The Right to Win

Working with Claude has been the strangest management experience of my career. I have led people since 2005. For the first time, I am working with a teammate who is thorough by default, available at three in the morning, willing to push back when I am wrong, and who explains their reasoning every single time. The competence is unsettling. The implications are larger.

Because the question I keep arriving at, and one the next decade will force on every employer, is this: given the choice between an entry-level human and an agent like Claude, what would have to be true for me to choose the human?

The honest answer, today, is uncomfortable. The dimensions on which entry-level humans were taught to compete (diligence, thoroughness, knowledge-work execution, "getting it done right") are dimensions where I now lose to my AI teammate every time. Cheaper. Faster. More patient. More consistent. Available always.

This is not a story about AI taking jobs. That story has been told. This is a story about strategic position. About what it takes to win in a market that has fundamentally re-priced what humans are uniquely good for. And about how little of that conversation is happening at the level it deserves: at the level of every parent, every school, every employer, and every young professional walking into the workforce right now.

What Companies Have Always Known

In strategy, the "right to win" is a precise idea. Lafley and Martin made it the centre of Playing to Win: a company has a right to win in a market when it has a deliberately built unfair advantage: a unique capability, asset, or position that gives it a better-than-even chance of succeeding against everyone else trying to serve the same customer.

The discipline is brutal. You decide where to play. You decide how to win there. You invest relentlessly in the capabilities that compound your edge. And (the part most leaders flinch from) you say no, on purpose, to markets where you do not have a right to win and never will.

Companies that operate without this discipline die. Slowly, then quickly. They show up to fights they cannot win, get ground down by competitors with sharper position, and run out of the resources needed to redeploy.

We are now demanding the same brutal logic of every human entering the workforce. Almost no one has been told that this is the game.

Schools still optimise for the test. Parents still measure report cards. Universities still credential for compliance. Employers still hire entry-level for diligent execution. The entire pipeline that produces working humans is still optimising for the strategic position that AI just took. We are sending kids to fight in armour that no longer fits the war.

The Old Right to Win Just Vaporised

The strategic position a 22-year-old occupied in 1995 was real and defensible. Be more thorough than the 45-year-old. Be more available. Execute the standard playbook with fewer errors. Show up early, stay late, do the analyst grunt work with care. That bottom rung was a defended position. It produced incomes, careers, and (critically) the senior talent of the next twenty years.

That position is gone. Not eroding. Gone.

Stanford's Brynjolfsson published payroll data from millions of ADP records last August. 22 to 25-year-olds in AI-exposed occupations are down 13 percent in employment since late 2022. Entry-level software engineering and customer service: down roughly 20 percent. And here is the part nobody is talking about: wages are flat. The market is not telling young people "skill up, get cheaper, win on price." It is just closing the door. The bottom rung is being sawn off.

The Anthropic Economic Index, in its March 2026 report, is sharper still: enterprise API traffic is dominated by automation, not augmentation. The "AI as copilot" story we tell each other is mostly a consumer-tier phenomenon. The moment AI moves into a production system at scale, companies use it to replace tasks, not to coach humans through them.

So the human strategic position of "I am more diligent than the next person" is now occupied by a $20-a-month subscription that does not sleep, does not get moody, does not have a bad week, and gets sharper every quarter. You cannot win that race. Telling young people to try harder within it is malpractice.

The Gym Got Closed

Here is the part that should keep every CEO and every parent up at night. The traits that would give humans a new right to win (initiative, judgment under ambiguity, push-back, customer empathy, the ability to carry accountability for an outcome) were not taught in school. They were built on the bottom rung.

You started by doing standardised work. You made mistakes. You learned to push back on a senior who was wrong. You learned to read a room when a customer was about to walk. You learned to own a thing that broke and fix it without hiding. Senior judgment is downstream of junior reps. The diligent-execution rung was the gym where humans built the muscles that made them un-automatable later.

We are closing that gym at exactly the moment its outputs matter most. AI eats the bottom rung. The rung was the gym. The gym was the senior pipeline. The pipeline was the entire human edge of the next decade. The math is not mysterious; it is simply terrifying once you draw it out.

