It seems to me that a big part of the problem is the dichotomy, something that biologists have been wrestling with on the definition of life vs non-life for a long time. The desire for a clean simple boundary that allows us to conveniently discount everything on the other side of it sharpens these academic debates beyond any available evidence.
A spectrum of sentience makes a lot more sense to me personally, especially since as a science fiction author I feel no need to place humans at the high end of that distribution. I can imagine aliens or future humans who are a lot smarter and more self-aware than we are.
My last two columns for IGMS were about extending this general framework to machines.
Randall, I agree that the problem isn’t where to draw the line, but the assumption that there is a line. What we call “sentience” looks less like a property on a spectrum and more like a process: action beginning to recursively condition itself over time. The moral questions follow from that self-binding, not from intelligence or resemblance to us.
Oh, yeah, even a spectrum metaphor collapses some multidimensional complex system of loops into one line on a graph. I tried to get into that with states of consciousness in the first of those pieces. But in some sense we gotta start from where we are, culturally.
Thanks for sharing! I find the idea of a spectrum attractive for its elegance, while also having some sympathy for those who worry that it means positing sentience much more widely than common sense suggests
Yes, I think Cabanac's work is a great contribution, even if it turns out that things are more complex than on the most ambitious version of his theory
Brillaint piece on neuron-scale sentience markers. That 5-neuron circuit in C. elegans handling motivational tradeoffs really upends Cabanac's theory that valenced experiance evolved specifically for weighing drives. When I was doing undergrad neurosci work, we always assumed complex decisions meant complex processing, but turns out even unconscious reflexes can look pretty smart.
Thanks! I guess I still value the ingenuity of Cabanac's proposal, and suspect it points at something important about the function of sentience, even though in its most ambitious form it does seem to be undermined nematodes.
What a fascinating take on earthworms and their incredible role in nature! It's intriguing how, much like worms transform soil, AI is transforming how we explore complex concepts in science—like uncovering the hidden possibilities in physics. If you're curious, check out my thoughts on how machines are learning to identify the seemingly impossible in physics: https://00meai.substack.com/p/machines-learn-to-spot-the-impossible.
It seems to me that a big part of the problem is the dichotomy, something that biologists have been wrestling with on the definition of life vs non-life for a long time. The desire for a clean simple boundary that allows us to conveniently discount everything on the other side of it sharpens these academic debates beyond any available evidence.
A spectrum of sentience makes a lot more sense to me personally, especially since as a science fiction author I feel no need to place humans at the high end of that distribution. I can imagine aliens or future humans who are a lot smarter and more self-aware than we are.
My last two columns for IGMS were about extending this general framework to machines.
http://www.intergalacticmedicineshow.com/cgi-bin/mag.cgi?do=columns&vol=randall_hayes&article=039
http://www.intergalacticmedicineshow.com/cgi-bin/mag.cgi?do=columns&vol=randall_hayes&article=040
Randall, I agree that the problem isn’t where to draw the line, but the assumption that there is a line. What we call “sentience” looks less like a property on a spectrum and more like a process: action beginning to recursively condition itself over time. The moral questions follow from that self-binding, not from intelligence or resemblance to us.
Oh, yeah, even a spectrum metaphor collapses some multidimensional complex system of loops into one line on a graph. I tried to get into that with states of consciousness in the first of those pieces. But in some sense we gotta start from where we are, culturally.
Thanks for sharing! I find the idea of a spectrum attractive for its elegance, while also having some sympathy for those who worry that it means positing sentience much more widely than common sense suggests
According to Notes, this was your piece? Congratulations. Very well written, I thought.
Thank you!
Cabanac correctly identified motivational trade-offs as one outward signature of sentience, not its generative source.
Cabanac correctly identifies where sentience becomes visible.
He misidentifies what generates it.
Yes, I think Cabanac's work is a great contribution, even if it turns out that things are more complex than on the most ambitious version of his theory
Brillaint piece on neuron-scale sentience markers. That 5-neuron circuit in C. elegans handling motivational tradeoffs really upends Cabanac's theory that valenced experiance evolved specifically for weighing drives. When I was doing undergrad neurosci work, we always assumed complex decisions meant complex processing, but turns out even unconscious reflexes can look pretty smart.
Thanks! I guess I still value the ingenuity of Cabanac's proposal, and suspect it points at something important about the function of sentience, even though in its most ambitious form it does seem to be undermined nematodes.
What a fascinating take on earthworms and their incredible role in nature! It's intriguing how, much like worms transform soil, AI is transforming how we explore complex concepts in science—like uncovering the hidden possibilities in physics. If you're curious, check out my thoughts on how machines are learning to identify the seemingly impossible in physics: https://00meai.substack.com/p/machines-learn-to-spot-the-impossible.