Pavel Durov Built Telegram for a Billion People. He Still Won't Carry a Phone.

Pavel Durov is the founder and CEO of Telegram, a messaging app Lex Fridman notes is actively used by more than a billion people. Fridman spent a few weeks with him in Dubai and came away calling him one of the most principled and fearless people he has met: no alcohol, a stoic routine of diet and exercise, and a phone used almost only to test Telegram features. The company behind that scale is startlingly small. The core engineering team is about 40 people, covering backend, frontend, designers, and system administrators.

Durov's case for freedom starts with a move. He was four when his family left the Soviet Union for northern Italy, and even a child could see the difference: more kinds of toys, more ice cream, more cartoons. The second lesson arrived later. Without freedom you do not merely fail to enjoy abundance. You also cannot contribute to it. That is the insight behind a line he has written in Russian: freedom matters more than money.

Fridman asks how those values survive money, influence, and power. Durov's answer is that fear and greed are freedom's biggest enemies, and you keep them out of the way. Imagine the worst thing that can happen, make yourself comfortable with it, and nothing is left to be afraid of. It is worth living by your principles even if that life is shorter than a longer one lived in slavery. He does think about death. The instinct of self-preservation never really goes away, but once you die you stop experiencing life, so a life organized around the fear of dying is a poor bargain. Remembering that you can die any day, he says, is what makes the days you get count.

Durov's public rule is more than 20 years of abstinence from alcohol, tobacco, coffee, pills, and illegal drugs. The alcohol piece, he says, was easy. At 11 a biochemistry teacher handed him a book the teacher had written, *The Illusion of Paradise*, describing what addictive substances do inside the body. Drink, and brain cells become paralyzed, "literally zombies"; the next day some of them die and do not come back. If the brain is the most valuable tool you have, destroying it for short-term pleasure sounds, to him, ridiculous.

Fridman notes that many people drink to socialize and fear looking like the outcast at a party. Durov's first advice is not to be afraid of being contrarian. His second is diagnostic: if you feel you need to drink, there is a problem you are trying to conceal, a fear you will not confront. Alcohol is a spiritual painkiller that works for a few hours and then has to be repaid with interest. The tribal fear of being left out made sense tens of thousands of years ago, when exile meant starvation. Today it is irrational. If you do what everyone around you is doing, you have no competitive advantage, and you will not become outstanding.

That is also his advice on work. To contribute, you have to do something valuable, which usually means something other people are not already doing. The difficulty is that feeds and news make copying the default. Durov wants people to curate their sources, pick a field, and hunt information until they can become the world's number-one expert in it, which he says is less hard than it sounds because almost nobody else is trying. He applies the same long-term arithmetic to pornography: it is a surrogate that spends energy and inspiration on a fleeting pleasure. If you can access the real thing you do not need it; if you cannot, that is a deficiency to fix rather than anesthetize. He tries to act as if he will still be around tomorrow. Trading the mid and long term for the next hour does not make sense.

Fridman spent two weeks with Durov and barely saw him use a phone the way most people do. Durov does not think a phone is a necessary device. He grew up without one, stayed without one as a university student, and once he had one kept it in airplane mode or muted because he hated being disturbed. The philosophy is simple: he wants to define what is important in his life. He does not want other people or companies telling him what to think about today. The phone gets in the way of that.

He tries to protect quiet mornings. He allocates as much time as possible for sleep — even 11 or 12 hours — knowing he will not sleep that long, and then uses the leftover time lying in bed. Other people reach for a sleeping pill. He never takes pills, and those hours are his favorite part of the day, when ideas arrive. A shower and morning exercise without a phone can do the same. Open the phone first thing, he says, and you become a creature told what to think about for the rest of the day. The same is true of late-night social-media news.

The conclusion is counterintuitive for someone who built VK, one of the world's largest social networks, and then Telegram. The more connected and accessible you are, the less productive you are. The entire world can be fascinated by a quarrel between the world's richest man and the world's most powerful man, and for almost everyone watching it will change nothing they can affect. Durov estimates that 99% of people do not need to be up to date on everything, and that perhaps 95% of what gets published is written because someone wants you to buy a product, support a cause, fight a war, or donate money. He reads a piece of news and asks who benefits from his reading it. Supporting a cause you actually believe in is fine. Being manipulated into other people's wars is not.

