[Iv]

A reason why I allow myself some skepticism on the feasibility of quantum computing is that if it were possible, I would have expected evolution to have used it somehow.

If it turns out that protein folding really use a quantum computation, I'll move quantum computers from the "too good to be true" category to the "probably revolutionary stuff that I will see in my lifetime" alongside with nuclear fusion and strong AI.

[knzhou]

This argument proves too much. By that same argument, nuclear fission, ordinary CPUs, steam engines, rockets, helicopters, jets, X-rays, radio waves, superconductivity, superfluidity, air conditioning, liquid helium, liquid nitrogen, the Haber process, the fractional quantum Hall effect, Doppler cooling, graphene, long-distance satellite communication, gravitational waves, and GPS corrections for relativity all don't exist, because nothing in life is designed like them or takes advantage of them. You've literally taken us back to the 1700s.

I know that it's fashionable to simply declare quantum computing is impossible, and there are some strong arguments in this direction, but this particular argument isn't one.

The general reason people believe quantum computing is possible is that it describes just about all the things I mentioned above absolutely perfectly, along with literally thousands of other phenomena, with no deviations ever measured. This gives us good reason to assume quantum mechanics actually works, and if it does, then it's possible for quantum computing to work. (Also, of course you need to account for quantum mechanics to account for protein folding. You literally can't have chemical bonds at all without quantum mechanics.)

[Iv]

The argument I propose is not a strong one. It is still an argument: if something is possible, why did evolution not use it? There are several possible answers:

1. Life may not have a use for it

2. It may be impossible to achieve with proteins and cells

3. It may not actually be possible

For quantum computer I (weakly) believe that 1 and 2 are wrong: evolution and cognition would hugely benefit from quantum acceleration and biology operates at a scale where quantum effects are visible. I thought 3. slightly more likely but I'll readily admit that I am nowhere near the knowledge to be categorical about 2.

And note that of the list of things you are giving, there are many that uses the same physical principles that are used by life: steam engine (expansion of heated gases), rockets (ignition of gas), jets (propulsion), helicopters (a rotating wing is a wing), radio waves/X-rays (the RF spectrum, which visible light is part of), etc... The rest, IMO, falls either under 1. or 2. For instance I doubt long-distance communication really offers a substantive advantage when you know whales can already contact each other at 100s of kilometers through shouts, and superfluidity may require conditions and materials that are impossible to reach for organic material.

Note however that this last one is actually a kinda good (if weak) argument: if superfluidity was achievable through organic material and conditions close to the temperature and pressure average on earth, life would probably have found it, as it is clearly a useful property. If tomorrow we find that you can get room-temperature superfluids that are made out of C,H and O atoms, wondering why it is not found in nature will be a very good question.

[sriku]

All that is assuming that you do know for certain the evolution did NOT exploit it.

There is some extrapolative argument that hints at the contrary. Adrian Thompson's evolved FPGA circuits exploited a single chip's underlying physics in a manner no digital circuit designer would. By that thread, it would seem possible that evolution has already exploited quantum computation ... just that maybe we haven't had the tech eye to see it yet.

After all, all systems are quantum mechanical.

[smilliken]

Rhodopsin, in our retina, exploits a quantun mechanical effect. You could argue that our brains are quantum computers if you consider our retina including rhodopsin to be part of the brain.

[whatshisface]

Apples are red because of quantum mechanical effects, but that doesn't make them computers.

[smilliken]

Anything and everything is because of quantum mechanics, sure. But you could've asked a classical physicist to design an apple "computer" (the fruit)— just arrange the right atoms— but you couldn't have asked a classical physicist to design a retina computer exploiting rhodopsin.

[simonh]

I don't think 2 is so easy to justify. Quantum state transitions occur all the time in a discordant mess of activity at room temperatures. The only way we can control for that and have quantum transitions occur in useful controlled circumstances is by operating at close to absolute zero, which isn't very conducive to exploitation by terrestrial life.

It may turn out to be a similar issue to jet engines, or semiconductors. The materials and conditions required for them to operate just aren't very easy for terrestrial life to evolve into.