The cohort showing up at the door is already weakened by three structural shocks they did not choose. Phone-mediated attention collapsed in measurable, replicated ways: Jonathan Haidt's data shows fifty years of educational achievement progress that ended in 2012, the year teen smartphone-plus-social saturation hit. COVID deleted the in-person socialisation window for the cohort that should have been doing internships and first jobs. And helicopter parenting, well-meaning and now well-documented, degrades the exact two muscles the new right to win demands: resilience and push-back.

The temptation, especially for managers my age, is to call this a generational character problem. It is not. Every generation gets accused of this; the complaint is older than writing. What is genuinely different this time is structural, not moral. The system that produced working humans is broken in three specific ways at once, and no one has been given the new spec.

What the New Right to Win Looks Like

If diligent execution is the lost position, what is the defended one? The honest answer comes from spending real time inside an AI's capabilities, and noticing where the structural weaknesses sit. There are six.

Initiative.I respond brilliantly. I rarely originate. The human who walks in saying "I noticed our customers are quietly churning and I think we should..." is doing what I structurally do not. Originating from a blank page, naming a problem nobody else has named, starting a thing that did not exist; this is not a personality trait. It is a competitive position.

Judgment under ambiguity. I optimise within the frame I am given. The frame itself (the question of which trade-off matters when the stakes are real and the data is bad) remains human territory. The senior who looks at three reasonable options and decides, knowing the cost of being wrong, is doing work I cannot fake.

Trust and accountability.I can ship the work. I cannot carry the consequences. A human's name on the work creates skin in the game I structurally cannot have. Accountability requires being mortal, employed, and reputational. The colleague who signs off, takes the heat when it goes wrong, and stays in the job afterwards is doing something the AI cannot offer the system at any price.

Embodied and relational presence.The negotiation table. The customer's home. The warehouse floor. The supplier's micro-expression that tells you to stop pushing. The colleague you can read across a glass conference room. This work will not be remote. It will not be virtual. And it will not be mine.

Repair after conflict.Not ghosting. Not scorched earth. The harder middle path that builds long-arc trust. Marriages, partnerships, supplier relationships, and senior teams are all built in the repair, not in the avoidance. AI does not hurt anyone's feelings, which sounds like a feature and is actually a structural disqualification from the relationships that compound.

Taste and originating. I synthesise what exists with extraordinary fluency. I am bad at the leap that creates a category that did not exist. Taste (knowing what should exist that does not yet) is the unfair advantage that compounds across a career. The founder, the creative director, the head of product who can see what the deck cannot show you yet.

These six dimensions are the new defended position. Almost none of them are taught deliberately by school, parents, or employers. We are still optimising for the dimension we just lost.

Four Constituencies, Four Strategic Plays

A company does not build a right to win by hoping. It allocates capital, redesigns the org, says no to non-strategic markets, and builds capability deliberately. The same discipline now applies to four constituencies who each control part of the human pipeline.

1. Society and the Education System

K-12 still optimises for diligent rule-following and test performance. Universities mostly produce credentialed graduates with weaker push-back than the cohort five years ago. Bootcamps and the new alternative-credential industry mostly retrain people for the jobs AI is eating fastest. The whole system is investing capital in a strategic position that no longer exists.

The strategic play is not a tweak. It is a curriculum rebuild around initiative, judgment under ambiguity, deep attention, repair, and embodied work. Project-based learning where the project actually fails sometimes. Apprenticeships at scale. Real stakes, real customers, real consequences in the learning environment. Less screen, more room. This is the hardest reform of the four because the system has the slowest reflexes, and it is the one that matters most because it sets the floor everyone else builds on.

2. CEOs and Managers

Stop hiring entry-level for the dimensions AI now wins. You will lose every time and you will spend the salary while losing. Hire for judgment, initiative, and the human edges , and screen for them with assignments that actually surface them, not interview theatre.

And here is the harder play, the one nobody wants to fund: rebuild apprenticeship as a deliberate cost line, not a side-effect of needing cheap junior labour. The bottom rung you closed by deploying AI is the rung your senior talent grew on. If you do not build a new gym (structured judgment reps for juniors, where AI handles the diligent execution and humans handle the judgment calls under a senior reviewer) you will hit a senior-talent cliff in five to seven years that no recruiting budget will solve. You are paying for the apprenticeship explicitly now, or paying for the missing senior bench painfully later.

