Isn't there still an explanatory gap? You know, the what-is-it likeness?

This is the inside perspective, right? And Buddhism to me is limited to this inside perspective because Buddhism has been invented at a time where people were focusing very much on the experience of reality and how to adapt to this reality, but not how to split the atom, how to deal with the reality at the lowest level with the foundations of physical reality.

So Buddhism is not so much interested in understanding physics.

And in a way, this is more a Western tradition that starts with Aristotle, who thinks of the psyche as something that imposes itself on physics and interacts with it.

And to understand consciousness from the outside, we need to make a model of physics that can explain under which conditions systems can make arbitrary models.

Under which condition could a system build a model of what it's like to experience? Models at all.

And this is something that computer science is doing today, that it looks at representations.

And maybe just start from there.

ai that altman's building

yeah i think that this notion of control implies that somebody needs to control us and i find that the us is not a very liberal society anymore uh it was quite shocked when i moved here from europe that in many ways how controlling the society is like you don't own your body you only rent it from the fda You're not allowed to enlist yourself in medical trials for new cancer treatments because that would be unethical if you just cure your own cancer in waste that is not sanctioned by the ethicists.

And I think it also leads to poverty.

It leads to not just a lack of freedom, but it leads to a situation where almost nobody is able to live in a decent house because you're not allowed to build one.

You're not even allowed to change the tiles in your bathroom without a permit.

It all exists because you have highwaymen at every corner that make sure that you have to pay some licensed professional who has captured the regulators to exert control over you so you cannot be free.

And this society is also a safetyist in many ways.

Basically,

I didn't know if it

was art.

No.

Do you believe in God? In some sense.

I think that God is an aesthetic.

It's multi-gods.

There are many.

I mean, it's like you can just

Book now at aa.com

aa.comSome intelligence happening there and some mechanism that makes things coherent.

And at some point, you get to plants that are themselves coherent, are able to communicate within themselves and across each other to some degree, and it's not quite clear how smart plants are, right? Every cell can send messages to the neighboring cells, and this can be conditional.

And so in principle, the multicellular plant is some kind of Turing machine.

And maybe it can discover a learning algorithm over a long enough evolutionary time spans that makes it generally intelligent in the same way as the software that runs on our nervous system, just slower.

So maybe the nervous system is not the thing that enables intelligence as we currently mostly think in the mainstream of, um, biology and neuroscience.

But, uh, maybe it's just an optimization that enables animal intelligence.

Um, okay, so if reality is like fundamentally this construction, uh, like some sort of cognitive model of what's out there, almost like a compression or a simplification or an algorithm of it, like, um, w-what would it mean to encounter something that was like completely resistant to modeling? Like could there be any kind of, kind of consciousness with no concept of time or anything that's outside, uh, modeling?

I mean, do we need to have consent before we shut things down? I mean, at what point are you c- comfortable with putting the genie back in the bottle if Roman's correct? I mean, your, your p-doom might be irrelevant, but how do you speak for the rest of, of humanity? Not you personally, but how do we as scientists or as technologists?

There are a number of considerations when we try to compare the, um, nuke situation, um, with the situation of artificial intelligence, and, uh, it makes it very difficult to compare even, and it's also difficult to dismiss the comparison out of hand.

Um, I, like many people in my generation, uh, Generation X, um, think that now that I've been wrong about, um, the impact of nuclear weapons.

I grew up thinking in the 1980s, um, that, uh, we are just a hairbreadth away from, uh, assured destruction and that the risk, uh, of, uh, nuclear armaments is way too high for humanity.

Uh, now I think the adults have been correct, and a generation of students, uh, who took to the streets and were very upset about this turned out to be wrong.

Uh, nukes have prevented several world wars that would have been happening between World War II and now, that would have, uh, wrought tremendous destruction on Europe and potentially, um, uh, large parts of Asia and maybe even America.