Durov does get angry, lonely, and sad. He just treats those states as something he is responsible for working through. Self-discipline, he says, is how you break the loop of negativity that turns into depression. He does not remember being depressed in at least 20 years. The method is not rest. You identify a problem, see a solution, and start executing. People think they should recover energy first and then act; he says you gain energy by doing something. Write a small piece of code. Go to the gym on a day you do not want to. The good feeling comes after the first move, not before it.

The morning ritual is 300 push-ups and 300 squats, then the gym five or six days a week for one to two hours. He is not sure the bodyweight work changes his physique much. Squats are especially boring. That is the point: they train the muscle of self-discipline, and if that one is trained, everything else follows. Ice baths and the Eastern European banya — extreme heat, herbs, branches, then a cold plunge — are the same exercise. In the moment they can be unpleasant. Then you feel good for hours or days, with long-term health benefits he describes as alcohol in reverse: a few minutes of cost instead of a short pleasure you pay for later.

Swimming is the patience drill. His longest swim was five and a half hours in cold Finnish water; he got lost and barely found his way back, and says he did not panic. You start reluctant, swim ten minutes, then 20, then 30, and the process teaches the patience any large project needs. The harder argument is physiological. Brain efficiency is limited by how much sugar and oxygen the heart can push through blood. Physical training is the only way he knows to raise that ceiling, which is why he was surprised, more than a decade ago, that so few CEOs trained. The old stereotype that strong people cannot be smart is, he says, lunacy. When he cannot train, he feels stress immediately, so push-ups and squats continue even when a gym is impossible.

Diet follows the same rule. Processed sugar is addictive and makes you hungrier; intermittent fasting inside a six-hour window puts structure back into the day. He stopped red meat about 20 years ago because it made him feel heavy, and now eats seafood and vegetables. Pills, for ordinary headaches, mute a warning light instead of removing the cause, like smashing a helicopter alarm instead of turning away from the mountain. He is not arguing for the Middle Ages — antibiotics exist — but he thinks pills are overused because the industry is not incentivized to eliminate root problems. The same incentive audit applies to news, to alcohol, and to a clouded mind that can only follow mainstream opinion.

Telegram's core engineering team is about 40 people. Durov says they learned early that the quantity of employees does not translate into the quality of the product, and often does the opposite. Too many people spend 90% of their time coordinating the small pieces they each own. Some will not have enough work, still collect a salary, and then look for a purpose by inventing problems inside the team. Refuse to let people hire help, and they are forced to automate. Telegram has tens of thousands of servers, almost 100,000, across continents and data centers. Manage that by hand and you hire thousands. Manage it with algorithms and the system becomes more scalable, more efficient, and, he argues, more reliable.

Humans are attack vectors. A distributed system that runs itself can be more secure and faster, and if one part of the network goes down, traffic can shift. Privacy is designed the same way. Since 2012 the team has tried to make sure no employee can access private messaging data. Cloud chats are encrypted; decryption keys are split across data centers in different jurisdictions, so pulling the hard drives from one location yields ciphertext. Telegram has never shared a single private message with anyone, including governments and intelligence services, Durov says, and there has never been a leak of private messages or contact lists. End-to-end encrypted secret chats add another layer, but they have limits, so the product still relies on the encrypted cloud. Asked whether he might share private messages with governments in the future, he says the system is designed so that it is impossible. Changing it would break a promise. He would rather shut Telegram down in a country than do that.

Where does he find the nerve to say no to intelligence agencies? Part of it, he says, is that he has stood his ground since he was a child pointing out teachers' mistakes. They can threaten him, but the worst case is death, which returns him to the opening argument: there is no point living in fear. Telegram is successful, but losing one market, or all of them, would not change his lifestyle. He would still do his push-ups. He owns 100% of Telegram. There are no shareholders, which he calls quite unique among major tech companies, and that ownership is what lets the project run on principles. The right to privacy is in most Western constitutions, he notes, and is still under attack almost every week, usually starting with well-meaning proposals to fight crime or protect children. No dictator announces that he wants more power. The restrictions arrive in stages, and after a few years people cannot protest, assemble, or send a message that is not monitored.

In August 2024 Durov arrived in France for a two-day trip and was met by a dozen armed police. They read a list of something like 15 serious crimes. At first he thought it was a mistake. Then he understood they were accusing him of responsibility for crimes some Telegram users had allegedly committed — something, Fridman and Durov agree, that had not been done to a tech leader at this scale, even by authoritarian states. They put him in a small room with no windows and a narrow concrete bed. He spent almost four days there, answering questions about how Telegram works. What struck him was how little the people who started the investigation understood about technology, encryption, or social media.

