold problem. Most surgeons only notice the nervous-system signal once it's already loud — jaw locked, hands shaking slightly, voice sharper with the team. By then the technical cost has already been paid for several minutes. The actual skill ESRT is training is catching the early, quiet version of the signal — a faint tightening, a breath that got 10% shallower — while it's still cheap to correct. This is also, not coincidentally, the same early-detection instinct you already use in the OR for a slow bleed versus a torrential one: the earlier you name it, the fewer options you lose.
The broader surgical coaching literature (a newer field than clinical medicine itself, still building its evidence base — treat specific numbers with caution and prefer the general principles) converges on a handful of things that separate coaching, which builds you, from criticism, which only evaluates you:
One more principle from the literature, easy to miss because it isn't about the coach at all: the coached person has to actively solicit, not just tolerate, the feedback. Passive openness ("sure, tell me if you see something") produces vague, hedged input, because the observer is guessing how much honesty you can absorb. Naming the specific thing you want watched — "tell me about my instrument-exchange timing at the sellar floor, nothing else" — produces sharper, more usable coaching, and it also does something for the coach: it converts an awkward, hierarchy-violating ask into a clean professional request. This applies outside the OR too — a vague "any feedback on the PhD chapter?" gets a vague answer; "does the discussion section overclaim from an N of 40?" gets a real one.
You ask a fellow, specifically, to watch only your instrument-exchange timing at the sellar floor. Afterward: "You crossed hands early twice — try leading with the left first."
You have spent years building things that hold up a mirror to you: a diagnostic framework for your own habits, an advisor prompt library, a "grill" script you wrote yourself to interrogate your own patterns before anyone else could. You built /grill-me. You self-generated the Pact diagnosis and it turned out to be correct. This month's texts are going to sound, at first, like more of the same — another framework for seeing yourself more clearly. Read them anyway, because they're not asking for another mirror. They're asking for a coach, and those are not the same tool.
Here is the distinction, stated plainly, because it is the one you are most likely to skip past: insight is in surplus for you. Execution data is scarce. An AI mirror — a grill, an advisor, a self-diagnostic — will always be available, always articulate, always willing to go another round with you at 11pm. It produces the feeling of progress cheaply: you understand something you didn't understand an hour ago. But understanding your elbow drift and having someone watch your actual elbow, in the actual case, and tell you afterward whether it's still drifting — those are different categories of information. The first is insight as a consumable: you can gather as much of it as you want, in private, without ever changing what your hands do at the sellar floor at 2pm on a Tuesday. The second is what a coach produces, and it requires something an AI mirror structurally cannot: a witness who was actually in the room, who has skin in the outcome, and whose feedback lands in your very next case rather than in a note you filed and moved past.
The tell you already named yourself is the one to watch for here: mirror-sessions outnumbering executed days. If this month produces a beautifully self-aware reflection on your plateau and zero named coach, zero scheduled review, you have executed the exact pattern this course exists to interrupt — a formation note that names no changed behavior is collection, not formation. The insight was never the hard part for you. Naming an actual, specific person, and asking them, this month, is.
There is a second, quieter reason the OR feels effortless and the monthly review feels postponable, and it is worth naming so you stop mistaking it for a character flaw. Your activation follows urgency, novelty, stakes, and audience — not importance. Surgery hands you all four of those, automatically, every single case: a real clock, real anatomy you haven't seen inside this particular patient before, a real complication risk, and a real team watching what you do. You don't have to manufacture motivation in the OR because the OR manufactures it for you. That is precisely why you are already excellent there, and precisely why it is the wrong place to look for evidence about your growth capacity in general — the OR was never testing whether you can sustain deliberate practice. It was testing whether you can perform under conditions engineered to activate you.
The plateau lives exactly where those four activators run out. A structured case review, alone, at your desk, with no clock, nothing novel, no real stakes attached, and no one watching — is designed by its very nature to lose the fight against whatever has real urgency that hour. This is not a discipline problem you can will your way past with more resolve; it is a wiring problem, and the fix is not "try harder," it's rebuild the missing activators on purpose. A witness — a resident sitting in on the review, Samilly asking about it at dinner, a filmed five-minute clip sent to one named colleague — reintroduces audience and stakes artificially, the same way a deadline reintroduces urgency for a manuscript that would otherwise sit FINAL forever. Without a witness, the monthly review is not lower-priority than an urgent case; it is simply invisible to the machinery that decides what gets done today. That is the entire justification for making the coaching relationship in this month's practice a named person and not a private ritual — a private ritual is a review with the one ingredient it structurally needs removed.
