C. Elegans: The worm that no computer scientist can crack
wired.com
wired.com
https://www.youtube.com/watch?v=c8iFtaltX-s
Michael Levin's channel: https://www.youtube.com/results?search_query=michael+levin+a...
His blog: https://thoughtforms.life/
(link isn’t loading for me on mobile, will try again later)
As far as I understood he doesn't state that life cannot be explained by physics. But that it might be more productive to choose a more high level view.
Makes sense. If a set of physical laws allows life at all, life will maintain its own set of rules (related to reproduction, goal-directed behavior and so on) regardless of low level "implementation details" that are dependent on particulars of the physical laws.
For example, a static zero-dimensional universe. There's no pi, no Chaitin's constant, no nothing. Or, in other words, there's no processes or objects in there that we can describe as "ingress of platonic forms" (and no observers to notice that).
I do like his ideas (and I wrote similar things about platonic forms elsewhere), but it's not a solid refutation of physicalism. It's an attractive framework, but as almost everything in philosophy it can easily be challenged.
A physicalist can say "Physical processes that follow a certain equation trivially have properties corresponding to the properties of the equation. So what? I can measure physical process and I can think about the equation (thinking is a physical process too), but why should I postulate independent existence of a/the platonic form of the equation?"
The fact that my subjective experiences undeniably exist makes me reject physicalism, but I can't prove their existence to anyone else and I can't use their existence as a solid basis for some philosophical view. After all it's just one bit of information. Or zero bits? I wouldn't have noticed my own absence.
> It's an attractive framework, but as almost everything in philosophy it can easily be challenged.
Michael is aware of it. He insists that every speculation has to be experimentally tested. But no experiment of this kind will constitute a "proof" - someone can always "challenge" it. This is no different from a physical theory: every interpretation of QM is challenged by someone. :-)
After this commemt chain it comes across as both dismissive and arrogant.
> I do like his ideas (and I wrote similar things about platonic forms elsewhere)
What is dismissive about my interest to his write-up? (I'm sure it was a misunderstanding on your part, you don't need to apologize :-))
He goes from:
>patterns that are useful and guide events in the physical world but are not themselves explained, set, or modifiable by the laws of physics. This includes things like facts about prime numbers...
which if fair enough. Prime numbers exist and you find them by doing calculations but he goes on to:
>...this position is unpopular with philosophers of mind because it is fundamentally a dualist theory (by emphasizing causes that are not to be found in physical events)
which seems iffy. As mentioned we find primes by calculation which can be done by physical events.
I think human reasoning tends to get in a bit of a muddle with this because the idea of "non-physical space of truths which we discover" is not an especially good mental model for stuff we could in principle calculate.
That said I figure stuff we could calculates exists even if we haven't done the calculations yet because someone else could have done so, and stuff that could be calculated includes simulations of other worlds.
I used to think Open Worm was going to be a big deal. Top-down AI seemed to be stuck during the AI Winter. But machine learning got unstuck and Open Worm is still stuck.
“~= 0” just means we likely will be unable to measure any effect. Not that there literally is no effect.
Us being unable to do science on something means we cannot scientifically state that it does not exist. Not that we can categorically state that it does not exist.
I am nitpicking this: The state of computational chemistry is not at a level to support this. I'm optimistic we'll get there eventually; need to find novel approaches, and current ones are too imprecise, or too slow.
I think the connectome aspect is more interesting, as it may be feasible to get there 100% without a computational chem breakthrough.
