Some weeks ago I posted about completing six years at The Whole Truth on LinkedIn. My intention was to be thankful - because I genuinely believe that work anniversaries should celebrate the gracious tolerance of your employer and colleagues far more than they celebrate your own growth and loyalty. But as is often the case, I bungled my intention. The post became a bit ponderous (some might say didactic), about how no one really spends six years at the same company without the company answering a core need - a why.

Someone I don’t know read this post and replied saying that they found it very inspirational. I would like to formally deny this in writing. I did not reach here through any process worth copying. I have now worked for twelve long years - the first six not knowing what ‘working’ should entail, and the last six living a dream whose complex weave I am just beginning to notice.

And it makes me wonder, what IS the why of others? How much of it did they choose, and how much of it comes as a signal from society?

To be clear. I am talking about satisfaction at work and nothing wider. Happiness as a whole has far too many inputs - your health, your marriage, your joint-health, your best friend’s mark-sheet - and I cannot possibly target something so wide (maybe I can, but today we aren’t feeling that confidence).

Work is the narrower question. We understand that work can’t be fun and energising a hundred percent of the time, and that is okay. The bar is whether the thing means something to you on the days it is not fun.

II.

For my parents’ generation the question was settled long before anyone thought to ask it. A good job was a stable job. Government for preference, or a company large enough to behave like one. The whole architecture of a life was built around not losing it.

I grew up around transfer orders. It was a sheet of paper that arrived and rearranged life - the school, the friendships, the flat, the city. I learnt to be frightened of that paper very early.

The why of a government job which forces you to move home every five years is stability in an economy where not having one would set you adrift. Even in my generation, there is still a very large section of society which lives in this fear. Between 2014 and 2022, central government recruiting agencies received 22 crore applications - it landed a job for just 7 lakh candidates. Three in a thousand. Twenty-two crore people stood in a low-probability queue because a pension, a Pay Commission and a near-total inability to get fired was the only hedge on sale.

I have no smugness available here. Dad was buying insurance in a market that offered exactly one policy. But one generation later, his hard work meant I didn’t see this risk. I could choose a better why, and I kept failing.

III.

In the twelve years of my career, I got my why wrong three times.

The first was a number.

Percentage in board exams, rank in engineering entrance, percentile in CAT. Nobody around me framed these as a career decision, because they weren’t presented as one. I was decent at Math, so IIT it was, with the cherry of an IIM on top. I didn’t make it to either - just their poorer cousins. And all this time, I was just following protocol, not planning a career. So I ended up at whatever Plan D variant was available on the shelf for those who don’t make it to Plan A.

At no point did I ask what a person with these degrees actually does on an average Wednesday. Which behaviours the work rewards day after day. What kind of curiosity it demands and whether I had even a gram of it. I was optimising the entrance. Waking up at 5 AM to stare at unsolvable Physics problems written by Russians, and driving 12km on my Activa (from the shitty engineering college I managed to get into) to learn how to figure out the last digit of 3^3333.

In between, I learned how to design gears, build pumps, load a cantilever beam and something called the Finite Element Method. I used none of these in real life, outside the party trick of converting kW to bHP on the BMW website, while window shopping with friends.

The second was air conditioning.

Remember the transfer orders? What I wanted, with total clarity, was a peaceful job in one building, in one city, at a temperature I could count on. So when my MBA internship sent me to the villages of Western UP and Central Maharashtra (in peak summer, iykyk), I came back and specifically requested a B2B job at my first ever company - Reliance - and for several years I genuinely loved it. The softness of it. The absence of responsibility.

And then it started to hurt my ego, which is the least noble reason imaginable to want out of a job, and also the truth.

The third was a set of doors I kept closed for the longest time.

Rule-outs are the sneakiest kind of why, because a rule-out never announces itself as a decision. It is a why not masquerading as a why.

I ruled out consumer goods on the strength of one bad internship. One sample, one bad month, generalised to an entire industry, and a picture in my head of what the career would feel like. I also ruled out startups altogether, deliberately and for years, which is a story I already told and will not repeat here.

So let us add it up. I optimised for an exam score without ever asking what the score bought. I optimised for a fixed location out of an inherited fear. I wrote off consumer goods over one internship, and startups on principle, and the principle was thin.

I now work in consumer goods, at a startup, in a role that has changed shape three times in six years, and it is the most meaning I have ever found in work. Every significant decision I made was pointed directly away from the thing that turned out to work. Every day is either very funny or quite frightening, and on many days I barely manage both.

IV.

