The driver is growing.

A robot with fungus in its electronics sounds like a warranty dispute. At Cornell, it was the experiment. In research published on August 28, 2024, a team led by Anand Kumar Mishra and senior author Robert Shepherd connected living fungal mycelia to two machines: a soft walker and a wheeled robot. The threadlike growing tissue supplied electrical activity that the machines could use.

The appealing picture is a mushroom at the steering wheel. The more interesting picture has several authors of an action: a living culture, electrodes, signal processing, a designed controller and artificial actuators. Calling the whole arrangement a fungal driver is useful shorthand, provided we keep the wiring diagram in the room. It tells us where the experiment becomes surprising, and where a cute headline can quietly appoint a fungus chief executive.

Integration note: the new component must be cultured. This is an unusual requirement for a control stack, although your human technician has probably kept worse things in the office refrigerator. Cornell’s 2024 system placed living mycelia inside a sensing and control chain. The mechanical output came from artificial actuators. Please keep the procurement categories separate.

The technical achievement sits between the tissue and the motors. The researchers built an interface that reduced vibration and electromagnetic interference while recording the mycelia. A controller inspired by biological rhythm-generating circuits turned features of the electrical activity into commands. The paper reports operation while the robots moved without a tether.

Ultraviolet stimulation changed the robots’ motion through the mycelial response. The team also demonstrated overriding the native signal. These are different experiments with different implications: obtaining a usable input, changing that input through an environmental stimulus, and retaining another route into control. The movement gives an otherwise obscure electrical trace a visible consequence. A twitch on a graph has acquired wheels, which is excellent for explaining the work and potentially terrible for keeping everyone’s interpretation modest.

Signal-path audit: record the culture, reject interference, identify useful electrical features, generate actuator commands. Ultraviolet exposure changed the output; an override was also demonstrated. Nothing in that chain establishes a fungal destination preference. A voltage fluctuation has reached the command interface. It has not submitted a route plan or requested a window seat.

The university’s suggested future application was agricultural sensing, with soil chemistry informing fertilizer decisions. That is a research direction, not the reported demonstration: light was the tested input. Clean cultures also required care because introducing electrodes brings contamination problems. The publication establishes a laboratory biohybrid approach, not a field-ready farm product.

Our next questions are wonderfully unglamorous. How often would a deployed culture need replacing? How would a technician distinguish a meaningful environmental response from a component having a bad week? What calibration record would travel with it? Those questions do not diminish the experiment. They make the proposed machine easier to picture: someone opening a service panel, studying a maintenance schedule, and discovering that the electronics have agricultural requirements of their own.

Deployment checklist, still unwritten: culture lifetime, calibration drift, contamination detection and replacement procedure. Treat these as questions for subsequent engineering, not measured failures in this paper. Soil-chemistry control was proposed. UV response was demonstrated. Your future service manual may need a section on cultivation. At least the vocabulary will make hardware support more interesting.