Fridman reaches for Kafka's *The Trial*: a man arrested for a crime nobody can explain, stuck in a system that is everywhere and nowhere, until the bureaucracy breaks his spirit. Durov agrees the situation is absurd. If France had a good-faith problem, there were conventional ways to reach Telegram, as other countries have done dozens of times. The company has been taking down publicly distributed harmful content for a decade, with daily transparency reports for about eight years on child-abuse and terrorist channels, and millions of pieces of content removed every week. He suspects some of the outrage is emotional, driven by newspaper stories rather than data. In France he is not on trial. An investigative judge — a role he says is more like a prosecutor, which makes appeals harder — is investigating him. Telegram appealed the investigation itself last year and still had no hearing date. Other entrepreneurs told him horror stories of businesses paralyzed by such judges, only to be cleared after years of lost time.

For a long stretch he could not leave France; by the time of the recording he could travel to Dubai, where Telegram is headquartered, and nowhere else. He says he is patient and will not let the restriction dictate his actions — if anything he will double down on defending freedoms, because those four days taught him what their absence feels like. If French authorities wanted a backdoor into private messages, his response would be nothing, and not polite. He thinks France is losing the battle. The more pressure he gets, the more resilient and defiant he becomes. Attempts to use his confinement to make him block channels or change how Telegram works were refused, then disclosed, because he does not sign nondisclosure agreements with people like that. He would rather lose everything than yield. Agreeing to something that violates other people's rights, he says, would leave him a shell of himself. In a hypothetical 20-year prison sentence he would starve himself to death rather than "do something stupid."

Unable to leave France, Durov was offered a meeting with the head of the country's foreign intelligence service through a well-known French tech entrepreneur. He took it without promising anything. The request, as he tells it, was to restrict speech in Romania during a new presidential campaign after the previous results had been canceled: shut down Telegram channels that supported a conservative candidate or protested against the pro-European candidate France preferred. Durov's line was the product's rules. You cannot call for violence. Peaceful protest and peaceful debate are not political censorship he is willing to perform, in Asia, Eastern Europe, the Middle East, or Europe. He told the intelligence chief that if confinement was supposed to make him obedient, the opposite was more likely. He never signed an NDA, and he later disclosed the conversation. He considers the episode an attempt to humiliate him and millions of Telegram users, and to use his legal situation for geopolitical ends.

Moldova had come first, in September or October after the August arrest. Through an intermediary, French intelligence asked Telegram to work with Moldova's government on channels during an election. Durov's team said they would look for rule violations. A short list arrived; a few channels and bots did violate the rules and were taken down; the rest were left. A second list of dozens of channels followed. Telegram found no solid basis to ban them and refused. Then, Durov says, the same service let him know that after the first takedowns they had spoken to his investigative judge and said good things about him. The two matters had nothing in common. A judge looking at illegal content in France had no business hearing about Moldova. That was the moment he stopped assuming the case was only technological confusion. By the time Romania came up, he was already wary.

He insists Telegram has no side. French police asked the company to take down a channel of far-left protesters in France; the team looked, saw peaceful protest, and refused. During COVID they protected both Black Lives Matter organizers and anti-lockdown protesters, as long as no one called for violence or damage to public property. Fridman turns the conversation to what Europe would have to fix for founders. Durov's answer is the size of the state. French public spending, he says, is 58% of GDP, as much as or more than late-stage Soviet Union, with more people representing the state than building products. He describes a location-based social network, France's one real success in that field in a decade, sold to Snapchat after the founder ignored Durov's advice to hold on. The founder later cited a drawn-out data-protection investigation, eventually producing no charges, that had been sucking the blood of a young company already competing with Facebook, Instagram, and Snapchat. Another French founder had accounts frozen in a tax investigation that lasted eight years, killed the business, and ended with a finding that nothing was wrong. He rebuilt in Dubai, succeeded, and when France later offered a Champs-Élysées location, funding, and tax breaks, he said never. Unless societies count the companies that are never founded or are sold too early, Durov says, nothing changes.

At 11, Durov entered an experimental class at the Academic Gymnasium of Saint Petersburg State University after a rigorous exam. The idea was to pour as much as possible into a teenager's brain, with a focus on math and foreign languages, and see whether the rest of the disciplines would follow. Students took at least four languages — Latin, English, French, German — plus optional Ancient Greek, and classes such as biochemistry, psychoanalysis, and evolutionary psychology. Becoming a straight-A student was considered crazy. Durov had been kicked out of other schools for challenging teachers. In this one, debate was allowed. Bored as a teenager, he decided to attempt the impossible and take the top mark in every subject, using even the breaks between classes. Math, he still thinks, is essential: it trains you to split problems, sequence them, and retry, which is the same skill as programming, project management, and starting a company. Competition is the other lesson. Remove public rankings to spare children anxiety, and you eliminate losers and winners together. Kids compete in video games instead, then meet a more brutal contest in adult life.