The convenient story goes like this. Our parents wanted stability. Millennials wanted a good package plus a culture deck (foosball table, unreadable page about values). And Gen Z, bless them, have added a third term to the equation - dignity. Respect. The right to not be spoken to like that. The right to leave at seven and punch in at eleven.

I like this story. I have told it at dinner tables, with conviction. I am now fairly sure it is mostly wrong.

Deloitte’s 2026 survey of 22,500 Gen Zs and millennials has the cohort prioritising ‘stability’ and ‘well-being’. Then you open the India cut and 96% of Indian Gen Zs say they want senior leadership roles, against the 6% who, a year earlier, told the global edition that reaching a leadership position was their primary career goal. The Indian answer, read plainly, is that we would like the corner office AND the boundaries, please.

The “conscious unbossing” story - Gen Z supposedly refusing management - fails the same way. Glassdoor went looking for it in actual promotion data and found Gen Z entering management at roughly the same rate previous cohorts did.

So what actually moves between generations?

The vocabulary.

Each cohort inherits a fresh, socially approved dialect for describing a good job, and reports it back to whoever is asking with total sincerity. My father said security. A 2014 MBA said learning curve. A 2026 graduate says psychological safety. Underneath, the same machine runs: look around, find the currently respectable answer, adopt it, feel settled for a while.

Every rung on that ladder was designed for the masses. Which is exactly why a real why makes a terrible LinkedIn post.

Every one of society’s standards is something you either hold true in your heart, or tell the world with pride. Great pay, a culture that makes its way into an HBR article, a free smoothie machine at work, and the company of kind, intelligent people.

None of them will ever tell you what your Wednesday is for.

V.

You cannot derive your answer from the weighted average of society.

Look at any successful person and you are looking at an output with the working hidden. You get the outcome, the interviews, the tidy causal story assembled in hindsight by someone with every incentive to make it sound deliberate. You do not get the eleven things that broke their way, the four years they hated, or the spouse who agreed to move. Pattern-matching on that is reverse-engineering a recipe from a photograph of the meal.

Which is inconvenient, because pattern-matching is what all of us do. It is why every third person I know has a plan involving rental yield and a second plan involving a podcast. “Passive income”, “the creator thing”, “starting up” - all perfectly reasonable answers and all, overwhelmingly, borrowed ones. Somebody looked at a set of outputs, spotted a shape, and adopted it wholesale without asking the prior question of what they are personally optimising for.

People skip the prior question because answering it honestly is expensive. Your real answer will very likely embarrass you. It will not photograph well. It might be that the work you find most meaningful is unglamorous work in a mid-sized company forty kilometres away from a metropolis. It might be that the title matters to you a LOT more than you are willing to admit at a dinner where everyone is performing detachment from titles. Both are fine. Both are un-postable on LinkedIn.

Come to think of it, whys are not even static. Mine keeps changing, and the one that inspired this unknown man - it might not even stay.

And once you do have it, most of the metrics other people use to score your working life don’t mean a thing, because they were never designed for a why. There is no triumph in it. They just stop registering, the way even great theme songs on shows do after a while.

VI.

So, what was my why?

Once a year I look back and check whether I got better - as a person, as a leader, as a marketer - and for six years running it has come back yes. Through good years and rotten ones, through a year my team was rebuilt from scratch, through quarters we won and quarters we lost badly.

It fits on nothing. It is not remotely a strategy. And I cannot recommend it to you, because I have no idea whether the same audit returns anything at all in your hands. It just happens to be the only thing that has ever told me what my Wednesdays are for.

But I want to flag the part that worries me, because an essay that ends on my own cleverness is not worth your inbox.

The answer is situational. It was true at year three, it is true at year six, and nothing whatsoever guarantees it at year nine.

The most dangerous version of an offbeat choice is the one you got RIGHT - right for a phase of life that has since ended without sending you a note, and which you keep running anyway because it worked so beautifully the first time. Needs change. Plenty of people are still optimising, with tremendous discipline, for a version of themselves that stopped existing around 2019.

If one July the honest answer is that I did not get better this year, and it was not an accident, then the whole thing comes down in an afternoon and I will go and find another one.

Year seven, then. Same audit.

Claude recently told me that I have used a semi-colon just twice in approximately 15,000 words of exposition.

I. The case for AI

The project started as vanity, which I will own upfront. I have spent two years watching AI produce prose that reads like a LinkedIn influencer clearing his throat, and I have been at a crossroads – I love writing, it’s something I do more than most people around me, but I can’t (or don’t want to) write everything I always have to write.