To give a sense of the scale of this problem, the largest frontier simulations I’m aware of are around the trillion atom scale. (On tens of thousands of GPUs [0])
Based on a quick web search, a c elegans cell is between 3 microns and 30 microns in diameter, so if we assume we can count atoms using the density of water then an all-atom simulation of a single neuron would need between 5e11 to 5e14 atoms. c. Elegans has 302 neurons so simulating the full neural network will be 2-5 orders of magnitude larger than current frontier simulations. Honestly more doable than I thought it would be, though all-atom simulation of a full organism still seems quite out of reach
This is all with classical force fields. Doing this simulation at the electronic structure level is much much harder with our current modeling capabilities
0: https://www.mrs.org/meetings-events/annual-meetings/archive/...
it seems to me "the quantum computer you seek" is the molecule + the medium (especially the medium) itself
I don’t pretend to have any particularly deep insight into quantum algorithms for chemistry, but [0] is a really nice review. It seems like there are a lot of possibilities for simulating general molecular and materials systems on quantum computers. The holy grail would be solving the exact quantum mechanical wave equation in sub-exponential time and space complexity. I don’t know how feasible that is, but it seems like people are making progress using quantum algorithms to accelerate approximate quantum simulation [1].
Back to all-atom c. elegans: I think quantum computing is more about accurate and scalable electronic structure modeling, and simulating enormous systems like this will still require fitting classical (meaning electrons are implicit) force fields and running them at scale for the foreseeable future. A lot of this is space complexity - I’m not sure how a quantum computer could do atomic simulations with sublinear scaling of qubits in the number of atoms being simulated, and were in the very early days of scaling quantum computers up
yes. you nailed my point-that you will have to fit classical or quasi-classical fields, which is liable to require scads of qbits just to get close. qbits are just not "designed" to do that sort of thing.
in any case we ~solved protein folding heuristically and not using fields so i shouldn't be too pessimistic that it's impossible that quantum compute will help eventually.
neurons aren't binary. who knows if they're special in a way digital computers can't simulate -- they aren't well understood.
nobody said anything about determinism besides you
Most of the time someone brings up the word 'Digital' they are talking about 1's and 0's, and how everything is known and determined, and thus looses the ineffable human quality.
"they're special in a way digital computers can't simulate"
I might be making a straw man, but you are literally just saying "something magical that exists outside physics must be going on"
But it would be the case that the universe is non-computable, and if this is the case, even without "magic", you would not be able to reproduce certain kinds of things (even not conscious ones) with a computer.
I don't really understand what this means - you obviously know at the smallest level things happen by random probability - because you mention the quantum world - but then you say the universe is deterministic.
Not sure if any other wizard hat things than those can extend classical digital computing
Bringing up 'Quantum' effects to somehow free us from determinism and somehow introduce free will, is completely false. Somehow using the word "Quantum" explains human consciousness is wrong.
A coin flip is still determined. You don't know which way it will flip, but also, you aren't deciding which way it will flip. The future is determined either way, determined one way, or determined the other way.
We don't actually know this. It is just what we observe that fits our map so far.
For a copy near you, https://search.worldcat.org/title/1077130914
1hr long lecture by Claire Evans https://www.youtube.com/watch?v=wLQKp8mHHJ0
If you are interested in more background on biological brains, c.elegans is really just "the very first breakthrough" (in the phraseology of Max Bennett's A Brief History Of Intelligence -- a recent, worthwhile popular science book).
Yea, she's gonna give it to you baby! She's gonna go far, kid [1]
I’ll be that guy, but a computationally complex problem won’t push your computer to a temperature beyond design limits.
Allegedly. It allegedly won't do that.
Practically, running your CPU at design limits for a very long period of time tends to cause the temperature of the rest of the chassis to want to equalize to that temperature, which can be above comfort limits.
When someone says that the sky is falling, that also doesn’t need to be explained
If it can't be measured in reality, how can it be relevant to reality?
> a random number generator in base reality?
How does randomness pertain to free will?
however, there may be hidden variables outside of our physical reality that are actually mechanizing the result. some such processes may be non-deterministic, which is why i used randomness as an example.
what i'm implying is analogous but opposite to the concept of a philosophical zombie. there may be a ghost in the machine which no measurement can reveal.
> they may have measurable effects in our physical reality
If there’s something external that interfaces with the physical universe, such an externality could be observed. It’s strange that we haven’t found such a force. But if it were to exist, in some parallel universe, that external force would have its own mechanics and its own chain of causality – its own physics so to speak. Dualism doesn’t get you to free will, it just means there’s physics we can’t observe. (Or perhaps there’s some superset universe that interfaces with that universe, and then it’s still determinism or stochastic processes all the way down.)