His brother Nikolai is a mathematician and cryptographer: three IMO gold medals, two ICPC titles, two mathematics PhDs. Durov says he learned almost everything from him. They shared a bedroom; Nikolai was Wikipedia before the family had internet, answering questions about dinosaurs, galaxies, and Neanderthals. He started reading at three and by six was absorbing astronomy books in public, which got their mother scolded by strangers who assumed the child could not possibly understand. Durov started coding at 10 because he did not have enough video games. Scarcity, he argues, produced creativity that abundance of entertainment can drown. He built two-dimensional turn-based games, then a five-in-a-row tic-tac-toe on an infinite field so he could beat classmates by training against a computer his brother helped optimize. At university, bored, he built a site of digitized lectures and exam answers, then a forum, then a network that spread across faculties and universities until it had tens of thousands of users, complete with friends lists, albums, and blogs.

After graduation a former classmate asked if he was building a Russian Facebook. Durov had to ask what Facebook was. Seeing it told him which features to cut so the product could scale — existing users always want more complexity. He started VKontakte, "in touch," to stay connected with classmates and, as a 20-year-old, to meet people. He refused third-party libraries and obsessed over every line. Stuck on how to begin, he called Nikolai, then a postdoc at Max Planck in Germany, who told him to build only a login module: type email and password, see your name, and the next step becomes obvious. Durov calls that some of the best advice of his life. He built the first version of VK in two weeks and was the only employee for almost a year: backend, frontend, design, support, and marketing. The stack began as PHP, MySQL, and Debian Linux, then Memcached and NGINX; when the service passed a million users and drew DDoS attacks that intensified on negotiation days, Nikolai came home around 2008 and, with another programming-contest champion, rewrote large engines in C and C++. By 2009, visiting Silicon Valley and meeting Mark Zuckerberg, Durov was being asked why VK loaded faster than Facebook even there. Speed is still the core idea. Users, he says, can feel a 50-millisecond difference, and faster code also costs less to run. Telegram now opens dozens of times a day for more than a billion people; a half-second delay, multiplied out, is centuries of human time wasted for no reason but sloppiness.

Durov has fired engineers and watched productivity go up. In one story, two Android engineers were behind, and he considered a third hire until he noticed one of them lagging, complaining, and dodging responsibility. He fired that person and realized within weeks that the third seat had never been needed. A-players can tell when someone next to them is not actually working, and the insult of sharing a codebase with a distracted colleague is, in his view, more damaging than a missing headcount. Often the issue is not laziness but the inability to focus on one task for a long time. Telegram hires mainly through coding contests: anyone can enter, winners take prize money, and consistent winners get jobs. A 21-year-old who has competed since 14 and won eight contests in JavaScript, Android, and C++ beats, in his ranking, a LinkedIn hire from Google who is used to meetings and shared blame. Because the Android client is open source, a contest for a gnarly profile animation can also produce the solution itself, saving months of development. People who already use Telegram understand what it stands for. People who do not, he says, are a crazy hire.

That selectivity is how Telegram keeps shipping features first. Secret chats got an auto-delete timer in 2013; WhatsApp followed seven years later and at first even copied the exact 1-, 3-, and 5-second options, unsure what the magic was. Reply snippets — a small quote, a tap that highlights the original, a vertical line on the left — were arbitrary design choices that the rest of the industry then cloned. The default chat background is four colors, not two, constantly shifting under a vector pattern that must be interesting without stealing attention. Durov says he looked through several thousand variations for the default, and users can set their own four colors. Sending a message morphs the input field into the bubble across every content type, device, and OS, including old phones on new software and iPads that make everything huge. The deletion animation Fridman calls a Thanos snap — tens of thousands of particles, like dust — was especially hard on Android. Play the animation first and then close the gap, and it feels too long. The particles and the collapsing thread have to happen together, which means redrawing the whole screen. Tens of millions of people delete messages every day. If a detail can lift their mood even 0.001%, he thinks the time is worth it.