Before the practice, slow down on the texts themselves. These are the five moments this month turns on — two retold faithfully from Gawande (his essay is in copyright; read the original, it is short), one from the ESRT work, and two verbatim from Osler, whose essays are public domain and belong in your hands whole.
The hinge of "Personal Best" is a single afternoon. Gawande, eight years into practice, complication rates plateaued, invites his old residency professor Robert Osteen — retired, nothing to gain, nothing to prove — to stand in his OR and just watch a routine thyroidectomy. Gawande thinks the case goes cleanly; he can barely imagine what there will be to say. Then Osteen fills a page with observations. The operating light had drifted off the wound and Gawande never noticed, working partly in shadow. His right elbow kept rising toward shoulder height — a mechanical tell that his control of the instrument was degrading — and the cause wasn't skill, it was setup: the draping had hemmed him in from the start. Magnified in his loupes, he'd lost awareness of everything outside the wound, including when the anesthesiologist was worried. Small, structural, invisible-from-the-inside things. Gawande's own verdict on that one debrief: it gave him more to work on than he had found on his own in the previous five years.
Em termos simples: one honest observer at a routine case — not a hard one — will find more for you to fix in twenty minutes than another year of solo volume at the EEA corridor will.
Atul Gawande, "Personal Best," The New Yorker, October 3, 2011 — the thyroidectomy Osteen observed, roughly midway through the essay.
Reporting the essay, Gawande asks why violinists stop having teachers once they're professionals — and Itzhak Perlman tells him the assumption is wrong. Perlman had a coach his entire career: his wife, Toby, herself Juilliard-trained, who sat in the hall at his performances and told him afterward exactly what she heard — where the tone thinned, where the phrasing went mechanical. The performer, Perlman explains, cannot hear himself the way the hall hears him; he needed, in his own phrase to Gawande, someone to be his "outside ears." The greatest living violinist did not outgrow coaching. He institutionalized it — quietly, permanently, inside his own marriage.
Em termos simples: the coach does not need a title or a department — a trained, trusted set of outside eyes who watches your real cases and reports honestly is the entire mechanism, and Samilly already holds a version of this role at home.
Gawande, "Personal Best" (2011) — the Perlman interview; the two-word fragment is his.
Carter Lebares, a UCSF surgeon, began by measuring what everyone privately knew: extraordinary rates of burnout and stress among surgical residents. Her response was not a wellness poster — it was to take MBSR, the most-studied mindfulness protocol, and rebuild it around the realities of surgical training, calling it Enhanced Stress Resilience Training. The pilot work reported signals — feasibility, less perceived stress, better executive function under pressure — in small cohorts. Treat the numbers as early. But the reframe is the passage that matters: stress regulation is presented as a trainable technical skill of surgery, sitting alongside suturing and dissection, not as a personal virtue some surgeons happen to have.
Em termos simples: your nervous system during a clival bleed is equipment, like the scope — and equipment can be calibrated with deliberate practice, not just endured.
Lebares et al., UCSF — ESRT pilot literature (see RESOURCES.md); pilot-scale evidence, framing over figures.
Month 1 gave you Osler's definition of imperturbability; this is the harder second half of the same essay — the dose-finding problem. Too little insulation and every emergency floods you; too much and the "human heart" hardens into the deformation *The House of God* mapped. Osler treats the right measure as something cultivated — titrated by deliberate practice, the way ESRT claims a century later — not a temperament you are issued at birth.
Em termos simples: a calma operatória é uma dose a calibrar — isolamento suficiente para agir sob sangramento, sem endurecer o coração que atende a família depois.
William Osler, "Aequanimitas" — valedictory address, University of Pennsylvania, 1889. Public domain. (Definition of imperturbability quoted in full in Month 1.)
Osler's most famous rhetorical trap: a full paragraph of buildup before the one-syllable answer. But read it inside this month's argument and it sharpens — the master-word is not volume of work (you have never lacked that) but the steadying kind: the same case reviewed deliberately, the same movement refined, the plateau attacked with structure rather than hours. "The brilliant student steady" is the clause aimed at you.
