Jonathan Birch
speaker
151 appearances
1 recordings
1 series
first heard Oct 2024
last heard Oct 2024
Jonathan Birch’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsNo recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.
Appearances
Yeah, I talk in the book about the open worm project, which I think is still going.
Yeah, where the aim was to emulate the nervous system of C. elegans in computer software, see if you can put the emulation in charge of a robot, see if it behaves like C. elegans. I suppose we've learned something from this, which is how difficult the task is.
There's a lot of stuff going on at the within neuron level in C. elegans that even knowing the entire connectome does not tell you very much about. So even that is a very, very hard challenge.
But to me, it's a good way into this topic of artificial sentience because you can easily entertain in imagination the idea that this project had succeeded very quickly and then moved on to open Drosophila, open mouse. Once you have open mouse, I think you have a sentience candidate. If you've completely recreated in computer software everything the brain of a mouse does...
Yeah, that's right. There's a lot we don't know from the connectome. One thing you can't read off from the connectome is the weights of the connections, which is hugely important, or how those weights are changed by learning. But also, even if you had all of that, what happens within the neurons is also important.
There are within-neuron computations that are really crucial to steering behavior, for example. And so you wouldn't expect to get the steering behavior in a emulation unless you'd actually emulated the individual compartments within the neurons and how they're arranged in space.
Well, I think they've been, they've been trying. Yeah. Um, I'd be in favor of this sort of work receiving more funding than it does. Cause it's to me that there's risks, there's risks of creating artificial sentience candidates, but there's huge opportunities as well, because you've got the potential to create a system that could replace a lot of animal research.
Cause you could be doing research on the, the emulation where you can actually intervene at a really precise level, without injuring or hurting. And you could be doing that instead of lesioning living animals. So I'd like to see much more of this, and I think it's been largely funding limited, I think, so far.
Well, the octopus has about 500 million neurons, so I don't know how that translates into synaptic connections. A lot. It's going to be quite a lot, yeah. Crabs' brains are much, much smaller, and it varies a great deal by species, but not dissimilar to insects in terms of the number of neurons. With bees, you have about a million neurons, Drosophila, about 100,000. Okay.
Yeah, yeah, indeed, yeah.
Yeah, these are very hard cases. I suppose when I started writing the book around 2020, not sure the large language models were even on my radar at all. And then they've jumped onto everybody's radar through things like ChatGPT. And I suppose I've been on a journey like everyone else during that time.
I initially thought, well, these are next token predictors and the sector has been moving away from brain-like forms of organization. So it's been taking out things like recurrent processing that on many theories of consciousness are absolutely essential, but transformers take that out. So I thought, well, here is something that is conspicuously unlikely to be sentient.
But then I suppose I'm not sure that's the correct view anymore, I suppose, because I've been quite astonished by the feats of reasoning they seem to perform today. where it's, well, it's reasonably evident that we do not understand how they work. They're incredibly opaque to us. We don't know how they do what they do.
And there seems to be some element of acquiring algorithms during training that were never explicitly programmed into them. So in a way that architecture that was programmed into them, the transformer architecture, no reason at all to think that would be capable of sentience.
But when you have these very, very large models where they've acquired algorithms during training, we don't know how and we don't know what they are. We don't know the upper limit on what algorithms they might acquire. And we don't know what algorithms are sufficient or not for sentience. And so we're not really in a position to be so sure anymore. that they couldn't acquire those algorithms.
So, for example, if you think a global workspace is what it takes to have sentience, as many have suggested, we don't know that they couldn't acquire a global workspace.
Well, this is Stan De Haan's theory. His book Consciousness and the Brain is a nice exposition of it. But it's this quite popular idea that consciousness has to do with a network that puts the whole brain on the same page, as it were, by taking inputs from many, many different sensory sources and integrating them into something coherent and then broadcasting that content back
to the input systems and onwards to other systems of motor planning, reasoning, etc. So it's the the bit where you know, the central coming together of everything in the brain. And well, of course, that is designed as a theory of consciousness in the human brain. But the basic architecture
where you have lots and lots of input processes competing for access to this workspace, where once a representation gets in, the integrated content will then be broadcast back and onwards. There's nothing about that architecture that is inherently difficult to achieve computationally. And so we did a big report on this last year, 19 of us.
It was led by Rob Long and Patrick Butlin and had some top AI experts in there, including Yoshua Bengio. And our conclusion was there's no obvious technical barriers for why AI might not achieve something like a global workspace in the near future.
Showing 121–140 of 151 · page 7 of 8
← Previous
Next →