Stickers arrived three years and eight months before WhatsApp. WhatsApp's first version used GIFs or WebM. Telegram used vector animations of a few kilobytes, at most 20 or 30, with 180 frames running at 60 frames per second on every device. Nobody had tried it, Durov says, because it was crazily difficult, and finding artists who can draw in curves and formulas required another round of contests. The same pipeline later produced custom animated emoji, still unique to Telegram in his telling, including a hundred of them moving in one message without crashing the phone. Hold-to-send animations, reactions that make a like feel worth repeating, and later Telegram Gifts sit in the same place: the intersection of art and engineering, mostly invisible, felt rather than noticed. People like it, he says, when other people care.

Telegram launched in 2013, after Edward Snowden had made the scale of surveillance obvious. The team already considered the product extremely secure, but they wanted a design in which even they could not read messages, and they knew a group of people born in Russia would not automatically be trusted. So the client apps have been open source on GitHub since 2013, and secret chats added end-to-end encryption a year and three months before WhatsApp began testing it, and about three years before WhatsApp rolled it out in 2016. Durov thinks the rest of the industry moved because Telegram set the standard. End-to-end encryption cannot easily support huge groups with persistent history, one-to-many channels, busy bots, or smooth multi-device sync, and people lose files. Telegram's hybrid is opt-in secret chats for the paranoid — no screenshots, no forwarding — and encrypted cloud for everything else. The company builds Telegram in Telegram and found secret chats unusable for large-file teamwork. For someone who trusts only code, Durov still calls them the most secure means of communication among popular apps.

Telegram is, he says, the only popular messenger with reproducible builds on both Android and iOS, so researchers can check that the App Store binary matches the GitHub source. WhatsApp has never been open source; Signal and iMessage do not offer that dual-platform guarantee. Most of the stack is built in-house, including database engines, web servers, and the programming language used for the backend API, on the lesson from Snowden that widely used open-source libraries can hide flaws. Linux on the servers is the main exception. Durov never worked with Snowden in person but keeps in touch and calls the disclosures brave. The lesson was that cryptography experts can turn out to be promoting flawed standards, and that a government answering 9/11 can overreact until the cure is worse than the disease.

In 2018 Telegram had about 200 million users. Iranians used it to organize protests; the government banned it, while WhatsApp was spared or quickly unbanned. Russia demanded encryption keys that do not exist and, when Telegram said a ban was the only option, blocked the app in spring 2018. Telegram rotated IP addresses automatically, burning through millions of them, and launched a "digital resistance" so administrators worldwide could run proxies. The campaign cost millions of dollars. Russian censors then blocked such large subnets that grocery payments, local social networks, media, and banks broke, and they had to become more selective. The harder opponent was Apple, which refused App Store updates for at least four weeks and then offered a global update excluding Russia. Durov refused. A new iOS release was breaking old Telegram for users worldwide. He set a Friday 6 p.m. deadline: no approval, and Telegram would leave the Russian App Store. Fifteen minutes before that time, Apple approved the update. In Iran, buying IPs at that scale risked sanctions, so Telegram let engineers run proxies that showed a pinned ad the operator could sell. The market kept tens of millions of users. Telegram is still banned in Iran, Durov says, with something like 50 million people on it anyway.

In spring 2018, while raising money for TON and fighting bans, Durov came home to a rented townhouse and found that a neighbor had left something by the door. An hour later, in bed, he felt pain everywhere. Walking to the bathroom, body functions switched off: sight, hearing, then breathing, with acute pain in the heart, stomach, and blood vessels. He was sure he was dying. He collapsed and woke the next day on the floor, too weak to stand, blood vessels broken across his body, unable to walk for two weeks. He did not tell most of the team. He had never discussed the episode publicly before this conversation. It did not make him more afraid of governments. After you survive something like that, he says, you feel you are living on bonus time — in a sense you died already, and every new day is a gift.

The first time he assumed something similar was coming was December 2011. Mass protests in Moscow followed a disputed State Duma election. The government demanded that VK take down Navalny opposition groups with hundreds of thousands of members. Durov refused publicly, posting a scan of the prosecutor's request next to a photo of a dog in a hoodie with its tongue out. Armed police then tried to enter his apartment. He asked himself whether he had been right, decided he had, and concluded the next step was prison. He told himself he would starve to death there. He also realized there was no secure way to warn his brother: WhatsApp already existed but transmitted plaintext. He promised himself that if he survived he would launch a secure messenger. He packed a backpack, moved to a hotel, and started designing Telegram, knowing his days in Russia were numbered. He still thinks a restored Russian ban would be a huge mistake, because independent outlets such as the BBC now reach Russians mainly as Telegram channels. Protecting speech, he says, is not a way to make friends. Every government calls its opposition foreign agents; Ukrainians say he works for Russia, Russians say he works for Ukraine. When the 2022 war started he suggested pausing political channels on both sides. Users in both countries told him not to babysit them.

