TRIZ or How to Mass-Produce Genius
In the modern West, when a company wants to innovate, it summons a workshop. Sticky notes appear. Someone draws a Venn diagram. A facilitator, paid handsomely, asks everyone how they might "ideate" if they were a child, or a dolphin, or possibly a child dolphin. By the end of the day, the wall is colorful, the catering is gone, and nothing has been invented.
Meanwhile, in the more disreputable corners of Korean electronics conglomerates and before that German engineering firms, people are quietly using a system designed by a Soviet naval officer in the 1940s, and producing actual patents at a rate that should embarrass everyone else. The system is called TRIZ (from Russian - ТРИЗ - теория решения изобретательских задач - teoriya resheniya izobretatelskikh zadach, lit. 'theory of inventive problem solving') and it is one of the strangest and most useful intellectual exports of the USSR, somewhere between vodka, chess grandmasters, and the Kalashnikov on the list of Russian things that turned out to travel surprisingly well.
The Engineer Who Wrote to Stalin (it did not end well - duh...)
Genrich Altshuller was born in 1926 and became, by his early twenties, a patent inspector for the Caspian Sea flotilla of the Soviet Navy. His job, in essence, was to read other people's inventions and decide whether they were any good. He was very young, very bright, and possessed of a fatal Soviet flaw, which was the belief that things could be improved if someone in charge would only listen.
He noticed something. Most of the patents he reviewed were unimpressive. The few that were genuinely inventive seemed to follow recognizable patterns, as if the inventors, separated by decades and disciplines and ignorant of each other entirely, were unknowingly solving their problems with the same handful of tricks. A chemical engineer in 1923 and a mechanical engineer in 1948 might independently arrive at the same conceptual move. This was suspicious. This was interesting.
In 1948, Altshuller and a friend wrote a letter to Stalin complaining that Soviet innovation methodology was a disaster and proposing reforms. One imagines the friend nervously suggesting they sleep on it. Altshuller, twenty-two years old and obviously as fits his age - immortal, did not sleep on it. In 1950 he was arrested, charged with the marvelously vague crime of "inventor's sabotage", and sent to a Gulag camp in Vorkuta, north of the Arctic Circle.
Vorkuta, in a turn of events Altshuller could not have predicted, was full of intellectuals. The camp contained engineers, scientists, philosophers, and at least one professor who had survived by playing chess for bread. Altshuller, with characteristic optimism, treated this as the world's worst graduate school. He debated, argued, and refined his theory in the snow. He was released in 1954, after Stalin's death, and continued the work. He published his first paper in 1956. The system that eventually emerged is known by its Russian acronym TRIZ, which translates, with classic Soviet bluntness, as the Theory of the Resolution of Invention-Related Tasks.
What TRIZ Actually Says
The premise is almost rude in its simplicity. Altshuller and his collaborators eventually analyzed something on the order of two million patents, sorted them by what kind of problem they solved and what kind of trick they used, and discovered the following uncomfortable facts.
Innovation, the real kind, is not what we tell ourselves it is. It is not lightning. It is not the gift of unhinged geniuses pacing in front of chalkboards. It is mostly, embarrassingly, the application of one of about forty recurring conceptual moves to a system that contains a contradiction.
The forty moves are called the 40 Inventive Principles. They have names that sound either banal or vaguely menacing. Segmentation. Local Quality. Asymmetry. Periodic Action. The Other Way Around (which means doing the opposite of what seems natural). Cheap Short-Living Objects (which means replacing one expensive durable thing with many disposable ones, the principle that gave us the paper cup and, less cheerfully, the cruise missile).
A "contradiction", in the TRIZ sense, is a situation where improving one thing makes another thing worse. You want the plane to be faster, but faster means more fuel, which means heavier, which means slower. You want the chocolate to melt in the mouth but not in the hand. You want the user to feel powerful but not responsible. These are the actual atoms of design, and most engineers spend their lives trading off between them, settling for a compromise. TRIZ insists that the truly inventive solution does not compromise. It eliminates the contradiction by changing the system so the trade-off no longer exists. The chocolate gets tempered. The plane gets different engines. The user gets a feature that obscures consequence.
