Comment by visarga | Hacker News Reader

Comment by visarga | Hacker News Reader

I have a theory that everything is search. Protein folding? Search for minimal energy. DNA evolution - search for ecological niche fit. Cognition? attention is search, memory is search, imagination is also search, problem solving - of course is search. Scientific progress? It is (re)search. Optimizing an AI model? Search for optimal parameters to fit the data. Reinforcement learning? Search for optimal behavior to maximize rewards. Even speaking is search - we output words in sequence, searching the next word like LLMs. Now I can add comedy to the list.

<rant>Search is a nice concept, it defines everything clearly - search space, goal space, action space. Compare it with fuzzy concepts like understanding, intelligence and consciousness. We can never define them, precisely because they gloss over their input-output domains and try to present a distributed process as centralized in the brain.

Search has a bunch of properties - it is compositional, hierarchical, recurrent (iterative in time) and recursive. This pattern holds across many fields, I think it is based on the fundamental properties of space-time which are also compositional, hierarchical and recurrent (object state at time t+1 depends on its state at time t)

Search can be personal, inter-personal, physical or information based. It can explain away much of the mystery of the three fuzzy concepts I mentioned. I describe cognition as two search loops - search externally by applying known behavior to collect experience, and search internally to compress experience and update behavior.</>

You're spot on that all problems can be interpreted as a search problem. Similarly, all problems can be interpreted as a compression problem. Or parsing, boolean satisfiability, or halting, etc. It's helpful to keep them all in mind because sometimes a different problem domain has a tool that your preferred one doesn't, or just the mindset shift can be useful to unblock.

Lately I’ve come to think of science as a large-scale search algorithm so I think there’s truth to what you’re saying and it’s interesting to think about.

Strong agree. I'm on a similar path as you, travelling through related thoughts

I think this hypothesis goes a long way to explaining why the math of transformers (doing mathematical operations on language) create something that rhymes so much with intelligent thought. Though I should clarify that LLMs do not share the same processes or verbs of our intelligence, only the snapshot moment-in-time of a mind-like object ;)

I think another way to put that really good idea is simply to say that humans are innately wired to explore and discover!

Not surprising I guess, since our survival literally depended on that capability for tens of thousands of years (and still does, to a lesser extent)

I think a more specific denominator you might wanna look into is free energy minimization as you mention in your first example. I really liked reading Active Inference and What Is Life? on the subject.

How does this explain away the mystery of consciousness?

One take: Chasing the Rainbow: The Non-conscious Nature of Being (2017) https://www.frontiersin.org/articles/10.3389/fpsyg.2017.0192...

You might get a kick out of this paper (though some may find it's proposal a bit bleak, I think there's a way to integrate it without losing any of the sense of wonder of the experience of being alive :) )

It analogizes conscious experience to the a rainbow "which accompanies physical processes in the atmosphere but exerts no influence over them".

> Though it is an end-product created by non-conscious executive systems, the personal narrative serves the powerful evolutionary function of enabling individuals to communicate (externally broadcast) the contents of internal broadcasting. This in turn allows recipients to generate potentially adaptive strategies, such as predicting the behavior of others and underlies the development of social and cultural structures, that promote species survival. Consequently, it is the capacity to communicate to others the contents of the personal narrative that confers an evolutionary advantage—not the experience of consciousness (personal awareness) itself.

I find these types of arguments very odd, though at one point in my life I would certainly have endorsed them.

What is it about the modern scientific mindset that makes people say "actually, the ubiquitous experience of being alive, having thoughts, feelings, and making choices, is actually 100% an illusion."

Don't get me wrong, obviously there is interaction between evolutionary functions, the brain, etc - I mean, there's anesthesia, there's being drunk, horny, fight or flight.. there's all sorts of ways that it's obvious there's a link.

But why do so many theorists want to go from "there's a link" to "this is 100% an illusion?" I just don't get it. Is it that uncomfortable to have something that is outside the reach of physical systems theorizing, or something that is unexplainable (i.e., the link) that we'd rather fit reality into the theory than the other way around?

We have to have the courage to live with something that is inexplicable, at least for now (and, honestly, maybe forever), rather than lose faith in our own existence.