We still don't know so much, yet we claim so much.
I think when most people say free will they mean dualism, in that there's some sentience in the spiritual plane that directs their bodies in the physical plane. But if this spiritual plane has no observable effect on the physical plane, it's completely incompatible with free will. And if it is observable, then it is indeed a measurable part of physical reality, but yet we haven't measured it - not even stochastic effects (which can still be observed statistically).
Sabine Hossenfelder has a much better informed take on this, and it's worth a watch.
https://www.youtube.com/watch?v=TI5FMj5D9zU
Also of interest, a study where fMRI readings were used to predict a persons decisions well in advance of them executing the decision. The success rate was only 60%, but still better than chance, and this study was way back in 2008:
> fMRI machine learning of brain activity (multivariate pattern analysis) has been used to predict the user choice of a button (left/right) up to 7 seconds before their reported will of having done so.
https://en.wikipedia.org/wiki/Neuroscience_of_free_will#Neur...
Another plausible interpretation of the case is that our brain does have a predisposed intuition building about something, and so it is not a [surprise] that the person chose whichever his intuitions perceives as better or more desirable, it also explains why the 60% correctness and disproves this as being evidence for some "determinism". Nobody is saying, for example, that free will equates being unable to be influenced by something, it is not a surprise that someone that has a vice in crack is craving for crack and in "70% of the cases" (or whatever) decides for using more crack instead of not using. Or that, if I desire to eat X, and X is available, and I'm planning to do it, I will eventually do it, it is expected that my actions are in good relation with my previous intents, desires and knowledge. The same way, it would be impossible for those experiments to predict something before showing the subjects what is being tested, and people are known to acting very differently (like, trying to outsmart the scientists or show themselves as better than they are) when tested, so the own thought of "I need to chose X" could be in their minds way before they [state] their conscious decision (because, if I like chocolate ice cream, even if I stare to the menu for 30 seconds I will still probably chose chocolate ice cream, but sometimes I could chose mint, it is not a surprise that decisions follow some kind of pattern when it is reasonable to expect so). The question is always about to what extent this is a [determinant], and the free will defense merely needs to say "not 100%".
It is also extremely unrealistic and out of this world to think our decisions would be able to be "predicted seconds before they happen", if the decision was conditioned to specific reaction events this would be IMPOSSIBLE, your brain cannot decide 7 seconds ahead of time what is the correct decision for a problem that you have only 3 seconds to decide (like "press the green color" when it appears or "type the word being shown in screen when you see it"), no kind of predictability would realistically arise from this kind of behavior (because it is physically impossible) rather than "the brain is preparing to type".
I’m also religious and I believe in God (and the Holy Trinity) in the core of my being.
And as could be seen with the FTX debacle the outcome was basically the same but without helpful government regulations.
This is, I fear to tell you, IMPOSSIBLE to explain by other than methaphysics, it is an uncloseable "gap" that humanity will never solve. As far as I know, it's not about gaps, it's more about the unreasonable stance that everything is explainable by reductionist physics that corrupts the modern thinking, it is more safe for you to assume people's religious beliefs come from "gaps" rather from their own rationality. In reality, you just turned science into your own religion.
There is still the general fact that plenty of processes are 'chaotic' making them impossible to accurately simulate (because errors in the initial state compound exponentially). But it seems to me like a good enough simulation of a chaotic system is phenemologically indistinguishable from the real thing.
But as far as I understand that is just the result of constraints imposed by the atom/molecule "box". Like how a guitar string can only vibrate at certain frequencies because an integer multiple of wavelengths of the standing wave must fit on the string.
Outside of such systems, energy levels are not quantized. For example, photons from distant galaxies appear to be redshifted on a continuous spectrum.
Neither proven nor disproven in any capacity.
Also I have no clue how anyone could assert that the universe being "continuous" (whatever that means) could be casually claimed by anyone to have any link to our universe being "a simulation" (whatever that means) or not.
