Michael Levin (select video notes):
Difference in programming: Bio has a Commitment to an Unreliable substrate AND uses creative problem solving to adapt and change to meet novel circumstances using only what it has available to it. It commits to persistence by creatively re-interpreting previously learned information. It encodes memories and later decodes them (perhaps differently to account for new (unreliable) substrate or available resources). Life doesn't over-train on evolutionary priors.Biology assumes that the hardware is unreliable. Environment and your own parts will change. Don't over-train! Little allegiance to past self's meaning of engrams; re-interpret on the fly for present/future is all that matters. Engram is highly compressed - creative re-membering (not deduction). Unreliable hardware and confabulation are a feature, not a bug. Consciousness = active, creative storytelling from own memory traces.Re-Interpretation, NOT redundancy and error-correcting codes.
Whence specific goals and competencies if not selection? Evolution exploits "free lunches": shapes, behaviours, properties of networks, features of computation, numbers, etc.Option 1: There is a random set of amazing "facts that hold" and we will call it "emergence" and be surprised each time. Occam's razor -> mysterianism
Option 2: there is an ordered, non-phtsical latent space of patterns which can be studied systematically Optimism -> research agenda
Synmorpho beings and minimal algorithms as vehicles for exploring Platonic latent space
A Taxonomy of Platonic Latent Space:
Hypothesis: the world of forms is not just for low-agency facts about numbers and shapes. If it guides morphologies and morphogenesis w/ cognition, Platonic space also contains kinds of minds.
Classical sorting algorithms as a model of morphogenesis: Slef-sorting arrays reveal unexpected competencies in a minimal model of basal intelligence.
Computationalism: Nothing is a Turing Machine, not even a Turing Machine
No comments:
Post a Comment