Harvard Extension Student Psychological Club
2 HN recommenders · recommended in 3 months, 2023–2024
1:17:16 · 24K views · May 18, 2023
Developmental and synthetic biologist Michael Levin speaks at a Harvard Extension student event on bioelectricity, body intelligence and regenerative medicine. He discusses how cell electrical networks guide growth and form.
Recommended as an approachable look at Levin's work on regrowth and collective cellular intelligence, and used as an example of intelligence beyond brains.
4 comments from Hacker News
Whenever morphogenesis comes up, especially with the usual "work from DNA and proteins up" approach, I feel obliged to mention Michael Levin. He also researches morphogenesis, except his approach is on a higher abstraction level. He manipulates the electrical network that cells use to communicate with each other to organize their work towards multicellular goals. Do watch his lectures, it's some really fascinating stuff. You don't need to touch genes at all to regrow limbs and organs. Neither do you need to operate on that low level to cure cancer. This whole "let's effect organism-level changes by observing and manipulating DNA and proteins" feels to me like "let's add a feature to a react web app by observing and modifying the behavior of individual transistors in the CPU". Doable given enough time, but you'd drown in complexity in the process. this video
The article makes no mention of Michael Levin [1], one of the main researchers/labs, around bioelectricity/xenobots. A recent talk of his [2]. Secondly, bioelectricity has nothing or very little to do with electric currents as the one from the socket. There is no "brain zapping" involved or required. In the paper [3] detailing an adult frog limb amputation and full regeneration over 18 months, Michael Levin's lab used a single bioreactor with silk hydrogels applied only for 24 hours (see BioDome fabrication in the paper). [1] [LINK] [2] "From Matter to Mind: Bioelectricity, Body Intelligence, and the Future of Regenerative Medicine", this video At minute 33, this video : "We don't use any applied fields, no frequencies, no waves, no electromagnetic radiation, no magnets, what we do is we control the native interface by which cells control each other, that is the set of electrical controls that they expose to each other, and you can do this with drugs, with ion channel mutants, with optogenetics and various others techniques." [3] "Acute multidrug delivery via a wearable bioreactor facilitates long-term limb regeneration and functional recovery in adult Xenopus laevis", [LINK]
I think most AI research up to this day is a dead end. Assuming that intelligence is a problem solvable by computers implies that intelligence is a computable function. Nobody up to this day has been able to give a formal mathematical definition of intelligence, let alone a proof that it can be reduced to a computable function. So why assume that computer science is the key to solving a problem that cannot even be defined in terms of math? We had formal definitions of computers decades before they became a reality, but somehow cannot make progress in formally defining intelligence. I do think artificial intelligence can be achieved by making artificial intelligence a multidiscipline endeavor with biological engineering at its core, not computer science. See the work of Michael Levin to see real intelligence in action: this video
It gets wilder. In a 3D printer, the commands come from a computer and the current state is maintained in its memory as well. In a multicellular biological system, that state is stored within the thing itself that is being built . And the commands come from it self-organizing. this video