Lessons

You change the button, sales rise, you conclude the button worked, and you may be completely wrong. Learn why comparing before and after is one of the most reliable ways to fool yourself, what a controlled experiment does differently, why randomization is the whole trick, and how the A/B test turns a guess about cause into something you can actually measure.

Network effects are supposed to produce monopolies, yet plenty of networked markets have several healthy competitors. Learn what makes a market tip toward one winner, the four forces that stop tipping and keep markets competitive, why multi-homing is the quiet enemy of every platform, and how to tell a genuine winner-take-all market from one that merely looks like one.

A marketplace with no buyers has no sellers, and with no sellers has no buyers. Learn how platforms solve the cold-start problem that kills most of them, why the answer is usually to subsidize one side heavily, how pricing in a two-sided market breaks the normal rules of business, and the concrete tactics, from faking one side to going town by town, that actually ignite a network.

Some products get better the more people use them, and that single property explains why a handful of platforms dominate the digital world. Learn what a network effect actually is, the difference between direct, indirect, and data effects, why Metcalfe's law overstates the case, and how the same feedback loop that creates a monopoly also makes it fragile at the start.

A creator can have a huge following and no business, because the following belongs to the platform, not to them. Learn the difference between a rented audience and an owned one, why platform risk is the defining threat, and the move every durable creator makes: use the platform to find people, then take them somewhere you own.

The creator economy is worth hundreds of billions of dollars, and most creators in it make very little. That is not a failure of the market; it is the shape of the market. Learn why creator income follows a brutal power law, how the same feed algorithm that finds hits also concentrates rewards, what the 1,000 true fans idea gets right and wrong, and what the math means if you are the creator.

A creator with a million followers can earn a fortune or almost nothing, depending entirely on how the audience is monetized. Learn the seven distinct income models, from ad revenue share to sponsorships to memberships, who carries the risk in each, which ones depend on a platform's goodwill, and why the mix a creator chooses matters more than the size of their audience.

Search is shifting from a list of links to a written answer, and a large and rising share of searches now end without a click. Learn what changes when an AI writes the answer, why 'being cited is the new ranking', how generative engine optimization differs from classic SEO, and which parts of search are genuinely new versus the same old mechanics with a new surface.

Thousands of pages match your query; the engine picks ten and orders them. Learn the four questions ranking answers, why links became a proxy for trust, what PageRank actually measured, how the query itself is interpreted before anything is scored, and why chasing individual ranking factors is the wrong mental model.

Before a search engine can rank a page it has to discover it, fetch it, understand it, and store it, and most of the web never makes it through. Learn how crawlers find pages, why a crawl budget means not everything gets visited, what indexing actually stores, why JavaScript makes a page harder to read, and the difference between crawled, indexed, and ranked.

Half the money spent on advertising is wasted; the trouble is knowing which half. Learn why measuring whether an ad caused a sale is genuinely hard, why last-click attribution flatters the wrong channels, how privacy changes broke the tracking that measurement relied on, and the three-part toolkit that replaced it: incrementality, media-mix modeling, and privacy-preserving reporting.

In the moment a page loads, a request describing you is auctioned to thousands of potential buyers and the winner's ad is delivered, all before you see anything. Learn the supply chain behind that: the buy side and sell side, real-time bidding, how you become a targetable audience, and how much of an advertiser's money is eaten on the way through.

Nobody sets the price of an ad slot by hand. An auction does it, billions of times a day, in the time it takes a page to load. Learn why advertising runs on auctions, how the second-price idea makes bidding honest, why the highest bid often loses to a better ad, and why the industry quietly switched from second-price to first-price auctions.

Two internets, same technology, different shapes. Learn the operating requirements that platforms in China work under, from licensing and real-name accounts to filing recommendation algorithms with the regulator, how compliance became a core product function rather than a legal afterthought, why micro-dramas became a format with no Western equivalent, and what all the differences add up to.

More than half of China's internet users buy things through live broadcasts, a habit the West has never come close to. Learn why livestream selling worked there and stalled elsewhere, the mechanics of a selling stream, why conversion runs far above normal e-commerce, the difference between renting reach and owning an audience, and the costs the model carries.

Douyin is not TikTok, Xiaohongshu is not Instagram, and Weibo is not quite Twitter. A guided map of the platforms 1.1 billion people actually use: what job each one does, which conglomerate stands behind it, why the ecosystems were walled off from each other, and which Western analogy helps you versus which one quietly misleads you.

Western phones hold a dozen apps for a dozen jobs. In China, one app is the messaging, the payments, the taxi, the bill, the government form, and the shop. Learn how the super-app model works, why mini-programs made bundling possible, why a market that skipped credit cards produced a different shape of internet, and what the design actually trades away.

How platforms decide what stays up, and why nobody can reliably tell you whether something was made by AI. Covers classification at scale and the precision-recall tradeoff, why AI detectors accuse innocent people at rates that make them unusable, and how provenance approaches like C2PA and SynthID try a different answer, plus where they break too.

A practical, honest account of using generative models to make social content. Covers the jobs they genuinely do well, the specific ways they fail, why a model that sounds confident can still be wrong, the disclosure rules that now apply in the EU, and why the tool that removes effort also removes the thing that made your work worth watching.

Before the tools, the map. AI does three separate jobs on a social platform: it decides what you see, it helps make the content, and it decides what stays up. Learn what each job actually is, why they are built and judged differently, and why the same word covering all three causes most of the confusion about AI and social media.

Why a brand new post can reach millions on one platform and nothing on another. Learn the cold-start problem and why collaborative filtering cannot solve it alone, how content embeddings let a system rank an item nobody has touched, the difference between a social graph and an interest graph, and what actually separates YouTube, TikTok, Instagram and Pinterest.

A feed does exactly what it was told to maximize, which is rarely what anyone wanted. Learn which signals rankers actually weight and why implicit ones beat likes, how proxy objectives produce clickbait and rage-bait as correct answers to badly posed questions, and the mechanisms platforms use to pull an optimizer back toward what people value.

Every feed has the same impossible job: pick ten items out of millions, in under a tenth of a second. The answer is a funnel. Learn why recommenders split into a cheap retrieval stage and an expensive ranking stage, how two-tower models make retrieval possible, and why nearest-neighbour search is the trick that makes the whole thing fit in a budget.

Why Unity runs physics on its own 50Hz clock instead of the frame rate, and what follows from that. Covers rigidbodies and colliders, moving things correctly with forces rather than the Transform, interpolation for smooth motion, tunneling and continuous collision detection, and where physics cost actually comes from.