Technoscientific Disruption: Multilevel, Multimodal

Two distinct waves of technoscience are reshaping medicine, and conflating them obscures both.

The first wave is endogenous, made up of technosciences developed within biomedicine, for biomedical purposes, and already recognizable to clinicians as belonging to them. It is the wave Chapter II described: the new biologies, multiscale and multimodal imaging, genome and epigenome editing, organoids and patient-derived systems, multi-omic measurement, and AI turned on biological discovery. Its vocabulary is taught, and the profession has institutions organized to absorb it; it is genuinely disruptive, but it is ours.

The second wave is exogenous: broader, less domesticated, and developing largely outside medicine, yet certain to reshape it over the coming decades. Quantum computing and quantum sensing. Advanced materials and metamaterials. Autonomous robotics and swarm systems. Neuromorphic and edge computing. Photonics. Energy-based ablative technologies. Additive and four-dimensional manufacturing. Ambient sensing and the instrumented environment. Cryptographic and privacy-preserving computation.

That list is easy to underrate, because each item sounds like somebody else's field.

• Quantum sensing. Optically pumped magnetometers and nitrogen-vacancy diamond sensors detect biomagnetic fields without cryogenics, wearable magnetoencephalography, and magnetic imaging approaching cellular scale, in instruments that could sit in a clinic and not a shielded basement.
• Metamaterials and photonics. Acoustic metasurfaces that correct for the distortion imposed by the skull, making transcranial focused ultrasound routine; flat metalens optics that shrink an endoscope's front end; Raman and photoacoustic sensing for label-free characterization at a surgical margin.
• Neuromorphic and edge computing. Inference at milliwatt power budgets, which is what allows a closed-loop implant to decide locally instead of round-tripping to a server.
• Cryptographic and privacy-preserving computation. Federated learning and homomorphic encryption, which let models train across institutions without records leaving them: the technical answer to the data-fragmentation problem of Chapter III.
• Ambient sensing. Passive measurement of gait, sleep, respiration, and speech from the built environment, turning the home into a diagnostic surface with no device to remember to wear.

None of these was developed for medicine. Each will arrive anyway, through supply chains and cost curves set by industries that do not consult clinicians. Technoscientific disruption transcends the biomedical, and a profession that scans only its own literature will meet it late, and as a procurement decision.

Meeting it earlier requires an institution rather than a resolution. What is missing is a standing horizon-scanning function, held jointly across the specialty societies instead of duplicated inside each of them. It would be staffed by clinicians who spend part of their time inside non-medical technical communities, and it would report on a fixed cadence, not when a technology has already arrived as a purchasing decision.

Its output would not be a literature review. It would be a short list of technologies that have crossed a cost or maturity threshold somewhere else, each paired with an assessment of which clinical levels it could enter and what would have to become true for it to do so. No single society can carry that cost, and none needs to carry it alone, which is one instance of the federation problem Chapter XII takes up.

Disruption from these sources is multilevel, operating simultaneously at the molecular, cellular, organ, individual, institutional, and population levels, so that an advance at one level propagates unpredictably to the others. It is also multimodal, arriving through diagnosis, therapy, monitoring, logistics, communication, and discovery at once, through no single channel a specialty could guard.

The two dimensions multiply; they do not add. A single advance enters at one level through one mode, a quantum magnetometer entering at the organ level through diagnosis, say, and then propagates along both axes at once: downward to cellular measurement, upward to population screening, sideways into monitoring, therapy guidance, and discovery. The exposure surface of any specialty is the product of the two dimensions and not their sum, and most cells of that product are unoccupied and unwatched.

Two implications follow for biomedicine. The correct unit of surveillance is the technology, not the disease. Attention organized by organ or diagnosis guarantees that a development entering three specialties at once is noticed in time by none. And propagation is unpredictable in detail even when certain in aggregate, since which cell a technology lands in first is decided by cost and regulation as much as by need.

One implication follows for the physician as explorer, in the sense Chapter VII develops. The richest territory is the unoccupied cell, where a technology mature elsewhere meets a clinical level nobody has connected it to. That is a search problem, and the clinician is unusually well placed to run it, because the limiting factor is recognizing which unglamorous problem a technology solves, not building the instrument. Most of the second wave's clinical yield will come from people who noticed a match, not from people who invented a device.