And, uh, this, um, basically more than half a, um, century of peace that we had is completely unprecedented, uh, for large nations like ours, and I think this was achieved due to, uh, nuclear armaments, which have been, uh, empirically quite safe.

But, but what do you make of this, this, this regulation push? It's, it's somewhat unusual in history.

Can you unpack that, please?

Mm-hmm.

They are looking at alternatives to the present ways in which machine learning is emulating the statistics of human media production, thinking about systems that can continuously learn and that basically project the architecture of the mind across the substrate.

And some of these ideas are also expressed by Mike Levin, who thinks that the processes that make an organism coherent with itself through morphogenesis are very much alike or maybe even the same ones as the one that make a mind coherent with itself.

And so we are actively looking into this idea of coherence maximization, that a system becomes coherent with itself, that identifies numerous constraints between local subsystems that are being projected back and forth until the system finds some kind of state in which it is behaving as if it is a single thing, that has a single architecture, despite being made out of lots and lots of middle agents that have to talk to each other to form this global structure.

By the way, I don't want to appear like I'm completely certain as if you're going to succeed and we can recreate consciousness on digital substrates.

This is a hypothesis that we have that we want to test.

Do you see application of your work around this that might help keep us safer?

How would you describe their purpose and yours?

My goal is to make headway on understanding the question, how does consciousness work? How does the human mind work? And my working hypothesis at the moment, might be wrong, is that consciousness is a particular kind of representation that is necessary to make the human mind form itself, to build its structure.

What we observe is that we don't become conscious at some point in our life due to the complexity of our mind that makes us unique in the animal kingdom, but that we wake up already as babies before our mind is fully formed.

And if you do not wake up, if the lights don't go on, then we remain permanently vegetative, and we do not become fully fledged human beings if we don't build our cognitive architecture.

And so I suspect, since we all are conscious, that consciousness is the simplest solution to train a mind like ours.

And what we observe about consciousness from the inside is that it always happens now.

It's an experience of present and presence, and this tells us that it's perceptual.

Looks like they understand, but do they or not? Where do you stand on that?

Right.

And so instead, we only look at behavior in the way that we can evaluate in the lab so we don't have to fit our curves in a way that statistically stops making sense.

And so maybe we leave this to the neuroscientists.

And the neuroscientists also didn't feel, uh, responsible for consciousness because their responsibility is to look at individual neurons and their interactions, and how this interaction pan out- out without thinking about the nature of representation.

And philosophers got lost with respect to understanding the nature of representation because they q- at some point started to no longer follow the discourse in the theory of formal languages that would allow them to understand what the computer scientists were up to.

And the computer scientists never realized that the philosophers don't un- actually understand what a representation is, what learning is, uh, what the mind would need to be to do things.

And so this leads to a vacuum of ideas, and we need to fill it again.

But I mean, because, because of the particular epistemology you've a- you've approached it as in terms of observable consequences of the thing, I mean, I know it's an interiority problem, but y- y- you also have this hypothesis that that particular interior configuration produces performance

What does that look like practically?

This technology is only a few years now in, in the public spotlight, and it's-- people are only starting to use it.

There have been some predictions.

You remember maybe the movie Her, which is now 12 years ago, and Her describes a society in which most of the material problems are being solved and people are still lonely because they're difficult to get along with each other.

And in this situation, um, a man installs a new operating system, and this new operating system is a vision of what AI would be like if it was a hot woman.

And he does fall in love with her, and she's set up in such a way that this is a mutual relationship.

But she keeps developing, and at some point she drops out of the scope that is accessible to humans, and she is bored by the long breaks between his thoughts and his voices and the frames of his video and moves away and humans are left alone again.

Mm

Um, but i- if you have a sense of the, the way things are moving right now, you know, what they're sort of converging towards, or is there just far too much sort of fog of war to be able to answer that?

Well, one thing that has changed is that ideas seem to be disseminating much faster than they did in the past.