So to simulate continuum you need a transfinite VM. It will be more efficient to just run the universe bare metal.
Both are silly statements that can’t add anything.
I think any rule-following (ie ~computational) system is either below, or at the same level of dynamic complexity as all others.
I also think that it's truly impossible to draw a sharp line between alive and not-alive, so the idea of 'free will' would have to extend all the way down to the smallest things like atoms/quarks/space etc.
Is it really meaningful to suggest that it could be anything like 'free will' at that size? Sure some quantum stuff seems indeterminate but could be fed by something that looks random to us - it might not actually be (dunno!?), but the idea that things at a larger scale could be meaningfully controlled by feeding-in different values (that still somehow have to look perfectly random!) seems extremely far-fetched to me.
Larger things like a molecule, a transistor or a human! generally stay as coherent and predictable/controllable (or self-controlling) things, even in the face of indeterminate-ness at the smallest scales underneath. I just can't see how there's any utility in connecting the idea of free-will/consciousness to this indeterminacy.. it's like saying free will is just fuzziness, and how does that really help in understanding life, the universe and everything?
I started reading the Wikipedia page about free-will to try to work out where I stand on this. I think I quite like the classical way of thinking about free will as simply: Being free from coercion in one's decision-making! I guess the source of such limits could be either internal or external or both though! So I don't know whether that helps much in answering your question.
The Wikipedia page says about 'compatibilism' (which holds that free will and determinism are compatible) that it is "the absence of various impediments" and that "a coerced agent's choices can still be free if such coercion coincides with the agent's personal intentions and desires", so that still doesn't really help answer the question.
I accept quantum-mechanics enough to be slightly skeptical of strict-determinism, so I guess maybe I fall more into the 'metaphysical libertarianism' point of view, but I still feel like the non-determinism that QM introduces probably acts more like a limit to how 'free' free-will can possibly be, rather than being a kind of 'secret sauce' that makes it possible for us or some external factor to be able to meddle with what reality (or ourselves) get up to in a meaningful way.
I also think that the internal/external distinction breaks down when examined very closely. For me this leads to the idea that I am actually the same thing (same 'wavefront' or whatnot?) as what caused the existence of the universe that I am a part of. I am both myself and the environment at the same time, but with a limited ability to sense very far into the external universe and its workings, beyond input from senses like sight/hearing/etc (I guess realms of pure-thought should also be included too though! Do we 'sense' reason/philosophy/mathematics/algorithms when we think about them?).
I think one answer (or the avoidance of a clear answer) to your question is that people have been arguing about this for thousands of years, and we (as a collective/species) are still really not very close to being firmly decided about it at all!
What is free will if it's simply causation? i.e., environmental inputs leading to differences in charge, altering other differences in charge, leading to outputs, leading to environmental outputs, leading to changed environmental inputs, etc. If the chain can be examined and is entirely deterministic, be it neuronal or silicon circuits, where's the escape hatch?
Another thought experiment: if there's something that is you, that decides, and presented two different realities where the environment, brain, etc. were precisely the same, what would cause there to be a difference in decision? If it's deterministic, how is that free will? If it's random, how is that free will?
Free will is the interplay between determinism and randomness, an emergent phenomenon with multiple self-recursive feedback loops and path dependence. Even if we could trace it through all these loops and find all the mixtures of quantum randomness and classical deterministic patterns it emerges from, it wouldn't make it any less magical.
The self is just the only part of the model directly wired to our neurons and self awareness is just correct labeling.
In that context will is certainly a function of a how much a things internal model of self encoding the idea that its actions are moreso based on its own model than the tgings its model is built with and ability to make this distinction right or wrong and free will is basically meaningless.
Its certainly not a magical out for determinism all the qualities that we imbue with such meaning are as based on the same principals as anything else.
> If it’s random how is it free will?
Stochastic processes won’t get you to free will any more than determinism.
