REFLECTION · Sylhet veranda · post-monsoon

The career advice that doesn't survive a year.

Most career advice has a shelf life of roughly eighteen months. A retired Sylheti headmaster spreads forty years of notebooks on the clothesline after the monsoon, reads what dried, what curled, what buckled. Six lines that didn't make it, and one rule that did.

Read what dried

The boys come to my veranda still, the way they used to come to the staff room in the 1980s, except now they are grown men with sons of their own and the same nervous half smile. They sit on the wicker chair, they ask about the prostate, they ask about my brother in Birmingham, and then around the second cup of cha they ask the thing they actually came to ask. Sir, ami ki right thing korchi? Am I doing the right thing? I make my face the right shape and I lie a little, and then in the evening, when they have gone, I take down a notebook from the shelf and write what I should have said.

The notebooks go back to 1982. There are 43 of them, marbled covers, the kind they used to sell on Bandar Bazar for nine taka and a smile. Three of them are mostly dry. Two are curling at the corners. One, the one from the long flood year, is buckled in half and the ink has bled into a brown lacework you can read sideways if the light is right. Yesterday I took them off the shelf and pegged them to the clothesline on the veranda. The monsoon had broken on Tuesday and the haor was finally giving the air back. I made a cup of cha. I sat in the wicker chair. And I read.

What I want to tell you, and what I should have told the boys forty years ago, is that most of what I wrote in those notebooks did not survive the next monsoon. The advice was right when I gave it. Then the world tilted half a degree and the advice was partly right. Then the world tilted again and the advice was wrong in a way that made me want to write the boy a letter, except by then the boy was a deputy commissioner in Khulna and the letter would have read like a complaint about his career.

Cal Newport wrote a whole book in 2016 telling people to be so good at one thing that nobody could ignore them. So Good They Can't Ignore You. I bought a copy in 2018 from the Etihad terminal in Abu Dhabi because the cover was clean and I had four hours. By 2022 the same Newport was writing about Slow Productivity, which is a different book with a different shape, and the boys I had told to specialise were now being told by their managers that the specialty had been folded into a tool the company bought from a Bengaluru vendor for ninety dollars a seat per month. They were neither so good nor so ignorable. They were, in the language of the haor, out in the middle of the channel with the tide falling under them.

The clothesline — six pieces of advice that didn't survive
1984failed by 1991
Pick the trade that pays well and stay in it for thirty years.
My nephew was a lithographer for the Bangla Academy press until offset killed letterpress in the late eighties. The trade did not last thirty years. It lasted seven, then it lasted nothing.
1992failed by 2008
Get a government job. The pension is the point.
The pension is still there, alhamdulillah. The buying power of the pension is not. A pension is a promise denominated in 1992 taka, paid out in 2030 inflation.
2001failed by 2014
Learn HTML. Every office will need one boy who knows the web.
Every office has fifty boys who know the web by 2014, and the office buys a Wix template for the price of a kilo of mutton. The skill ate itself the moment it became furniture.
2010failed by 2019
Pick the thing you would do for free.
Has only ever been good advice for people who already had a sabbatical, or a flat in Gulshan, or a wife with a CSO job. For the rest, the thing you do for free is called a hobby, and hobbies do not buy the inhaler.
2017failed by 2024
Specialise. The deep narrow well beats the shallow wide pool.
The deep well got automated, the boys who specialised in PDF extraction or copy-editing or front-end CSS got an apology email from a manager in Singapore, and the shallow wide pool people who could explain three things to four kinds of stranger kept their desks.
2024failed by 2026
Learn to prompt. The model is the new pen.
The pen is still the pen, but the model now writes its own prompts when you ask it to. The boy who built his identity around prompt engineering in late 2023 is, in May 2026, watching a meta-prompt agent do his job for the cost of a cha-and-shingara from the corner shop in Kalighat.

You see the shape now. Roughly eighteen months. That is the half life of the average sentence I wrote with a fountain pen in the staff room. The 2017 specialise note had a slightly longer run because the world moved more slowly that year. The 2024 prompt note had a shorter one because everything in our trade is on a faster monsoon now. But the median came out at eighteen months, give or take a wet season, when I sat with the calendar last Friday and counted properly.

This is not a complaint about the advisors. Newport meant what he wrote in 2016, and he was right for the engineers who got into Google in 2017 on the strength of one deep skill. Paul Graham's What You Can't Say makes a related point, that the most dangerous beliefs of any era are the ones nobody around you can yet see are wrong. The advice that aged worst was, almost always, the advice that everyone agreed with in the year it was given. Graham's piece was written in 2004 and reads now, in 2026, like he was sitting on this veranda with me last Tuesday.

The Bo Burnham line, which a former student sent me on WhatsApp in 2023 and which I have not been able to scrub from my head since, was something to the effect of, don't take advice from people like me who got famous before they turned twenty, because the dice are still in the air for them and they are mistaking luck for a method. That sentence is the closest any modern person has come to what I would write at the top of every one of these notebooks if I were starting again. The dice are still in the air. Be careful what you write down.

The trouble with writing in ink is that the ink lasts longer than the world the ink was describing. I have a great uncle who advised a boy in 1962 to go into jute. Jute. The boy listened. He died poor in 1991 in a flat in Mohammadpur with the curtains drawn. My great uncle, by then, was advising other boys to go into pharmaceuticals, which was the right advice for that decade, and would later be wrong by 2010 in the specific subspecialty of generic antibiotics, which is how Pran ended up owning half the market for agricultural chemicals in Sylhet division and an old pharmacist in Zindabazar ended up selling his shop to a cousin who turned it into a mobile recharge counter.

So the rule, if there is one, is not a rule. It is a small instruction I would write at the front of every notebook if Maryam were still here to laugh at me for it. Keep the question alive longer than the answer. The answer rots. Newport's specialise, Graham's heretic belief, Burnham's dice, my own 1986 accountant note: all of them are answers that worked until they didn't. The question they were all answering, what should this boy do with his life given what I can see today?, is the only thing in any of these notebooks that has held its shape across all 43 years.

I have started doing something the boys find irritating. When they ask me what they should do, I ask them what they think the question is. Not the answer. The question. Half of them have not noticed that they were asking a different question from the one their friends were asking, and a third question from the one their mothers were asking. Once you can say the question out loud, in your own voice, in your own sitting room, on a Friday afternoon, the answer is almost always already there, sitting on the table like a samosa nobody wanted to be the first to finish, already half eaten by the time anyone admits it.

The boys leave. The cha cup cools. I take the notebooks down from the line. The 1984 one is dry. The 2010 one is still curling. The 2024 one will need another full day in the sun, and the call to Maghrib is already rolling in from the mosque on the corner of Bandar Bazar Road, the one with the green dome that the floods of 2004 nearly took.

It will dry in time, the way the others did, the way the buckled one from the flood year eventually dried enough to read sideways, and I will peg it back on the shelf next to the rest of them, waiting for whatever a boy asks me next spring, by which time the mosque on Bandar Bazar Road will have called the Maghrib prayer a great many more times than I have bothered to count.