Consciousness is just the result of a search for "a more or less linear story that makes sense of the way I act and react"

We're good at predicting states of minds of others (helpful when trying to exploit limited resources, and very helpful for either predator or prey), and we can cheaply gather a lot of data on ourselves, so why should the capability for inferring states of mind not, as a side effect, also provide us with our own inferred state of mind: our own "I"?

This is a theory that would work, except for the fact that I know I exist. Why does materialism so desperately want to ignore that knowledge? Is that really simpler than the idea that we have a soul that interacts with the body and brain, but which also has its own nature that is separate from biology and evolution? Do we have to go 100% and say things that start with "consciousness is just"?

I'm passionate about this because I know -- from personal experience -- that this type of philosophy can really go along with denying ones own existence in a deep way. It feels great to know you're a soul that exists. I don't know why it's supposed to be "rational" to convince yourself against a simple truth that we all know intuitively. And I say that as someone who used to feel that way.

The quality of consciousness and the existence of a non-body non-brain soul seem to me like two completely orthogonal issues (I can easily imagine creatures without conscious awareness but with a soul; I can certainly imagine creatures with brain-driven consciousness but without a soul; a rock has neither; in your model people have both) so it seems unlikely that answering your two questions would move the conversation forward.

For what it's worth, I know I exist as well; can we agree that we both exist, but we have differing models for what the necessary constituents of that existence are?

> Can we agree that we both exist, but we have differing models for what the necessary constituents of that existence are?

Of course, yes.

> I can certainly imagine creatures with brain-driven consciousness but without a soul

This would probably be the heart of the disagreement. I don't believe this is possible. Such a creature would not have qualia.

And, as a species, I don't think we're any closer to resolving this question "objectively" than we ever were. fMRIs say where blood is flowing in the brain, but that's hardly enough to explain the phenomenon of subjective experience.

By the nature of the question, we won't be able to attack it from the outside, and I don't think I could generally convince another person that they have a soul that exists, if they're inclined to explain themselves using materialism, which at this point has become flexible enough to be unfalsifiable, with the everlasting faith that someday science will fill in all the gaps.

That's why my approach now is just to poke holes in the seemingly impenetrable confidence that materialism is the only "rational" way to think.

(By the way, I'm not saying you hold that position.)

> This would probably be the heart of the disagreement

OK, sounds like we're agreed there.

If souls are required for consciousness, then I guess we could try to decide which creatures are conscious by first deciding which have souls? Would that question be any easier to answer that way around?

I thought we were talking about "what is consciousness" rather than "which creatures are conscious." The conversation started with "consciousness is just [a series of material processes resulting in an illusion of interiority, rather than a genuine phenomenon of -- for lack of a better word -- personhood]"

I would probably say "consciousness is the soul" rather than "souls are required for consciousness," but either way I don't see how that helps the fundamental issue that it's impossible to physically prove another creature's interiority, including humans.

What I said was that it's all search. We search and learn, search and learn. This feels like consciousness because it is a recurrent process that feeds on itself. We create relational representations from data, and these representations encode the structure of our experiences. In other words embeddings explain away the qualitative aspects of qualia.

Just declaring that they explain it away doesn't make it so. You've come up with a theory that, from 10,000 feet up, could correspond with certain observations of consciousness. Why should it be true?

We were talking about "what is consciousness" but rapidly hit an end.

Sorry for the lack of clarity: "which creatures are conscious" was my attempt to switch topics to a line which I had hoped might be mutually interesting.

Cheers!

Before we stop, if you don't mind, could you answer why you are dissatisfied with the idea that there's a soul, and why we need to do away with it 100%?

To be clear, I really do take issue with the 100% aspect. There are many psychological functions that are clearly at least mediated if not outright caused by the brain and body.

But I think some people say "there's so many functions that are physical, that probably 100% are and we just don't know it yet." But that doesn't seem logically any more forced than "there's a soul that's mediated by the brain," so I don't know why people are so willing to give up the soul.