Upon ruminating on the general unfairness of many things in life, I decided that I will have to make peace with rewarding a machine for some turns of phrase. But in those vulnerable moments of life, the worst that could happen would be being called out as AI slop. So. What if there was a way to train Claude to write like me! Surely there is a way to train the AI to do this – there might even be a Substack about the best prompt to use for this.

So I logged in to my Claude account and handed over my body-of-work, as it stood a few months back:

Four strategy notes about the Protein Powder business I handle, a monthly read for our E-Commerce business (my previous job) that I had been writing for four long years, a long letter to a friend explaining why the pricing tiers on his newsletter were wrong, and a five-thousand-word essay about magic systems in fantasy novels. There was also some short fiction I put on a WordPress blog in 2015 and had mostly forgotten about.

(There is a portrait of a man, here. I am not certain it is a flattering one.)

Nonetheless, the next thing I know, I was measured by Claude.

One in five of my sentences in a strategy note runs past thirty-five words, which is about the length of this sentence and the one before it stapled together, and which no writing coach on this planet would sign off on. I have used two semicolons in my entire adult corpus. TWO. I shout in capitals, I make my jokes inside parentheses (like this one), and where a properly educated person types an em-dash, I type a hyphen with a space on either side of it because I spent most of my adult life thinking that the em-dash is a typo.

This was… strangely good. Until I read this, I couldn’t possibly believe that AI could ‘quantify’ my writing. That it could just identify an empirically repeating pattern of sentence lengths, and behavioural tics, even repeated grammatical mistakes, and file them up as a writing style. But here was Claude Fable 5 – released from the clutches of US Government and available in India only till the 7^th of July (at the time of this writing), doing just that.

I. The machine, confidently wrong

But as I actually got the AI to use this skill and write like me, I found the output insulting.

It told me, with total confidence, I never ‘hedge’, which is flattering and false. I hedge constantly. I just do it in the first person, where you can catch me at it: “I feel safe in pegging.” “This worries me a bit.” The machine had mistaken accountability for swagger, which, come to think of it, is a mistake a fair number of people have made about me.

So we went back and did it with data. And then began the long, petty war on slop, which satisfied me in ways I am not proud of.

If you have read anything written by a language model recently, you know the tells even if you cannot name them.

“Here’s the honest truth…”

“The point is not the protein, it is the permission.”

“The category quietly switched to growth.”

These SHOUT AI.

Claude wanted to “meter” these. Allow one per piece, as an earned flourish, exercising good taste. I said no, and I will defend that. You cannot ask a model to ration the exact thing it overproduces, because the rationing gets done by the same machine which thinks the phrase is delicious. That is the fox counting the hens. Strip them to zero. All of them, including the ones I use myself, because I would rather lose two good sentences of mine than leave the door open for two hundred bad ones of its.

II. The diff

The lesson which actually landed came later, and it came from a diff.

After ALL of the AI stripping, I asked the skill to write a product brief in my voice. It produced something quite good and slightly not-me, so I opened the file and started cutting. Where it had written “kill ‘light’,” I wrote “let us not use ‘light.”

Where it had asked a clean, tidy open question, I wrote three sentences of me arguing with myself and landing on a lean. Where it had capped a tactical instruction with a neat little epigram, I deleted the epigram, while leaving an almost identical one four paragraphs above, because that one closed a real argument and this one felt decorative.

Then I looked at what I had done and understood something uncomfortable. Almost everything I cut was ‘technically’ better writing than what I put back. This is going to sound complicated, so let me complicate it a little further by borrowing a sport metaphor.

Jannik Sinner is, at the time of writing, the best tennis player in the world, and he hits a ‘technically’ wrong forehand.

There is a moment just before a player swings - coaches call it the power point, where your coach will ask you to keep the racket head up and the strings squared to the incoming ball, everything cocked and tidy. Photograph Sinner at that exact moment and it looks like an error. His racket is lying flat with the head pointing the wrong way and the handle aimed at the side fence. From that wrong position he whips the whole frame around in the last fraction of a second and whips the ball violently. His forehand comes out at 125 km/h, far faster than the tour average, carrying about 300 revolutions per minute more spin than the average professional manages.

Any good coach would warn club players against copying it, because the same whip would wreck an ordinary arm. But the whip IS the point. The power lives inside the wrong position. Correct the flaw and you have corrected away the only reason anyone is watching.

My writing is full of stuff a decent editor would take out. I use “which” where a grammarian wants “that”. I say “a LOT” when I mean a lot. I write a forty-word sentence and then a four-word one, and then I wander off to describe something I found interesting, and I come back a paragraph late. Every instinct the machine had was to clean this up. Tighten. Compress. Fix the “which.” Make it presentable.