[soulofmischief]

Nature works on the premise of emergence. It's entirely possible that the energy gradient required for a biological system to make successful, permanent changes beneath its fundamental layer of operation is just too much.

Suppose this all started with a few self-replicating proteins. From that we got organelles, and then cells. Then multi-cellular organisms, then tissues, organs, etc.

But working backwards, from protein -> molecular chemistry -> quantum phenomenon, may simply not be the path of least resistance and thus for the overwhelming majority of life in the universe, was not an evolutionary path.

[knzhou]

The answer is a combination of 1 and 2. Not every computing device is actually useful. For example, you can find plenty of brain parts that look vaguely like GPUs or FPGAs. None that look anything like a standard CPU. This would be basically impossible to build out of cells, and not useful anyway.

The same thing applies to quantum computers. They’re much much harder to build because they’re more delicate. We’re talking about effects that usually are completely destroyed by a single unwanted atom coming in and hitting something. And there are a lot of atoms flying around in cells. Propagating any quantum signal from even one cell to an adjacent one is impossible. Finally they’re less useful. I can’t think of problems a biological brain needs to solve that require even a moderately fast CPU, let alone a quantum computer, which provides speedups over the CPU for only certain specific problems.

But none of this really matters, because your comment is one long isolated demand for rigor. You wave away my long list of examples because you think something very distantly related exists (in which case, with those low standards, quantum computers already exist), or because the examples are clearly impossible or not useful (without equally seriously considering the same for quantum computers). This is what I mean by skepticism of QC being driven mostly by intellectual fashion.

[peteradio]

It's 2. There is no way for cell/proteins to maintain a useful small scale quantum state.

[raverbashing]

I think in your list of examples there are plenty of cases where nature, while not building the exact thing uses the same underlying principles.

There are a lot of seeds that are aerodynamic so that they get spread more widely for example. Even seeds that have controlled falls due to "autorotation"

But of course on planet earth at least it is hard for nature to use things requiring too high or low temperatures. Nature doesn't need the Haber process, but it does fixate nitrogen.

[simonh]

Right, so life could make use of quantum transitions that are used in quantum computing, without actually doing or using the results of quantum computations.

[knzhou]

If your standards are low enough that you would say superconductivity appears in nature, then quantum computing certainly does.

[raverbashing]

> But of course on planet earth at least it is hard for nature to use things requiring too high or low temperatures

So, no, superconductivity probably does not exist in nature

[knzhou]

Most candidates for quantum computing also require extremely low temperatures — in fact, lower than the ones for superconductivity.

[kuschku]

> nuclear fission

https://en.wikipedia.org/wiki/Natural_nuclear_fission_reacto...

For pretty much every thing you list, nature has something pretty close to it in it. Even nuclear fission.

[gilleain]

Although natural, that reactor is geological not biological.

It's unlikely that enzymes could catalyse fission reactions. It would be amazing if they could...

[knzhou]

If you really think you can explain all examples, why did you pick one (by far the easiest one) instead of letting me pick? Let’s see some natural examples that use general relativity, superconductivity, the fractional quantum Hall effect, and superfluidity.

[emilfihlman]

Wtf? Non even remotely the same argument.

[teraflop]

> if it were possible, I would have expected evolution to have used it somehow.

That seems reasonable as long as you're consistent in applying the same skepticism to other inventions, such as the internal combustion engine and the wheel.

[Iv]

The physical principles behind combustion engines and the wheel are displayed in many life forms. Their engineering constraints and goals differ.

It is, however, reasonable to expect evolution to find a way to exploit quantum algorithms as it is very useful to several fitness advantages and is something that you would expect to be achievable through protein manipulations.

[aaron_m04]

Bacterial flagella are rotary, so very close to being wheels.

[SolarNet]

Fossil fuels are primarily biological in origin. Certainly we extract the energy in a much different way, but that energy capacity is there biologically (slow burn, turns out high speeds aren't actually that useful in nature).

As for wheels, hip joints are a more efficient design of a rotatory system (multiple degrees of freedom). Wheels are much too simple a design to see much use in nature.

