Should Physicists Study the Question: What is Life? (msn.com) 89
An astrophysicist at the University of Rochester writes that "many" of his colleagues in physics "have come to believe that a mystery is unfolding in every microbe, animal, and human." And it's a mystery that:
- "Challenges basic assumptions physicists have held for centuries"
- "May even help redefine the field for the next generation"
- "Could answer essential questions about AI."
In short, while physicists have favored a "reductionist" philosophy about the fundamental laws controlling the universe (energy, mattery, space, and time), "long-promised 'theories of everything' such as string theory, have not borne significant fruit: There are, however, ways other than reductionism to think about what's fundamental in the universe. Beginning in the 1980s, physicists (along with researchers in other fields) began developing new mathematical tools to study what's called "complexity" — systems in which the whole is far more than the sum of its parts. The end goal of reductionism was to explain everything in the universe as the result of particles and their interactions. Complexity, by contrast, recognizes that once lots of particles come together to produce macroscopic things — such as organisms — knowing everything about particles isn't enough to understand reality...
Physicists have always been good at capturing the essential aspects of a system and casting those essentials in the language of mathematics... Now those skills must be brought to bear on an age-old question that is only just getting its proper due: What is life? Using these skills, physicists — working together with representatives of all the other disciplines that make up complexity science — may crack open the question of how life formed on Earth billions of years ago and how it might have formed on the distant alien worlds we can now explore with cutting-edge telescopes. Just as important, understanding why life, as an organized system, is different at a fundamental level from all the other stuff in the universe may help astronomers design new strategies for finding it in places bearing little resemblance to Earth. Analyzing life — no matter how alien — as a self-organizing information-driven system may provide the key to detecting biosignatures on planets hundreds of light-years away.
Closer to home, studying the nature of life is likely essential to fully understanding intelligence — and building artificial versions. Throughout the current AI boom, researchers and philosophers have debated whether and when large language models might achieve general intelligence or even become conscious — or whether, in fact, some already have. The only way to properly assess such claims is to study, by any means possible, the sole agreed-upon source of general intelligence: life. Bringing the new physics of life to problems of AI may not only help researchers predict what software engineers can build; it may also reveal the limits of trying to capture life's essential character in silicon.
- "Challenges basic assumptions physicists have held for centuries"
- "May even help redefine the field for the next generation"
- "Could answer essential questions about AI."
In short, while physicists have favored a "reductionist" philosophy about the fundamental laws controlling the universe (energy, mattery, space, and time), "long-promised 'theories of everything' such as string theory, have not borne significant fruit: There are, however, ways other than reductionism to think about what's fundamental in the universe. Beginning in the 1980s, physicists (along with researchers in other fields) began developing new mathematical tools to study what's called "complexity" — systems in which the whole is far more than the sum of its parts. The end goal of reductionism was to explain everything in the universe as the result of particles and their interactions. Complexity, by contrast, recognizes that once lots of particles come together to produce macroscopic things — such as organisms — knowing everything about particles isn't enough to understand reality...
Physicists have always been good at capturing the essential aspects of a system and casting those essentials in the language of mathematics... Now those skills must be brought to bear on an age-old question that is only just getting its proper due: What is life? Using these skills, physicists — working together with representatives of all the other disciplines that make up complexity science — may crack open the question of how life formed on Earth billions of years ago and how it might have formed on the distant alien worlds we can now explore with cutting-edge telescopes. Just as important, understanding why life, as an organized system, is different at a fundamental level from all the other stuff in the universe may help astronomers design new strategies for finding it in places bearing little resemblance to Earth. Analyzing life — no matter how alien — as a self-organizing information-driven system may provide the key to detecting biosignatures on planets hundreds of light-years away.
Closer to home, studying the nature of life is likely essential to fully understanding intelligence — and building artificial versions. Throughout the current AI boom, researchers and philosophers have debated whether and when large language models might achieve general intelligence or even become conscious — or whether, in fact, some already have. The only way to properly assess such claims is to study, by any means possible, the sole agreed-upon source of general intelligence: life. Bringing the new physics of life to problems of AI may not only help researchers predict what software engineers can build; it may also reveal the limits of trying to capture life's essential character in silicon.
Viruses (Score:5, Insightful)
Just try and decide if a virus is alive or not.
Re: (Score:2)
it's a philosophical/epistemicological question. physicists have often dwelled on it but that's still not "physics".