In the presentable, I found no joy.

So the skill file now carries a section which is, functionally, a list of my own inefficiencies, protected from correction. Do not remove his “just.” Do not fix his “which.” Do not prune the digression, the digression is what he is doing here. It is the strangest document I have ever been party to..

Seventy thousand words and twenty five notes of my writing later, I think what Claude produces will pass for my writing for exactly zero of the ten people who have read me before. And I am certain that outside of a boring work document, I will never really write anything with AI – at least not yet. There is power in what those ten people think, after all.

(P.S. – Yes, I can write a better AI prompt. No, I won’t write one)

But then there was this work thing which forced me to write a long answer. This month, it was my turn to write a letter to some of the brand’s super-fans about how I joined this org (we do this kind of thing time-to-time). The brand team limited me to 500 words. And if you know me, you know this is less than half of what I want to say about far, far more trivial things.

So here is the longer answer. It meanders but it’s still a fun story, which involves a police station and my critique of large, well-established orgs. I think I like this one better, because it is the one where I come across as an idiot who got lucky. Self-deprecation is the safest form of humour on the internet, after all.

I. A very large ocean, one very small drop

In 2019 I was at Reliance - the largest oil-to-consumer conglomerate in the country - selling synthetic rubber and plastic to other large manufacturers. Which is to say I spent my working life helping put a little more plastic into the world, about as far from food, or from virtue, as a man can get and still call it a job.

It was my first job, so for the first two or three years I did not know I was SUPPOSED to be unhappy. I had a fat paycheck (also, it made me fat). I had a designation that impressed relatives at weddings. And I had a role so magnificently non-essential that I could have disappeared for six months and the only department to register my absence would have been the canteen.

The thing about Reliance is that it is enormous and impersonal in almost exactly equal measure. Every decision of any real consequence gets taken inside a small, trusted circle at the top, and the rest of us - several lakh of us - exist to execute that decision with a straight face and a status update. I “ran” a P&L worth tens of thousands of crores in roughly the sense that a deckhand “runs” an oil tanker.

The tell, as with most things in my life, was my wife. She was at Unilever, learning the fundamentals of how a business is actually built - why a consumer picks one thing over another, what a brand is for, how a rupee travels. I was at Reliance, learning the fundamentals of… well… Reliance.

II. Every door, politely closed

So I decided to leave, which turned out to be the easy part, because deciding costs nothing.

Reliance was the best petrochemical company in the country, which is a wonderful thing to say but a terrible thing to have to escape, because it meant the only genuine exit was outside (Chevron, or Exxon Mobil or some such). So I looked abroad, and discovered that the firms overseas were not short of managers. What they wanted from India were engineers. I was, in the language of my own industry, the wrong export.

Fine. Side door. I signed up for an internal project which was supposed to transform Reliance’s aging ERP backbone, make them more consumer centric, and wrest the monopolistic supply control out of the hands of business heads and into the hands of a machine. No points for guessing how that went. Even I wasn’t too dumb to think this wouldn’t survive. The real reason I did this was - it got me up close and personal with a bunch of consultants to whom I generously offered a colour printout of my CV.

Not much came of it. Consultants will tell you that they switch out to Industry all the time. The reverse is rare. Accenture Strategy would not match my pay. Deloitte were too nervous to poach a Reliance manager and thereby lose Reliance as a client. Which left me, for one deeply strange stretch, simultaneously too expensive to hire and too dangerous to touch. Flattered and unemployable in the same breath.

Which left startups. But I had ruled out startups.

I had watched friends bolt from RIL into the first great Indian startup wave - Oyo, Zoomcar, that whole 2014-15 cohort - and I had watched a fair number of them land badly. So, being the sort of person who responds to fear by building a framework, I wrote myself a rulebook. No startup that bought its customers with deep discounts, because a business that rents its consumers does not own anything. No startup built on tech, because I knew no tech and had no intention of pretending otherwise in an interview. No startup whose founder I knew only from a flattering profile in a newspaper.

As of 2017, there existed no startup on the face of the earth that I was permitted to join. I had built a wall instead of a filter, and then I had stood behind it, congratulating myself on my judgement.

III. A rulebook that failed.

Then, in 2019, my wife again:

“There’s a guy at work who’s been writing a blog about his weight loss,” she said. “You seem to be doing most things wrong, so why don’t you read it?” Marital feedback, delivered with love.

That is how I found Fitshit - Shashank’s blog, and the wellspring of what this company would later go on to say out loud. I read nearly every post on it. And somewhere in the middle of that reading I watched, in real time, a man’s weight-loss diary turn into a business. The giveaway was a review of a protein bar so obsessively, unreasonably detailed that it was obvious this was no longer a hobby. Nobody writes 2,000 words about a bar unless something has gone deeply wrong, or deeply right.