[jyounker]

Not quantum search, but sizable fraction of a second quantum coherence at room temperature:

https://arxiv.org/pdf/0906.3725.pdf

I suspect we’re just really bad at identifying quantum processes in living things.

[quotemstr]

Your argument doesn't stand up to cursory examination. There are all sorts of tricks of physics that evolution never found out how to use. Natural systems don't use the (macroscopic) wheel either, but that works great. Refrigeration works great, but nature doesn't use that either. Complex life has barely been around 540 million years ago. Why would you suppose evolution would have had time to explore the entire space of techniques that physics allows?

[Iv]

Actually, rolling things, from seeds to eggs, including whole shrimps is a locomotion technique used by nature since a long time. The axle is something it does not use, and I would argue does not need. In all natural terrains, legs are a superior option.

The physical principles behind refrigeration are witnessed in several places in nature.

[xenadu02]

Use of the vast majority of the radio spectrum is not used by any form of life that we know of.

Being able to communicate non-visibly without giving away our position audibly would be a huge advantage (until your predators/competitors figured it out).

Life isn't generally suitable to the use of really high energies like X-rays because it damages cells. It isn't suitable for low energies because it is difficult to create an individual receiving/broadcasting element at such a small scale.

There absolutely are physical phenomena that life does not take advantage of.

[SolarNet]

There are sea bacteria that "communicate" in the giga-hertz frequency. But the general point being that large portions of the range are used, like the visible spectrum and the thermal spectrum.

x-rays aren't a unique physical phenomena, just a wave length of electromagnetic waves, which is a physical phenomena that life makes excessive use of (photosynthesis, vision).

[goodmachine]

> There are sea bacteria that "communicate" in the giga-hertz frequency.

Wow. Link?

[SolarNet]

https://ucsdnews.ucsd.edu/pressrelease/can-organisms-sense-v...

[Iv]

Well, if we exclude all the visual spectrum, which is the most useful part of the RF spectrum for animals: transparent to water and atmosphere, reflects well on a lot of surfaces, has kilometers of range if there is line-of-sight, wavelength of a size making it possible to have lens and sensor cells small enough to fit an animal.

But yes, the idea is that if there were a physical principle that would be super useful to animals yet not witnessed in nature, that is something that needs to be explained. There are reasons for evolution to miss a solution.

In the case of quantum computing, I could not really see it: the effects happen at a scale where evolution operates, and could easily be integrated, e.g. in nerve cells to create a workable signal. I thought reasonable to make it an element to feed my skepticism, though not a ultimately strong argument to deny any possibility of quantum computing.

[TheOtherHobbes]

Evolution is a relatively poor search algorithm that tends to get stuck at local maxima. Unlike science, it doesn't build maps and models for exploration and extrapolation.

So far as I know there's no general theory of theories which quantifies this, so there's no way to make predictions about the cut-off point for evolutionary invention.

But in a hand-wavy way, evolution's only feedback loop is first-order and binary - mutate and reproduce at a positive replacement rate, or not.

The feedback loop in science is more complex. Instead of being driven by a random search, "mutations" are guided by a creative model. This creates momentum in the model space which isn't available to evolution - which in turn makes it possible to discover more complex and less immediately accessible solutions.

It also makes it possible to build systems whose value is guaranteed, or at least strongly suspected, before resources are diverted to making them physical.

The bottom line is evolution is only ever going to find a small subset of all possible biological configurations, and that space is going to exclude many features that are available to science-driven search.

(Of course you can argue that scientific meta-search was a product of evolution anyway, so the distinction is academic.)

[josalhor]

The properties of quantum computing require incredible low temperatures. We are talking close to 0 K, not -few degrees. That is unfeasible for nature to develop naturally as the advantages of using a quantum algorithm would be far lower than the natural disadvantages of energy requirements, weight, natural complexity, lack of versatility... On top of all that, quantum computing and algorithms always require some kind of non quantum processing, so on top of that we would include non quantum processing...

[xenadu02]

You seem to be engaging in a "No True Scotsman" argument here.

I could point to MRIs and SQUID, but I guess you could claim some animals sense magnetism (for wayfinding or determining north). I don't consider those to be in the same league but then we're back to arguing semantics under your definitions.

