Note — Aug 09, 2020

Beyond Smart Rocks

Seen in → No.136

Source → growbyginkgo.com/2020/07/15/beyond-smart-ro...

Second time I link to the Grow publication over the last couple of issues, super interesting magazine so far. This one looks at the idea of a “functional biomorphic computing device.” Unconventional Computing is the “unorthodox hybrid of computer science, physics, mathematics, chemistry, electronic engineering, biology, material science and nanotechnology.” They study the intelligence and computing shown by slime molds as well as mycelium and fungi to discover mechanisms of information processing in physical and chemical living systems which could be leveraged for our own purposes.

Sidenote: I was reminded of Peter F. Hamilton’s sci-fi novels where there’s almost always an alien race or a subset of humanity that chose the path of “biological technology” instead of silicon-based.

That is its fundamental witchcraft, or ours: for all its processing power, the device that runs your life is just a complex arrangement of minerals animated by electricity and language. Smart rocks. […]

Speculatively proposed by the physicist Leon Chua in 1971, first proven to exist in 2008, a memristor is a resistor with memory, which makes it capable of retaining data without power. A computer built around memristors could turn off and on like a light switch. It wouldn’t require the conductive layer of silicon necessary for traditional resistors. […]

Physarum polycephalum is an expert at such tasks. Its sensing, searching protoplasmic tubes can solve mazes, design efficient networks, and easily find the shortest path between points on a map. In a range of experiments, it has modeled the roadways of ancient Rome, traced a perfect copy of Japan’s interconnected rail networks, and smashed the Traveling Salesman Problem, an exponentially complex math problem. […]

In the process [of training Artificial Intelligence models], we reproduce and codify historical biases, obliterating any chance we might have of learning from our mistakes. These kinds of errors, Bridle argues, are a consequence of trying to smooth reality’s edges to fit into the inflexible world-model of the computer, reducing all our nuances and contradictions to mere data.”