> fMRIs say where blood is flowing in the brain, but that's hardly enough to explain the phenomenon of subjective experience.

fMRI scans correlate well with neural net embeddings. That is a great hint. We just need to look at the semantic spaces developed in these models, by a purely mechanistic process, to see how it goes from data to semantics.

The idea that we will just need to do that is a form of faith, no more rational (in my opinion, less so) than belief in a soul.

Making up nonsense will always be easier than actually understanding reality.

Just because we can’t explain something right now does not mean you can insert whatever you want into that hole and assert it’s just as valid as any other explanation.

It's funny how much I agree with your whole comment ;)

Also, "making up nonsense" is a very disrespectful, and intellectually dishonest, way to approach someone's understanding of life which is hard-earned through experience and thousands of hours of introspection and study. Consider that people who disagree with you on this may still be just as educated and smart as you are.

Not if they think making up supernatural explanations is valid.

It’s the same logic as God of the gaps. Science doesn’t understand something yet so better fill it up with feel-good made up stuff and pretend it’s just as valid as actual science.

To me, this theory is Science of the gaps. "Science can't prove it yet, but it will." :shrug: That's faith, not logic.

One of the stupidest things I’ve read on here. Science has an incredible track record. Faith doesn’t.

Science, in fact, does not have an incredible track record on consciousness. Science does not even claim to be able to answer every question that can be posed. So you are making a leap of faith assuming it will continue to make progress on this question. You are in fact, being very irrational, as well as aggressive and unpleasant. Good day.

We have a dual search loop - outside, we act using our experience to gain new data. Inside, we compress this data and update our experience. We search for experience and search for understanding. Acting is search for new insights, learning is search for error minimization.

I think the way we encode our experiences is relationally, like neural networks. We relate new experiences against past experiences, this creates a semantic space that is highly dimensional. Any concept is a point or a region in this space. It has consistent semantics, which leads to the unified experience. We can relate anything to anything in this space without having a central understander. Encoding your own experiences creates a first person perspective from 3rd person data, which was always a "hard" problem to explain in philosophy.

The serial action bottleneck adds to the illusion of centralization. But it's still a distributed process, no neuron is conscious or understands by itself. And even in society, no human can recreate even a 1% of human culture individually. We are not that smart on our own. We should always look for the larger context where we develop, not just the brain.

Search has the virtue of not hiding the environment, it is social and distributed, unlike more personal concepts like consciousness, intelligence and understanding. But as I said above, even inside the brain there is nothing but distributed processing, no homunculus.

I think the core of my argument is "there is no centralized consciousness, understanding or intelligence, they are distributed processes, they act across neurons in the brain and across people in society". It seems like a hard pill to swallow, if that is true then there is also no centralized understanding or truth.

I commend your diligence, your theory may even be internally consistent, but I don't see why I should believe this rather than the comparatively simple and intuitively true (to me, anyway) notion that there is a genuine soul, it has its own volition, and it certainly interacts with the physical body and brain through a link that is currently (and may always be) mysterious.

What you say that the soul acts like a homunculus, "interacts with the physical body and brain through a link that is currently mysterious". It is a "center" thing that has volition, and semantics, and is genuine.

And it is understandable to do so. It certainly feels unified, and genuine. And it is simple and nice to be so. But intuition fails us hard when it comes to introspection. We feel alive, and conscious, but we are a few billion neurons connected by trillion branches, all wrapped in a bio-robot, put inside a complex environment full of living things in a state of cooperation and competition.

Again, that is your theory. Just because you've created it using the primitives of materialist science doesn't mean it's true or that it's more likely than another theory. The fact is there's no scientific basis for believing what you've said. There just isn't.

If it feels like there's a scientific basis for what you're saying, that's because you're invoking the metaphysical principle of strict materialism, which is a common (though not inherently valid) way of thinking about science, and if that were true then yes, something like your theory would have to be true.

But be clear - that is not something proven by science, it is an axiom that is far from obviously true or universally accepted.

It may also feel that your theory should be true because so much about how the body and brain works is, I agree, materialist in nature, so why not hypothesize that everything is materialist? But that is very much a hypothesis and an intuition, not something that is scientifically proven or logically forced. And, as you say, intuition can be flawed.

Reply on news.ycombinator.com