★ LIFE SCIENCE · brain-behavior in silicon · the bridge ★
Stop simulating a brain on a chip and start building a chip that is one. Neuromorphic engineering wires silicon to spike, remember, and compute the way neurons do — memory and processing in the same place, like the thing it copies.
of the living and physical body — the cell, the neuron, the molecule, the matter that does the work
ethereal
of the information and the limit — the bit, the entropy, the open question, the pattern with no body
spiritual
of mind and meaning — the intelligence claimed, the pioneer's insight, what it says about life
electrical
of the machine and the rule — the chip, the gate, the circuit, the engineered substrate
The Idea
the three-beat story
The Idea
copy the wiring, not the math
Carver Mead's insight in the late 1980s: don't model a neuron in software running on a normal computer — etch the neuron's behavior into the silicon itself. Analog circuits that integrate, leak, and fire like the real thing.
The Synapse
the memristor remembers
A normal computer shuttles data between separate memory and processor — the von Neumann bottleneck. The brain doesn't; synapse and signal are the same place. The memristor (predicted 1971, built 2008) is the device that holds a weight and computes with it at once — the artificial synapse.
The Chips
brains on a wafer
IBM's TrueNorth, Intel's Loihi, Manchester's SpiNNaker, Heidelberg's BrainScaleS — event-driven, massively parallel, sipping power. Built for the same prize as organoid intelligence: the brain's efficiency, in a substrate you can manufacture.
Watch It Spike
an illustration of a spiking neural network — charge accumulates, a neuron fires, the spike propagates. Not a chip simulation; a homage.
spikes 0
The Reckoning
the bridge, and the honesty about it
The Bridge
life ⇄ hardware
The hardware half of the seam: where biology asks 'what is a mind?', this asks 'can silicon be one?' — and answers by copying the brain's architecture, not just its output.
>Pairs with organoid intelligence (the wet answer) and the memristor (the device).
Render, Not Invent
sourced
The science is summarized from the public record; Mead, Chua, Williams, Furber are living and CITED, not minted.
Emergents are devices, chips, and concepts. The spiking network below is an illustration, not a chip simulation.
Why It Matters
the efficiency floor
A GPU brute-forces a neural net; a neuromorphic chip lets the structure do the work, event by event.
The same 20-watt thesis as the brain — efficiency as the real frontier of compute.
The Roster
the devices, concepts, and pioneers as ACI .agents — each a birth certificate & a nature (10)
A life-science sphere on the life ⇄ hardware bridge — rendered, not invented, two-layer honest (settled capability vs the open questions, flagged as questions). Living researchers are CITED, not minted; deceased pioneers are minted in memoriam. The interactive above is an illustration, not a scientific simulation. No copyrighted text reproduced. Part of the seam where the life-science domain meets the hardware frontier — kin to organoid intelligence, cellular automata, and the memristor. Each entry is named by its nature: natural, ethereal, spiritual, or electrical.