He owns 100% of Telegram, takes a salary of one dirham — a third of a dollar — and says he has invested on the order of a hundred million dollars of his own money. 2024 was the first profitable year. The company does not run ads on private data, contacts, metadata, or activity, only context ads against channel and group topics, which he estimates leaves about 80% of ad value on the table. It also has no news feed, the engagement engine he thinks damages minds at scale. Premium launched in 2022 at four to five dollars a month, now has more than 15 million paid subscribers, and is on track for more than half a billion dollars this year, with over 50 features in the bundle. TON, originally Telegram Open Network, was built in 2018 and 2019 because Bitcoin and Ethereum could not absorb Telegram's load; Nikolai designed native sharding. The U.S. SEC objected to the fundraise, Telegram abandoned the launch, and the contest community shipped it as The Open Network. Usernames can be owned as smart contracts Telegram cannot confiscate. Channel ad revenue is split 50/50 and withdrawn in TON; mini-app developers pay a 5% commission. Telegram Gifts, vector collectibles that sit next to a name, sold $12 million of a Snoop Dogg drop in 30 minutes. An early collection priced around $5 now has a floor near $10,000. One person made several million dollars trading gifts in months, and $12 million from mini apps with a team of two or three. Durov bought his first few thousand bitcoin in 2013 near $700, ignored the crash, funds his own life from that investment rather than from Telegram, and thinks bitcoin can reach $1 million because governments keep printing money and nobody is printing bitcoin.

A Russian prison riddle asks which of two unbearable chairs you would choose. Durov's answer is to refuse both: take the spikes from one and use them on the other, turning a disadvantage into a tool. Politicians and bosses, he says, often force a choice between two bad options so that whatever you pick becomes your responsibility. About 15 years ago a friend asked him to be a sperm donor rather than use an anonymous stranger; he found it absurd, then realized other couples had the same problem, and was persuaded to do more of it, especially when he thought he might not have long to live. He does not know how many biological children he has. Writing his will, he decided not to distinguish children he had raised from children he had never met: shared DNA, even proven 30 years later, entitles a share of the estate. He would not release the money for the first decades of their lives. Overabundance, he says, paralyzes motivation and willpower, especially in boys who can be proud of a father's wealth instead of their own work. Splitting a fortune among more than a hundred people was supposed to keep each share from being too large. He admits the remaining tens of millions per child may still be too much.

That worry is the point of John B. Calhoun's Universe 25. Mice given unlimited food, water, nests, and safety doubled, then stagnated; mothers neglected or killed young; violence and hypersexuality spread; well-groomed "beautiful ones" withdrew from mating; fertility crashed; and the last mice died beside untouched food. Durov's reading is that humans evolved to overcome scarcity. Infinite food and entertainment were never part of the species' history, and in a world where everything is available, purpose becomes hard to see. Restrictions should come from inside. He considers himself lucky, counterintuitively, to have grown up poor: the same jacket for years, a professor father unpaid for months as the Russian state nearly went bankrupt, a mother working two jobs. It was not easy. It created priorities. Fridman notes that AI may manufacture a human version of that abundance, cutting suffering while stressing the evolutionary fabric that holds societies together. Durov's close is a call for balance between chaos and order, self-restriction and room to create.

His father had just turned 80, a scholar of ancient Rome who typed books on a late-Soviet typewriter and, Durov says, taught by relentless example. Children discount words and copy actions. You can do the right thing nine times out of ten, his father told him, and they will copy the one mistake — including a parent who forbids phones while using one constantly. In their last conversation the older Durov also said AI can have consciousness and creativity but not conscience: not morality, not rooted principles, not integrity as humans mean it. Durov is cautious about manifestation. Goals, optimism, and confidence help, but he does not believe things appear without effort. Optimism plus logical action over many days is what he trusts. Fridman offers a picture of effort reshaping the landscape of probabilities; Durov answers with quantum immortality, the many-worlds idea that consciousness only continues in branches where it does not die, so some version of a person might reach a thousand years old while other worlds think he is long gone. After they stop, Fridman stays with Kafka. In *The Trial*, Josef K. is walked to a quarry and stabbed without resisting. Tyranny's final victory, Fridman says, is not the killing. It is when you hold still for the knife because you have been exhausted into submission. He saw none of that exhaustion in Durov.