MBA Class of 26D · New York City
Twenty-five of us spent five days in Manhattan hearing 20-plus founders, operators and investors, across finance, consulting, consumer, tech, health, hospitality and impact, on how they actually built their careers. This is the summary: what they said, where they disagreed, and what is worth taking home.
Before you read
Every speaker here was invited because they are both successful and interesting to listen to, selected twice over. So when a lot of them say the same thing, read it as a prerequisite, not a recipe: the things you would probably need if you want this kind of career, not a formula that produces it. Plenty of people do all of it and it still does not work out, and we never hear from them. The most useful signal is in the examples, the contrasts and the disagreements, where thoughtful, accomplished people looked at the same question and answered it differently.
Day by day
Five days, more than twenty speakers. Tap any name for the full summary.
The people
Each card opens a full summary: who they are and how they got there (context), then their actual takeaways and advice. Filter by day or sector.
The big themes
Every talk was read from its transcript and coded on nineteen themes at three levels: a main thread, touched on, or not raised. The themes are grouped into five families; within each, they are ordered by how many speakers made them a main thread. Tap a theme, or a speaker, to read it, and tap again to clear.
Incidence, not mention-counts: a single reader's judgement of which speakers emphasised which theme, coded from the transcripts. N = 17; the two corporate sessions and the alumni panel are each one session. Descriptive only.
The trek in numbers
What actually drove each speaker, the shape of each career, how they split on AI, and whether they planned their path or were pulled along it.
What this shows. How strongly each speaker optimised their own choices for money, impact or interest, and how those combined. About their careers, not their talk topics, a speaker can talk about money all day without living for it.
What this means. Interest, not money or impact, was the primary driver for eleven of the sixteen readable speakers, and a strong one for fourteen. Only Rachel Ravens was strong on all three at once; Anthropic and Linda Rottenberg were strong on impact and interest together; Frank Brown was the one money-first case, and even that was really about status and access. So "follow what interests you" was less a piece of advice than a description of how this particular room already lived, which is exactly why it should be read with the survivorship caveat in mind.
Levels: strong / moderate / weak, a single reader's read. Jinal Shah shows no clear read, her talk was a leadership narrative with no clear money / impact / interest optimisation.
What this shows. Each career as a ribbon of roles, one segment per job, coloured by industry. A colour change is an industry jump; a gold dot above a join marks a change of country; and ◆ marks a company of their own. Ordered by how many industries each crossed. Nearly none of these was a straight line.
A colour change is a new industry; a gold dot marks a change of country; ◆ = founded it; the green INSEAD tag marks where each alum did the MBA. Right column shows how many distinct industries each crossed.
A few that stand out
Homelessness as a teenager, then Club Med kitchens and a Michelin saucier station, then Head of Ops at a YC startup, then founding Thesis. A very low start to three industries.
Five careers in one life: consulting, then running philanthropy for UBS's billionaires, then author, then painter, now psychotherapist.
Broke into private capital without ever working inside a finance firm, from journalism and the UN to leading LAVCA and now running GPCA.
The counter-example: one industry her whole career, drinks and consumer goods, but across Moscow, Bratislava, Poland, Hong Kong, Copenhagen and New York. Mobility, not reinvention.
Industry and function are read from the meeting notes; country changes are marked only where the notes state a move (geography is under-recorded elsewhere). Craft roles like kitchens and bartending are bucketed as operations, so a cook-to-cook step is not a function change. Education gaps omitted. The two corporate sessions are shown by both speakers (Krifcher and Caine for Anthropic; Waldron and Wayne for Acumen); the alumni panel is omitted.
What this shows, and what it doesn't. This is one lens for reconciling some of the AI talk, mainly Anthropic's and Jean Boivin's, not the week's verdict. The idea: capability (the frontier) races while diffusion (adoption inside real organisations) lags, and your vantage shapes which one you weight, though an allocator has to track both: how fast the frontier moves and how fast adoption catches up.
Builder
“Don't bet against the exponential. Build for where the models are going, not today's.”Alon Krifcher · Anthropic
Allocator
“The largest capital buildout in human history.”Jean Boivin · BlackRock
“AI's capacity is being overrated; they'll wind up hiring a lot of people back.”Cate Ambrose · GPCA
Operator
“All you've got is leadership. AI is going to be doing most of it.”Jessica Spence · Edgewell
Where the room actually sat
The rest did not take an AI position at all, they were talking about other things. Which is itself the point: most people do not forecast AI, they just live the adoption curve.
Read it with the disagreement, not instead of it. This lens does not settle the argument. Allocators watch the frontier clock too, they just price the lag, and Cate Ambrose made a different claim entirely: that the capability itself is overhyped and firms will rehire the analysts they cut. So treat "two clocks" as Anthropic's and Boivin's framing, and see Where they disagreed for the genuine three-way split.
A conceptual diagram, not a plot of speakers.
What this shows. A simple, observed split. Some speakers planned each move deliberately; others were pulled along by curiosity or circumstance and seized what came. Same ambition, opposite operating style. This is just where each one sat, not a scoring system.
What this means. There was no single "right" way to navigate. The deliberate planners (Zoia Kozakov, Jean Boivin, Frank Brown) and the opportunists (Dan Freed, Patti Simpson, Cate Ambrose) were about equally successful, which quietly undercuts any advice to "be more strategic" or "just go for it." Both worked. What mattered was self-awareness about which one you are.
The practical residue
The concrete, repeatable advice, separated into what most speakers agreed on, specific tools you can use tomorrow, and the advice that was actively disputed.
The interesting part
The speakers did not sing from one hymn sheet. On six questions, accomplished people gave directly opposing advice. These are unresolved on purpose, the disagreement is the point.
If you remember five things
The lessons the cohort kept coming back to in the reflection sessions, and what Prof Carmon drew together at the close.
01
The loudest, most-repeated takeaway of the week, anchored to Jessica Spence's rule: take the job you are only 60% ready for. One classmate turned down a promotion she felt unready for, and three months later her manager pointed out she was already doing the work. The readiness you are waiting for mostly arrives after you jump, not before.
02
Prof Carmon's closing line, and the thread through every reflection: there is no single formula for a good career, so build more than one professional identity and define success for yourself. In Herminia Ibarra's phrase, you do not think your way into a new identity, you try your way in.
03
Almost every path looked planned told backwards, and several speakers admitted it was not. Told "you always knew what you wanted," the CHANEL speaker replied that it was not true, she was just good at storytelling. The tidy narrative is built afterwards, worth remembering both when you measure yourself against someone else's and when you tell your own in an interview.
04
The value Prof Carmon pushed hardest: the INSEAD network is the thing you actually got from the year, but only if you are deliberate about it. Frank Brown's habit of tracking every contact and reaching out on a cadence became a running reference. Most of the room found LinkedIn self-promotion uncomfortable; the answer was a system and a frequency of your own, not louder posting.
05
The one cross-industry commitment named again and again on the final day, one classmate reckoning they were "only touching about 30%" of what the tools can do. The speakers framed why it matters: as AI absorbs the technical work, judgment, taste and leadership are what is left. Several of us walked out having promised ourselves the same thing, to spend the summer actually building with it.
A summary of the week, put together for the trek group. Speaker takeaways are distilled from session notes and are our own read, not verbatim records. The theme grid and the charts code all seventeen sessions by a single reader; treat the numbers as descriptive, not exact. Figures attributed to a speaker (for example the "$24 trillion in dry powder" line) are that speaker's own claim, not ours.