And this leads to a more rapid development because ideas are begotten by earlier ideas.

And so if the turnaround cycle of ideas is increasing, the memes are coming higher frequency, then, uh, the future is changing faster.

And, uh-Things are implemented more quickly, ideas are spread more urgently, and, uh, they go into, uh, many more years.

One thing that I find interesting is the extremely negative view that people have taken on artificial intelligence, and I suspect that's largely due to two factors.

One is, um, that, um, large media organization, um, starting from The New York Times, decided that they see artificial intelligence as an economic and social threat to themselves.

They're motivated, you know, politically and in terms of storytelling and being able to create these hype machines that say, "Oh yes, this technology can do anything, and it will be able to do even more as long as you plow $1 trillion into, you know, funding, you know, the infrastructure for it and the learning for it." Um, w- what's the implication there, and, and how concerned are you by that sort of ownership model? And also, I guess by the fact that if we have this, this layer now in our economy that, that the, the, the ownership of it is so centralized versus being decentralized across, you know, a much broader worker base.

Why?

At the moment, we don't have AI as it existed in the horrible science fiction stories.

What we have are statistical text generators and some multimodal models that can help us with image recognition and generation of media.

But, uh, it's not clear if, uh, these models automatically become self-improving.

At the moment, the AI companies are working very hard at, uh, getting these models to a level where they approach the capabilities of human experts, and the models are, at the moment, below this.

They are, uh, in some sense, idiot savants.

They have read everything.

Where do humans fit into this world, and what happens if, if we just get displaced before AI can self-design? But what happens if we just become kind of redundant?

What is your worldview and, and how does it relate to, to really not just the end of days, but theories of re-resurrection, redemption, and the ultimate meaning, you know, to some people is to be reincarnated, right? So where does this fit in? What, what did you mean by you don't die because you were never really alive?

I think that nihilism is a frustrated sense for a need of meaning, right? When you feel that you need meaning to be alive, some kind of deep connection to existence and your being in the world, and if this connection is not visible, then you feel like an ant without its hill.

And this thing that when you separate an ant from the anthill, the ant just withers and dies instead of trying to strike on its own and enjoy its newfound freedom.

And this is a condition that is also present in most of us because humans are a social species, somewhat similar to social insects.

Our meaning does not just exist in ourselves, doesn't end in the ego and our own organism for most people.

But for the vast majority of people, right, for, for the people who are not sociopaths, it exists in the connection of the superorganism.

So in this way, we are like cells in an organism, and the meaning of the cell is contribution to the organism.

What do you make of this theory?

Do you think brains are needed for consciousness? Do you need a brain to have a sufficiently information processing system to do this second order perception? Or do you think this could be happening in even simpler systems that don't have brains or individual cells? Even, you know, is my kidney cell conscious? Is my kidney cell capable of second order perception? Or do you think you need a brain to do this?

I have no idea.

So I really don't know.

And I'm agnostic unless I have a criterion to find out.

And so I don't see a way to completely rule it out.

As a computer scientist, if I look at the cell, it's processing information, right? And it can perceive messages from its environment and get conditional messages in response to other cells in its environment.

So in principle, you can build a computer in any kind of multicellular organisms.

But also maybe we're just moving too fast and none of us know how to act and none of us know how to as a society to adapt to it because it's just moving too quickly.

How do you think and where do they fundamentally fall short in their assumptions?

To me, the greatest shortcoming is when we pick a belief despite being unable to prove it.

As long as there are logical alternatives to a certain belief, to a worldview, we must acknowledge this and remain agnostic.

And we can maybe...

in some sense quantify our agnosticism based on evidence that we have.

But as long as something cannot be logically ruled out, it's possible.

And so we cannot simply say, well, it's possible and I don't see a logical reason against it, but I don't believe in it because I like this view better.

And I'm curious what you think about that and the futility of us being able to try to create our understanding of how this all works while being nested within it.