https://www.youtube.com/watch?v=9pG6V4SagZE

[These are incomplete notes]

transcript

Transcript - Toward AI-Driven Discovery of Electroceuticals

Primary topics:

• Individual cells are highly capable, including sophisticated computation that lets them navigate and operate on their environment (despite having no brain)
• Cells coordinate in organs to solve complex problems
• The coordination takes place via electrical signaling
• The electric signaling in brain synapses uses those same ion junctions, but all cells have them. Computation in the brain uses repurposed mechanisms for computation that is ubiquitious at many scales within the body
• The high-level computational goal (ex. build an eye) is encoded in electrical patterns
• Information is persisted outside the brain: once body type is set (or reset), planarians can be diced up and will create the new body type
• Some caterpillar memories survive the deconstruction of its brain during metamorphosis
• Cells building an organ have a high-level concept of what they are building and can choose different routes to build it (ex. kidney tubules: standard cells, large cells (use fewer), single cells (decide to wrap their cytoskeleton into a tubule)
• The electrical pattern of latent possibility may be the precursor to planning (considering hypotheticals) in high-level cognition
• Pertubing the electrical network allows exploring the morphological state space (ex. creating other possible types of heads)
• Cancer is cells disconnecting from the collective and pursuing individual goals (narrowing scope of self)
• Cells working in collaboration have greater memory and greater ability to compute
• Dr. Levin has developed a model for directly comparing highly diverse intelligences
  • Defines scale by the scale of its maximum goal
• Living things hack other living things all the time
  • Cells w/ other cells (to form and maintain a collective)
  • Wasps hacking plant cells to convince them to make a gall vs. leaf
• Self = boundary of goal