To find which of the forty principles is likely to help you, TRIZ provides a contradiction matrix, which is exactly what it sounds like, an enormous table where you look up your problem like a verb in a Latin textbook and it tells you which conceptual tools have historically worked. It is, in a sense, the world's first crowd-sourced creativity guide, except the crowd is dead and the guide is in Russian (some of the crowd were dead precisely because they were Russian and too creative).
Behind all this sits the most quietly subversive idea Altshuller produced, the Ideal Final Result. Before solving a problem, TRIZ asks you to describe what would happen if the system did exactly what you wanted with zero cost, zero complexity, and zero side effects. The function performs itself. The seatbelt fastens you. The drug delivers itself only to the sick cell. You are not allowed to mention how this happens, only what it is. Then you work backward toward something a human can actually build. It is unbearably useful. It is also a small philosophical bomb, because once you start asking "what if the thing simply happened", it becomes painfully clear how many products exist mainly to justify the company that sells them.
The Philosophical Bombshell Nobody Wants
Here is where TRIZ gets uncomfortable. If invention follows patterns, then "genius" is doing a lot less work than we have been led to believe. The romantic story of innovation, the one starring lone Edisons and barefoot Jobses and rumpled Einsteins, starts to look less like history and more like marketing. Altshuller's quiet claim, made from a Soviet apartment by a man whose monthly pension was probably insufficient fo a good set of clothes, was that the structure of breakthrough is impersonal. Anyone trained in it can do it. The cult of the founder is mostly theater.
This is, of course, the kind of claim that Silicon Valley cannot afford to take seriously, because its valuations depend on the opposite belief. So TRIZ tends to live in places that care about results more than mythology. Samsung built a whole internal program around it in the 2000s and has credited it with billions of dollars in savings and a noticeable bump in patent quality. Hyundai uses it. Intel has used it. Boeing, Procter and Gamble, and a long tail of less glamorous firms quietly use it. In every case the playbook is the same. Hire a few people who actually understand TRIZ, ignore them in public, deploy them in private, watch the patents pile up.
Why It Matters Now, In The Year Of Our LLM
We have arrived, more or less, at the moment when large language models can be aimed at problems and asked to brainstorm. They are quite good at producing plausible-sounding ideas at a volume previously available only to undergraduates on coffee. They are very bad at noticing which ideas are actually inventive, because they have no internal compass for what an inventive move even is.
This is exactly the gap TRIZ was designed to fill. A model that has internalized the forty principles, the contradiction matrix, and the discipline of the Ideal Final Result is doing something categorically different from a model that is just sampling adjacent tokens about turbines. It is doing what Altshuller did at the patent office, but at a speed he would have found indecent. A small but growing number of researchers are working on exactly this combination, and the early results are the kind of thing that should make anyone selling generic "AI innovation consulting" nervous.
There is a broader cultural point too. TRIZ is a rebuke to the idea that creativity is sacred and unteachable. It is also a rebuke to the opposite idea, currently fashionable, that creativity is just remixing whatever happens to be in the training data. TRIZ says creativity is the resolution of contradictions, and contradictions are real structural features of the physical and social world, and you can get better at resolving them with practice and structure. It is a profoundly grown-up theory of invention in a culture that mostly prefers childish ones.
The Strange Inheritance
Altshuller died in 1998, in Petrozavodsk, having outlived the system that imprisoned him by seven years. He wrote science fiction in his spare time, under the pen name Genrich Altov, and managed to predict a number of technologies that later showed up in the real world, which is the kind of thing TRIZ practitioners take as evidence and TRIZ skeptics take as coincidence.
The system he left behind is one of those small ironies of intellectual history. The USSR, which was supposed to industrialize everything including the human spirit, accidentally produced a workable theory of personal creative breakthrough. It then exported this theory through emigration, samizdat, and the slow leakage of useful ideas across borders, until it ended up improving the next generation of Korean phones and German cars and possibly, soon, the AI systems that some people think will replace inventors entirely.
If those systems do replace us, they will be using Altshuller's principles to do it, which means that a young naval officer who once wrote an angry letter to Stalin will have, in the end, helped shape the machines that succeed the human inventor. This is either deeply funny or deeply sad. TRIZ, naturally, would suggest the contradiction can be resolved. We just have to imagine the ideal final result first.