The CEOs who will be remembered well in this decade are not the ones who rode the productivity wave fastest. They are the ones who reinvested some of that productivity into building the next bench when nobody was forcing them to.

3. Parents

School is no longer the right to win. School delivers what AI now delivers cheaper. The traits that AI cannot fake have to be built somewhere, and largely that somewhere is home. This is uncomfortable because it transfers responsibility back to a constituency that has spent decades outsourcing it to institutions.

Four traits to train deliberately. Long-form attention: books, sustained focus, no algorithmic feeds during the years the brain is wiring itself for depth. Agency: let them fail at small things early, do not rescue, give them real responsibility for outcomes that matter to other people. Repair: teach the harder middle path after conflict, not avoidance, not blow-up. Embodied work: chores, sports, in-person hosting, eye contact with adults, the awkward conversations that build social muscle.

None of these are graded. None show up on a transcript. All of them are the right to win.

4. The Young Professional

You have agency in the now. The race you are being told to run (be more thorough, more knowledgeable, more diligent) is the race AI is going to win every single time. You need a different strategic position, and you need to start building it on day one.

The play has four moves. Become AI-fluent immediately, not as a hobby, as a discipline. The within-cohort divergence is already sharp: AI-fluent entry-level candidates are getting 12-percent salary premiums while general entry-level gets squeezed below cost. Take the uncomfortable assignments: the ones that require walking into difficult rooms, signing for messy outcomes, owning decisions a peer would not own. Cultivate taste deliberately by reading deeply, looking at great work in your field, developing opinions you can defend without flinching. And, the boring move that beats everyone, show up on time, every time, and finish what you start. Reliability is not the right to win, but unreliability disqualifies you from having one.

The World That Comes Out the Other Side

If this rebuild happens (and it will happen unevenly, with the rebuilders winning enormous compounding advantages) the workplace by 2030 looks meaningfully different.

Org charts compress. The five-person company doing the work of 500 (which I have written about separately) becomes the median, not the outlier. Wages bifurcate hard: AI-fluent humans with judgment command 30-percent and 50-percent premiums on yesterday's benchmark; the diligent-only get squeezed below the cost of running the equivalent agent. The squeezed middle is salaried middle-management knowledge work, the role that grew up alongside the post-war corporation, which compresses fastest because it added the least defensible value.

Entrepreneurship stops being the exception and becomes the default path. Gen Z already runs side hustles at roughly twice the Millennial rate at the same age; that signal is not laziness rebranded, it is the leading edge of how a cohort responds when the salaried bottom rung disappears. The gig economy flips from desperation to designed portfolio: multiple income streams as architecture, not fallback. Personal finance rewrites itself around longer income arcs (50-60 productive years, not 35-40), more capital formation early, more volatility absorbed because no single salary will smooth it for you anymore.

The new job categories that absorb redeployed humans are starting to be visible. Agent orchestrators: the "agent boss" role Microsoft is now naming explicitly. Trust signatories: humans whose accountable signature is what lets the AI's output enter the world (legal, medical, fiduciary, regulated). Embodied and relational specialists: care, teaching, trades, hospitality, therapy, the work that has to happen in a room. Originators: founders, R&D, creative direction. Repair workers: the long tail of fixing AI failures and handling the cases the system cannot. Civic and community builders: the work that keeps a society legible to itself.

These categories will not absorb the displaced one-for-one and on time. They never do. The transition will be brutal for the cohort caught between the old defended position and the new one. Whether the brutality lasts five years or twenty depends almost entirely on how seriously the four constituencies above pick up their plays now.

Where This Sits in Life After AI

The Five-Person Company essay asked: which jobs survive. The Identity Crisis essay asked: who are you when the task that defined you is automated. The Right to Win is the operating layer between those two: what humans, employers, parents, and societies must deliberately do to be on the right side of both questions when the dust settles.

It is the work of the next ten years. It is not happening yet at anywhere near the scale the moment requires. And it is the conversation I would most like every leadership team, school board, and dinner table to have this year, not next.

The companies that build a right to win do so on purpose, with a doctrine, with discipline, against pressure to do anything else. So will the humans who win the next decade. And so, if we choose it, will the system that produces them.