Em termos simples: mais horas você já tem — o que quebra o platô é o trabalho estruturado e repetido sobre o mesmo gesto, com um olho externo, até a brilhantina virar consistência.
William Osler, "The Master-Word in Medicine" — address, University of Toronto, 1903. Public domain. (Day-tight compartments quoted in full in Months 1 and 12.)
The written practice for this month is concrete and repeatable: choose one case per month — not your hardest, not your easiest, just a real one — and give it a structured review, either from a video recording (if consent and institutional policy allow) or from careful mental reconstruction immediately after. This is not a morbidity-and-mortality review (outcome-focused, institutional) and it is not the Month 01 post-case Stoic review (character-focused, daily, quick). It's slower, technical, and aimed specifically at the plateau: what would a coach, watching only this case, tell me to change?
learning-records/
1. What I planned. The approach, the anatomy expected, the contingency plan, the team briefing — write it as if explaining to a fellow before the case.
2. What actually happened. Where the case matched the plan, and where it diverged — anatomy that wasn't as imaged, an unexpected adhesion, a bleed, a change of plan mid-case.
3. Decision points. List each moment where you had to choose between two real options (push through vs. stage the case, coagulate vs. dissect free, extend the corridor vs. accept subtotal resection). For each: what did you choose, and why, in the moment?
4. One technical detail to refine. Not five. One. The elbow-drift-sized thing — instrument angle, hand tension, timing of a maneuver, a habit in exposure — that a coach watching only this case would flag first.
5. What a coach would have seen. Write this in the third person, as if you were standing behind yourself. This is the hardest heading and the actual point of the exercise — it forces the outside eye you don't yet have in the room.
"Done" for this practice is not "I thought about the case." Done is one dated file in learning-records/, per month, with all five headings filled — even if heading 5 is short and uncertain the first few times. The skill of writing heading 5 honestly is itself the plateau-breaker; it will feel awkward before it feels natural, which is exactly the sign it's working.
The surgeon's definition of done, applied here: this month is not done at "I read the essay and reflected on my plateau." It is done when the file exists in learning-records/ and one behavior demonstrably changed — a case chosen and reviewed, a coach named, a message actually sent. A note that names no changed behavior is collection, not formation, no matter how well-written heading 5 is.
The monthly review trains the internal eye. It is not a replacement for an actual external one — and this month's real behavior change, alongside the written practice, is identifying at least one person who could plausibly watch you the way Gawande's coach watched him: a trusted colleague at HSJ, a mentor from the outreach network (Dr. Raza and similar), a fellow who has scrubbed enough cases with you to have a real opinion. It does not need to be formal. It needs to be someone specific, named, and asked — this year, not eventually.
The ask itself is smaller than it feels. It does not require a conversation about vulnerability or plateaus or growth — it requires one specific, professional sentence, close in spirit to the sharper solicitation from the coaching-literature section above: "Would you scrub in on one case a month and tell me one specific thing you notice — nothing formal, just an honest eye". That sentence is a request any competent colleague recognizes and most will be glad to say yes to, because being asked to coach is itself a form of respect, not an imposition. The size of the ask is not the reason it keeps not happening; the size of what it exposes is. Naming that honestly — out loud, to yourself, before you write anyone's name down — is worth more than another week of refining the case-review template.
A senior colleague, unprompted, tells you during a case debrief: "That took longer than it should have."
A fellow who scrubbed with you says: "Your left hand tensed up right before the leak repair — want to look at that clip together?"
This month's texts plug directly into the OR mindfulness toolkit you already carry (Month 01 and the syllabus's practice systems), and sharpen a few pieces of it specifically for growth rather than only regulation:
The next level of my craft does not live inside my own hands — it lives in whoever I let watch them closely enough to tell me the truth.
Bring it back to me. Pick one real case from the last two weeks — not your worst, just a real one. Walk me through the five headings out loud, especially #5, what a coach would have seen. I'll push on whatever feels vague or too easy on yourself.
Then name one actual person — by name — who could become your Month-10 coach this year. What is stopping you from asking them this month?
Primary text: Atul Gawande, "Personal Best" (The New Yorker, 2011). ESRT: Lebares et al., UCSF — qualitative framing only, no fabricated statistics. Full shelf in RESOURCES.md.