The fact remains that physical phenomena exist that evolution did not discover. There are perfectly good reasons for that, but I don't think it is fair to say if nature doesn't use it then it is "questionable".

[knzhou]

What's the explanation for why living things don't contain superconductors?

Of course: you need an extremely cold and clean environment, which is very hard to generate in a cell. And it wouldn't even be that useful for cells anyway.

The exact same reasoning applies to quantum computers.

[changoplatanero]

lasers would be another one, right? Is there any lifeform that can make a laser?

[Iv]

Is there any lifeform that needs it? With light sensors that basically give us line-of-sight range, I don't see which problem organic lasers would solve.

[wnkrshm]

What single-mode lasers do offer is a means of inconspicuous communication. If you can aim your communications channel at a receiver, you cannot easily be detected.

But life could do similar things with collimating reflectors, even for sound. Are there organisms that use geometry to shape sound to send a signal? I'm only aware of ones that use geometry to boost/tune their reception (e.g. owls).

Edit: Dolphins apparently

[ben_w]

The ability to dazzle an attacker or prey seems useful when one is hunting the other.

[xaedes]

Yes! Humans. So transitively evolution invented lasers and all the other cool stuff mentioned here.

[quotemstr]

Rolling is not the same thing as a wheel and axke. You're playing word games.

Arguing that while nature doesn't might not use some technology, it does do something else that relies on the same principle. Who's to draw the line between a technology and a principle? You can say nature has done anything so long as your definitions are uselessly vague.

[arthurcolle]

You didn't explicitly mention it, but planthoppers use gears (as one example of a "man-made" macroscopic invention) that we then discovered as a locomotion mechanism in this species within Insecta. Not a perfectly round wheel, but a wheel with teeth.

https://www.livescience.com/39577-insects-with-leg-gears-dis...

[chopin]

Sweating uses the same physical principle as your refrigerator and for the same reason.

[api]

A bit of a further tangent but:

I continue to believe that it's very likely that brains somehow leverage quantum computing.

The human brain is able to accomplish computational tasks on only around 40 watts of power that destroy what we can accomplish using pretty much any known machine learning algorithm using tens or even hundreds of thousands of watts of power. Maybe our algorithms are primitive or wrong, or maybe the brain just is not a classical computer.

The power of brains is so "unreasonable" that I've long suspected that there are only three possibilities:

(1) Brains are quantum computers.

(2) P=NP, or for a weaker form perhaps there exist large numbers of undiscovered algorithms that offer massive speed improvements over any currently known algorithm.

(3) Brains or intelligence are "supernatural" or at least tap into something about nature that we fundamentally do not understand.

I think option #1 is by far the most likely, especially given that brains excel most of all at search and quantum computing seems to really be able to speed up search.

There is just no way classical computation can do what the human brain does on 40 watts. To me that strains credulity much more than any of the three options above. It just cannot possibly be so.

[lonelappde]

Illogical. 40watt computers can also do things that destroy what human brains can do. I can load up a raspberry pi with logic to control robots and apply computer vision models and do some pattern learning. And the power of software is still rapidly improving. And computers can run on much lower power too, and likely even lower in future. There is no particular reason to believe that quantum is what makes brains work, as known quantum computers are super energy inefficient, and quantum effects are not observed in high temperature matter (not counting quantum effects of photons) like humans. How could you think our algorithms are not primitive and wrong? They've been improving at exponential rate since almost forever, with no hints that they've reached optimality beyond tiny pockets.

[api]

I think you're ludicrously underestimating what brains and biological neural systems in general can do.

[zamalek]

You'd be absolutely correct if we knew how everything in nature works. We haven't explained everything that evolution has done and QC has been proposed for some of those things, for example the quantum mind hypothesis (where there are good arguments both for and against it).

In a way, your argument makes the same mistakes that it claims these scientists are making.

> strong AI

I'm not sure what you're arguing here; evolution has achieved this in the form of the human mind. I strongly doubt we are as close as some believe to creating SAI, certainly not in the lifetime of any living human being, but arguments that hinge on natural demonstrations require that this is eventually possible.

[e2le]

How do you know it doesn't use it?