How a Professional Can Stay Motivated to Keep Learning

Alexander recalls that as a child he meant to become a scientist: he chose MIPT over Moscow State University because you could get in through olympiads in March and May and sit nothing in the summer. At his lyceum he saw a wall covered in physics medals, and studied at the correspondence physics school. On the applied mathematics faculty it turned out that a top scientist, in his own words, was not going to come out of him, and not much money was arriving from Severodvinsk. So from his third year he worked at a small firm near the university: three months answering user calls, which he disliked, three more assembling and servicing computers, then lead of the support group because student turnover was frantic, and finally an invitation from a new head of department to write code. Next came a master's programme run by IBS for MIPT students and a systems analyst diploma on business process engineering; analysis was not to his taste and he went back to development.

The turning point, he says, came a little over a decade ago: a divorce from his first wife, a firm that was not much fun and, by his own account, a weak technical level of his own. He joined woman.ru as a lead, where the launch was, as he recalls, due in March 2013; the contractor ran out of money that summer and left them to it, and the system went live in March 2014. Alexander hired for potential rather than knowledge of a framework — the team peaked at about ten — but in 2015, with the advertising market down, he could not raise their salaries, the team dispersed, and he had no wish to repeat the loop. At banki.ru, instead of the promised interesting project, he inherited enormous technical debt: the debt was paid off, it never got interesting, and he left in under a year and a half. Then came three and a half months at an advertising startup where the pitch did not match reality at all — ever since, he asks where the previous person went. That advertising domain brought him to Tinkoff, to a vice-president: front end, an experimentation system, a little backend and an ad-buying engine left from Tinkoff Digital.

The host brings up 68 certificates on LinkedIn; Alexander replies that the number is beyond counting and that they came out of unfinished business with university study: combinatorics and what is now called ML were taught by people responsible for risk at a large bank, so divorced from applicability that his brain rejected it. When massive online courses appeared, someone recommended Model Thinking — the course behind a book later translated into Russian — and after it he signed up for everything: programming languages, economics, game theory, graph algorithms and effects in social networks, voting and the division of common goods, Scala, a Stanford database course, logic and ML. He ran six to eight courses in parallel, and over eighteen months of what he calls a learning fit he counted up 1,200 to 1,400 academic hours. He paid for the courses so that stinginess would make him finish them.

The most memorable and one of the hardest, he says, was the course on voting and dividing common goods, with its example of splitting a cake so nobody leaves aggrieved: one person cuts, the other chooses; the cutter makes the halves equally valuable to himself, the chooser takes the one that looks sweeter, and total satisfaction comes out above a hundred per cent. A generalisation of the approach, Alexander notes, was patented for dividing property in a divorce — which amused him, as he had divorced relatively recently himself. The same course covered the theorem that no ideal voting system without a dictator exists: a candidate with no chance of winning still pulls away part of the attention and swaps the top two around. His motivation to keep learning is simple: the more you know, the more blank spots appear on the boundary. People, by contrast, tire him out — as a lead he bought camera gear, walked around Moscow and read some thirty books on photography.

Stopping, he argues, is not an option: he is reading a book on the brain's neuroplasticity — the original came out in 2007 and was a bestseller, while the Russian edition appeared only in 2023 — and retells its central idea: unused skills lose their neural connections to other activity. On his Telegram channel he now writes about books less and about papers more: how microservice architecture is arranged in big tech, how Google approaches its architectural processes, what Amazon recommends for reliability and failure isolation. White papers strike him as denser: a book author has to fill three hundred pages, and the idea gets smeared thinly across them — Nassim Taleb, in his view, is famous for exactly that. Reading papers is no strain for him, RFCs being no simpler, but in his reviews he often redraws the diagrams: authors show a before and an after with the elements moved around so that the two cannot be matched. A harmful category of its own is IT books translated by Russian publishers: he read the third edition of Tanenbaum's distributed systems textbook pen in hand, against the original.

He started writing reviews as external memory: from a new book he extracts the increment rather than its structure, so he needs a snapshot to return to and point people at — instead of a ten-hour book, a ten-minute text, as happened with Team Topologies, which he now opens less often than his own article about it. Talks grow out of a work problem: saying "we had a problem, we did one-two-three" is not enough, you need a survey of the alternatives, the practice and conclusions a listener can apply — and giving the talk helps him let the topic go. He began speaking regularly in 2018, deciding that if you want a story told the way you like, you tell it yourself; for one conference he assembled the slides overnight in a hotel in Ulyanovsk on half an hour or an hour of sleep, and from next year he plans to slow down and write a book. As the industry's main problems he names the tools and hardware no longer available, the shortage of people, and the cargo cult of digital transformation.