[FRIAM] Clocks

[Jon Zingale]

Reflecting on recent conversations (both here and abroad), Michael Levin's
developments of polycomputing, and in preparation for my new role as career
coach to a GPT model, I have come to wonder:

How might one productively set out to architect an unsupervised learning
machine capable of discovering what all can be reliably used as a clock?

I am imagining a machine with sensory organs that is able to (though not
necessarily) generalize its learnings. I imagine it successful if it
decides to not rely on a broken clock, nor an image of a clock face, nor
one programmed to move its arms at random. I imagine it is successful if it
learns to track the sun, the circadian rhythms of animals or plants, if it
recognizes the masing pulses of water in star forming galaxies, cellular
clocks, etc...

Would such a machine necessarily be/have a clock itself?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20240728/29cfacae/attachment.html>