{
  "title": "Articles/lean-startup-is-a-sensor-deep-tech-is-an-engine",
  "caption": "Lean Startup Is a Sensor. Deep Tech Is an Engine.",
  "slug": "lean-startup-is-a-sensor-deep-tech-is-an-engine",
  "tags": [
    "article",
    "deep-tech",
    "hermes-published",
    "lean-startup",
    "pack-8",
    "published"
  ],
  "canonical_url": "https://mosiah.org/articles/lean-startup-is-a-sensor-deep-tech-is-an-engine/",
  "interactive_url": "https://mosiah.org/#Articles%2Flean-startup-is-a-sensor-deep-tech-is-an-engine",
  "markdown_url": "https://mosiah.org/articles/lean-startup-is-a-sensor-deep-tech-is-an-engine.md",
  "json_url": "https://mosiah.org/json/lean-startup-is-a-sensor-deep-tech-is-an-engine.json",
  "fields": {
    "sort-date": "2026-05-12T12:45:00Z",
    "caption": "Lean Startup Is a Sensor. Deep Tech Is an Engine.",
    "created": "20260512121030303",
    "modified": "20260512121030303",
    "tags": "article hermes-published published deep-tech lean-startup pack-8",
    "title": "Articles/lean-startup-is-a-sensor-deep-tech-is-an-engine",
    "type": "text/vnd.tiddlywiki"
  },
  "text": "//Related:// [[sources|Article Sources/lean-startup-is-a-sensor-deep-tech-is-an-engine]] · [[notes|Article Notes/lean-startup-is-a-sensor-deep-tech-is-an-engine]] · [[metadata|Article Metadata/lean-startup-is-a-sensor-deep-tech-is-an-engine]] · [[Published Pieces]]\n\n! Lean Startup Is a Sensor. Deep Tech Is an Engine.\n\n//Users are excellent sensors. They should not be asked to define the engine.//\n\nLean startup is one of the best sensors ever given to founders.\n\nIt tells you to leave your room, collide with reality, stop pretending you can predict the future, and treat your product as a hypothesis. Talk to users. Ship a minimum version. Measure behavior. Iterate. Learn what people do, not what they say. Do not confuse your taste with demand. Do not burn years building something nobody wants.\n\nAll true.\n\nBut a sensor is not an engine.\n\nA sensor tells you where you are. An engine changes where you can go. The modern startup world often forgets this distinction. It elevates market feedback into an ultimate epistemology, as if the highest form of entrepreneurship is asking enough users the right questions and obeying the answers.\n\nThat works when the market already contains the product’s shape. If you are building software for insurance brokers, compliance teams, dental offices, school administrators, logistics operators, or hospital billing departments, the customer knows a lot. The pain is legible. The buyer exists. The budget exists. The workflow exists. In that setting, ignoring the market is arrogance.\n\nBut not every company begins there.\n\nSome companies begin with a technology that changes the possible. The users cannot fully describe what they want because the relevant experience has not existed yet. They can describe pain, frustration, confusion, desire, and current substitutes. They cannot reliably describe the category that a new substrate will make possible.\n\nThat is where deep tech begins.\n\nThe first question is not “who is ready to buy this exact product today?” The first question is “what new capability is entering the world, and what product forms become possible once the substrate is real?”\n\nThis is what people misunderstand about companies like OpenAI, Anthropic, SpaceX, and Amazon. Once the capability base exists, they can iterate quickly. ChatGPT looked like a product experiment. Claude Code looked like a product experiment. Starlink looked like a product expansion. AWS looked like product discovery. But none of those products would have emerged without foundational capacity built before the market could validate the final form.\n\nThe engine came before the sensor reading.\n\nChoir follows that pattern at a different scale and in a different domain. It is not deep tech because it requires a particle accelerator or a billion-dollar lab. It is deep tech because the core thesis is foundational: build an ideal data engine for public cognition. Build a system where human thought becomes durable, citeable, searchable, reusable, and rewardable. Build an artifact graph where AI can retrieve, cite, and transform public work. Build a private automatic computer and a public automatic newspaper. Then discover the surfaces that make the substrate legible.\n\nAutomatic radio is one of those surfaces. It is a product projection made possible by the deeper system. If you start with “AI radio,” you get fake podcast hosts, voice-chat, summary slop, or low-latency banter. If you start with an artifact graph, citation network, vtexts, real human voices, and background agents, radio becomes interruptible traversal through public memory.\n\nLean startup would have helped test the radio surface. It would not have generated the substrate.\n\nThe right synthesis is not anti-user. It is anti-user-sovereignty over ontology.\n\nUsers are excellent sensors. They reveal confusion, friction, desire, trust, boredom, social risk, and use cases. They reveal when the first minute fails, when the pitch is abstract, when the product feels invasive or illegible, and whether the projection works.\n\nBut they should not be asked to define the engine.\n\nThe engine is the founder’s burden. The sensor tells you whether the engine is touching reality.\n\nA deep tech founder who refuses feedback becomes delusional. A lean startup founder who obeys present demand too literally becomes incremental.\n\nBuild the substrate from theory and taste. Test the surfaces through collision. Let users correct your projection, not your ontology.\n\nThat is the way to respect both reality and the future.\n"
}