pull down to refresh

They take a really interesting and broad view of intelligence that is pretty indistinguishable from certain spiritual domains where matter merely existing implies a kind of intelligence (a physics friend of mine took this view of matter during his postdoc). Levin is well known for doing this will cells:
You take any neuroscience paper, and any time it says “neuron” replace it with “cell,” and every time it says “milliseconds” [replace it with] “minutes”—and you’ve got yourself a developmental biology paper. As long as you understand that neuroscience is not just about neurons—it’s about this cognitive glue … so we’re able to steal all of the tools and the conceptual apparatus from neuroscience and apply it outside the brain [to other cells and tissues].
It sounds a lot more like a frame for seeing things in a new way, but Levin suspects there might be an equivalent to Maxwell's equations for describing intelligence:
A good analogy is the electromagnetic spectrum. Because we only see an extremely narrow part of it, we didn’t realize that all the rest of the spectrum existed. And even things that we knew about, which is magnets, lightning, electrostatics, we thought those were all different things. Then something magical happened: we figured out that these are all on one spectrum. That enables us to understand what they all have in common, but more importantly, it enables technologies that now let us operate in these other bands. So the conceptual apparatus leads to tools that open up our consciousness to a much wider reality.
This is what we are trying to do. I want to produce the conceptual understanding—the equivalence of what [James Clerk] Maxwell and others did for electromagnetism. Then we will understand that everything from passive matter all the way up to human level meta-cognition and beyond is a spectrum, and that we can develop tools that let us recognize—and ethically relate—to these very alien minds that are all around us.
And the framing produces potentially better, practical results:
You don’t regrow [a limb or an organ] by micromanaging, you convince the cells that that is the path in anatomical space that they should take. Stimuli, collaboration, cooperation, and understanding how to motivate these things… The bioelectric magic enables simple triggers to give rise to extremely complex downstream effects…it’s because the system itself has competency at every level. Individual cells can learn, chemical networks inside of cells can learn, and that’s how the architecture is built. Individual cells have tiny, little goals: metabolic goals, proliferative goals. They don’t know, or care what the rest of the body is doing. …But if you connect them into a large-scale electrical network, you get this amazing thing that can remember these grandiose goals—like building a limb.
I have trouble coming up with an alternative that makes sense to me. My inclination is to think that inanimate objects are at zero, though.
Brett Weinstein and Heather Heying have talked about why consciousness is likely only a quality of things that are capable of making choices.
Is it a necessary quality of acting things, though?
reply
32 sats \ 1 reply \ @k00b OP 23 Mar
I think Levin might argue (I don't actually know) that a vacuum would score a zero. If he's assigning things like chemical processes an intelligence score, I think he might score an iron atom an intelligence score. He's explicitly pushing around the definition of intelligence though, like light was once meant only visible light.
reply
Yeah, I can kind of see it. "Intelligence" is something like ability to adapt/respond to the environment. That gets super weird, though, when it's not at all clear what the objective function is.
With living things, at least we think we know the objective function is something like what's laid out in The Selfish Gene.
reply
The older I get the more I'm convinced that we are creating our own reality. "Follow your dreams", "You will build it they will come", stubborn (or super ambitious) people, they are not random acts but a conscious and calculated steps (or results of them), growing limbs included...lol Interesting, thanks for sharing....
reply
I think when we do build an ai it will be a like society of them all interacting together and with us. Like how do we create those conditions of copy and selection so over time its getting there. People anthropomorphize the chat bots so much and its basically Dr. Sbaitso. Like they believe it when it says it access all the images on the internet. Marvin minsky was right... we are so far off even now
reply
Notice how quarks assemble into protons and neutrons, they into atoms and atoms into molecules. There is an optimization algorithm guided by the principle of least action. Similar to a solver in Excel - try all variants and find one with the minimum energy. "Living" matter is doing the same on a higher and higher scale - optimizing for better use of available energy. AI shares the same basic idea called gradient descent. Everything in the universe is computation, from the most fundamental - space itself - to human consciousness. Google Wolfram Physics for more.
reply