PRODUCT LEADERSHIP|GROWTH-HACKING|UNIT ECONOMICS
all things value: creation → delivery → capture → scaling
Sber · Magnit OMNI. GMV, ARPPU, reach, retention.
Press a lens to highlight what is relevant, for example Growth & Economics, or just read.
Structural analysis to find the 20% of moves that carry 80% of the impact, and a defensible no to the rest.
AI made prototypes cheap; judgment is the scarce asset. Opportunities framed in commercial terms, demand validated before building.
Most problems are discipline problems in tooling costumes. Heavy machinery only for hands-on AI builds and synthesis at scale.
Judge products by the behavior they install, not the features they list. Name what becomes scarce when capability makes something abundant.
Probes, not opinions. Hunt the unnamed pain; run every idea through an adversarial loop ending in the cheapest disconfirming test.
Wrapped this methodology into a Claude skill. Try it yourself. It runs the four kill-gates on your own roadmap or build-vs-buy question.
Things I build to keep the judgment sharp.
My methodology as an installable skill: four kill-gates that decide what not to build. Willing to answer "don't build anything."
Turns Claude into a technical illustrator: exploded views, cutaways, explorable figures on faint graph paper. A figure becomes a model that yields many views.
Faculty of Physics, NSUSemiconductor Physics, a statistical foundation