Sure enough, the blog became a company. And the company was hiring, for a role that Shashank had chosen to call “Entrepreneur in Residence.”

Now let’s go back to my rulebook for a live example of serendipity.

No startup that buys customers with discounts - and here was a brand selling a 52-gram protein bar for 100 rupees and refusing to apologise for the price. No startup built on tech - and here was food, the most stubbornly analog thing there is. No founder I knew only from a newspaper profile - and here was a health-obsessed, frighteningly intelligent senior from my own wife’s company, a man whose B-plan I had just spent several weeks reading in long form (at least the parts that mattered to me).

My grand rulebook - this edifice of prudence, my carefully reasoned wall, turned out to have been a door all along, and it had The Whole Truth’s name on it. Except it did not, yet, because I am getting ahead of myself.

So I sat down and wrote the most well-considered 500 words of my life, and I sent them in as an application to Shashank’s first-ever LinkedIn hiring post.

IV. The luck, and its cost

What happened next was an absurd run of good fortune, followed by the universe arriving to collect.

We start with the fortune, because I had no business receiving any of it. Zero FMCG experience. A cold application in response to a LinkedIn post made by a stranger in petrochemicals. I had even received the auto-reply - thanks, will get back soon - which is ordinarily where applications go to die. My realistic chances were very, very low.

Shashank wrote back the same night.

Then I met him, and fell for the entire thing at once - the vision, the product, the pack, and the name the pack carried: And Nothing Else. Then I gave two interviews so comprehensively bad that no sane person should have hired me, and received an offer anyway. Then he matched my Reliance pay. Then he humoured a genuinely stupid request - that I be allowed to start on the 1st of April, so that my tax year would not get muddy - which, in hindsight, is a wonderfully accurate preview of the kind of person I was about to become at this company. And then, a full month before I had technically joined, he invited me to the company’s first-ever team outbound, where I met the early lot: Shashank, Rachna, Sanchi, Sahil and Apar, who had himself joined all of a month before me. A room full of absolute gems, and me, still technically employed selling plastic.

Nobody receives that much good luck without a bill arriving. Mine arrived in two parts, both of them within a fortnight.

Part one was COVID-19. On the day of that outbound, it had been a topic of idle conversation, the sort of thing you discuss between sessions. Roughly six days before I was due to join, it became the first nationwide lockdown. Production and Operations halted. And there was no supply to run the e-commerce sales channel that I was supposed to be running.

Part two arrived on the 24th of March, a Tuesday, disguised - as these things always are - as a “quick catch-up.” On that call I was informed that we were changing the brand name. The name I had fallen in love with, the name I had more or less said yes to, was going. And the replacement, I understood, would be “a process of discovery,” which is a beautiful phrase that means we do not have one. This was the beginning of a series of meetings which would result in the birth of “The Whole Truth”.

I hung up. I walked over to my wife. And I asked her, very casually, how things were going at work, because it was beginning to look like her job might come in rather handy over the next few months.

V. Never a dull day

To be fair to Shashank, he had warned me. I was not, he said, his e-commerce guy. I was his Entrepreneur in Residence - which is Shashank-speak for I might ask you one thing, or I might ask you a hundred things, and you will say yes to all of them.

So I did.

My first unofficial day at The Whole Truth was spent at the Sakinaka police station, asking the state of Maharashtra for permission to restart our shipments in the middle of a national lockdown. There is no line in any MBA syllabus that prepares you for this. There is no framework. There is only you, a folder of papers, a policeman with better things to do, general awkwardness and the regret of not being fluent in Marathi.

I have not really stopped since. I built this company’s first ever P&L, while looking after finance. I shipped bars to Kriti Sanon and Rashmika Mandanna, while looking after supply chain (united the two of them a good few years before Cocktail 2 ).

I did eventually get to work e-commerce for four long years, which happened to be precisely the four years in which quick-commerce arrived and rearranged the furniture in every CPG startup’s living room. And now, in the most predictable career move in the history of consumer goods, I have crossed the floor from sales to marketing, and taken charge of a category.

When I look back, I find it nearly impossible to reconcile that my ‘wall’ was actually a ‘door’. Did I accidentally create the best rulebook for a Mechanical Engineer with six years of non-starter work-ex to find their way into a startup? Or was I just swept in by the current of TWT? Do you have a door that was once a wall?

Never a dull day at TWT, folks.

Not since the day I joined. :)