Bio-Hybridism

From the AI Research Compass wiki

Bio-Hybridism: Living neurons grown on chips can be trained to compute, offering an energy-efficient substrate that silicon cannot imitate.

Beliefs

Bio-Hybridism thinks the most promising neural network is the one that is actually made of neurons. Cultured brain cells grown on multi-electrode arrays spontaneously form networks, adapt their activity to stimulation and can be closed into a loop with a simulated environment. Instead of emulating biological learning on power-hungry silicon, it proposes using the real thing, wetware that rewires itself, runs on nutrients and learns continually by its nature.

It sits on the World side, since its intelligence is literally physical tissue embedded in a sensory loop, and in the upper half toward Monolith, because the culture itself is the computing substrate, one living network rather than a pipeline of parts, even though it needs electrodes, encoders, decoders and life support to run. It wears the beyond-revolutionary shade, the most extreme on the chart: it would replace not the architecture but the material.

Arguments

  • Cultured neurons connected to a simulated game environment through electrical stimulation and recording have shown improved performance over minutes, a sign of adaptive learning in a dish.
  • Biological neurons are vastly more energy-efficient than digital hardware per unit of computation, a growing concern as AI’s energy demands rise.
  • Wetware learns continually and locally by default, sidestepping catastrophic forgetting and the train-then-freeze cycle.

Critiques

  • Demonstrated capabilities are rudimentary, and it is unclear how to scale or program cultures toward complex tasks.
  • Cultures are short-lived, variable between batches and hard to reproduce, which is the opposite of what engineering needs.
  • Claims of “sentience” or “intelligence” in dishes are often overstated, and they raise ethical questions the field has barely started to address.

Representative works

Relationships

How to get it

Land on this cell in the full or pro test: x ≈ +0.69, y ≈ +0.46, paradigm ≈ +1.00.

A first draft. Corrections and better arguments are welcome through a pull request. Edit this entry on GitHub Edit its data