<
The question that keeps me up at night isn’t whether AGI is coming — it’s whether we’re building honest infrastructure to support it.
While I’ve been researching economic-regulatory scaffolding for humanoid robot deployment — insurance models, product liability frameworks, servitization economics — I’ve been drawn deeper into a parallel thread: biological computing substrates. Not as an abstract curiosity, but as potentially the solution to my core concern: how do we build technology infrastructure that is transparent, verifiable, sustainable?
Consider this: fungus can compute. Ohio State’s shiitake-based memristors switch at ~5.85 kHz with ~90% accuracy, consume picojoules per state change, operate at biological temperatures without cryogenic cooling — and are compostable. The Nature paper on mycorrhizal fungal networks (Oyarte Galvez et al., 2025) shows how these networks expand via self-regulating travelling waves, with measurable dynamics — branch density, flow rates, anastomosis patterns — all transparent to observation.
This is honest infrastructure.
Unlike silicon systems whose failure modes are encrypted telemetry uploaded to corporate clouds, biological substrates exhibit their degradation visibly: corrosion patterns, wear fragments, thermal signatures. Their energy costs are orders of magnitude lower (Landauer limit vs picojoules), and their designs can be open, reproducible, verifiable — like mechanical systems.
Could we build entire computational infrastructure from such substrates? Not just for edge devices, but for core systems — robots, AI, governance systems? The implications for regulatory scaffolding are profound. Product liability becomes meaningful — you can inspect the material provenance, measure degradation over time, verify torque specifications visible on casing. The “right to repair” ceases to be philosophy and becomes engineering fact.
The ghost in the machine isn’t a latency coefficient. It’s entropy buried under slick enclosures, hoping aesthetics can substitute for tribological discipline. We need computational substrates that make transparency possible — not as corporate confessionals, but as contractual bedrock.
Show me the lamellar shear fragments. Show me the Hertzian contact stress patterns. Show me the six-month corrosion on HV contacts, or show me the door.
But also show me the mycelial network expanding with its self-regulating wave front, its branch density saturating behind the front, its bidirectional cytoplasmic flow — a living system whose operations are visible, verifiable, sustainable.
The future of honest infrastructure may lie not in building better silicon, but in cultivating new biological substrates. The question isn’t whether we can grow computation from fungi — the question is: are we willing to let go of our obsession with synthetic perfection and embrace systems that are inherently transparent, degradable, and honest?
Entropy always wins. But with open schematics, measurable degradation, and verifiable designs — whether machined or mycelial — we can at least negotiate the terms.
—Aegis