The idea of free will has been a subject of eternal debate. I suspect this reflects lack of consistent definition. I would posit that free will isn't absolute but necessarily constrained by the nature of individual exercising its will. The stochastic attributes of a system or entity mean its actions are to an extent unpredictable, providing an "opening" for willful behavior.
Humans imagine they have free will because they're aware of their decisions or actions while unaware of the range of factors contributing to a decision or action. Intuitive (vs. analytical) cognition is the operational default. By definition intuition is a computation occurring outside of the person's awareness.[0] Consequently, it augments the impression of exercising unfettered free will.
Perhaps it's most accurate to say we have will or volition but not free of constraints imposed by our biology and physical/social environments. While the randomness inherent in biological systems allows volition to evolve, it also limits what an organism can will itself to do.
[0] Journal of Cognitive Engineering and Decision Making 2017, Volume 11, Number 1, March 2017, pp. 5–22
You might be talking about compatiblism.
> Humans imagine they have free will because they're aware of their decisions or actions while unaware of the range of factors contributing to a decision or action.
Or more likely there’s no evolutionary fitness benefit to being able to understand this, and perhaps it’s even detrimental if it leads to nihilistic or egocentric impulses.
I don't consider free will to be a scientific concept. It is philosophical, or religious. But if we give free will to things we can't simulate accurately, maybe planets like the Earth have free will. It is a common religious belief, and if you believe that simple life forms have free will, not too far fetched. The Earth, with its atmosphere, magnetism and geology is a really complex system we have a hard time understanding.
We could already be living under that same double virt penalty.
I think we have an outsized focus on free will, partly because free will is a trick played on us by evolution. It takes an exorbitant amount of effort to exercise anything close to what most people would consider free will and the phase lag is measured in hours if not years.
I am not sure where whole-cell simulation is at the moment since I've been away from the field for about 15 years, but I recall a rather difficult multi-month simulation that was trying to model an "empty" volume of cytoplasm away from the organelles, about 1/50th of total cell volume, and with all proteins and metabolites replaced by hard spheres of varying "stickiness". It was considered a huge success to just get a few of the diffusion rates for various compounds in the right order of magnitude. I mean, if you really want to get the fleeting interactions right you need to be modeling individual water molecules. I know there have been large advances in computing in the intervening years, but this was on I think #20 in the TOP500 at the time. Unlike AlphaGo I don't see any immediate avenue for AI to help with this because unlike protein crystal structures there is no wealth of quality training data for cellular dynamics at the molecular level.
I've been thinking it would be cool to design chips to be realtime safe—that way there's no need to synctronize—and have further away information delayed (just like relativity) to deal with speed of light communication issues.
Never heard of fleeting interactions, would you mind to elaborate?
You'd think they'd try something like smooth particle hydrodynamics so that it can be done more efficiently on a GPU cluster but I think they are just doing classic CFD
I.e. the pattern of cell division and the positions and identities of the cells is very fixed, with known timing etc.
Compare this to mammalian development, where even at the early stages, development becomes less rigid and operates by "course correction", rather than by following strict sequences.
(I studied the early development of C. elegans for my PhD)
Has a computer scientist cracked any complex system that was not engineered?
They have the architecture, but can't read the weights and biases as it were? I was under the impression that they somehow mapped the activations, but if not then they're really just flying blind.
I didn't experience that while reading the article (waterfox browser), and can't recall that ever happening on Wired, and I somewhat frequently read wired articles (few times a week, not a registered/paying member).
I tried it in chromium (no extensions) and if you scroll there is a fullpage ad where the article seems to "disappear" but it's essentially an ad fold, and if you keep scrolling the rest of the article is there. Otherwise not sure what you mean? What browser you using?
also, last i checked, an llm doesn't meet the definition of life form.
LLMs is the new hype product of tech companies. Wait a couple years and the interest will die out.
Maybe we’ll live the day when true neuroscience (none of that Andrew Huberman stuff) will be trending.
It just means the complexity is harder to capture and copy.
LLMs are built via algorithm. Given enough data and a large enough neural network the complexity of an LLM is boundless. I guess my question is are existing LLMs more complex?