Re:Viruses (Score:5, Informative)
Just try and decide if a virus is alive or not.
it's a philosophical/epistemicological question. physicists have often dwelled on it but that's still not "physics".
"Epistemicological?" Anyway...
Generally, viruses are not considered to be alive, by scientists who study them (a.k.a. virologists.) They have no metabolic function (i.e., they're dormant for the most part) and they need other organisms to reproduce.
And if physicists have dwelt on whether viruses are alive, then it's no more than what other scientists who aren't virologists have done.
Re:Viruses (Score:5, Interesting)
are not considered to be alive
exactly. that's unrelated to the laws of physics. it is epistemology; it's a subjective definition based on how we decide to classify stuff. different people will classify it differently, and although there may be commonly accepted requirements there is no such thing as a universal concept of "life" with a compelte set of hard and unquestionable boundaries, nor will there ever be. you may say that viruses are alive or that they are not, both works in different contexts, because they do exhibit features that we attribute to life. in any context where those features are relevant (and maybe others not so much), you can consider them alive. a virologist may disagree, and that's fine.
then it's no more than what other scientists who aren't virologists have done.
indeed, but tfa literally quotes a(n astro)physicist inviting his colleagues in physics to philosophize. now, i haven't rtfa, but if his intent is in producing a (better) set of parameters to search for life in the universe more efficiently that might be interesting, but that is not really what life "is".
Re: (Score:3)
[The consideration of what "alive" means] is epistemology; it's a subjective definition based on how we decide to classify stuff.
Epistemology is the study of knowledge, not classification.
[Y]ou may say that viruses are alive or that they are not, both works in different contexts, because they do exhibit features that we attribute to life.
Over-attribution of such "features" can lead to nonsense, such as animism. [wikipedia.org] There are some scientists who think viruses are alive (or are a different kind of life) because they have genetic material, can adapt and replicate, and in some cases may show evidence of bacterial ancestry. But these scientists are in the minority, at least for now.
[If the astrophysicist's] intent is in producing a (better) set of parameters to search for life in the universe more efficiently that might be interesting, but that is not really what life "is".
Sure. And still on viruses: the ability to detect cosmic viruses may help, because even if you don't think viru
Re: (Score:2)
Re: Viruses (Score:3)
Epistemicological is the study of the nature of knowledge that occurs spontaneously when you consume the right type of mushrooms
Re: (Score:2)
Sometimes, a question like this can span multiple disciplines.
If you feel hot, for example, it might be because the temperature is high, or it might be because your skin is in contact with hot peppers. The sensation of heat is a feeling, but it also involves biological processes and physics. We know that heat is not just a figment of the brain because it can be measured and it affects the state of matter.
The concept of life might be a philosophical / epistemological question. But the process of life also in
Re: (Score:3)
Reminds me of the question of whether Pluto is a planet. The terms "planet" and "alive" are simply words that we define; they do not necessarily correspond to real distinctions in nature. And these two terms in particular fall into that category of man-made definitions that don't necessarily correspond to real categories.
Another example is "cloud": is fog a cloud? Is the condensation when I exhale on a cold day a cloud? Is the iron (or titanium?) vapor on planets that are really close to their stars clo
Re: (Score:2)
Re: (Score:2)
Prions (Score:2)
Viruses are arguably parasites, which are generally considered "life". An even trickier are prions. They have even been caught evolving (adapting), so "just a protein" doesn't hold water (no pun intended).
Re: (Score:1)
Correction: "An even trickier case is prions"
Re:Are they stupid? (Score:5, Insightful)
It is common that people who devotedly study science don't also study philosophy. As such, there is a lot of misunderstanding among scientists (academic and/or professional) as to what philosophy is and, relevantly, what kind of material has already been thoroughly covered.
This sounds like a group of scientists are simply wandering into philosophical territory, and could benefit from an education on that front with emphasis on such questions of life and consciousness.
Re: (Score:2)
There is more to the Universe than mass and electromagnetic fields. "What is life" is a basic question that can lead you towards something other than mass and electromagnetic fields. In short, it sounds like physicists may be trying to find things that they can not see with any microscope or sensor that currently exists.
Re:Are they stupid? (Score:4, Interesting)
First, figure out how slime mold [science.org] can generate a reasonably efficient transportation network model. And figure out how much intelligence that must take.
Or is this just what they do when they need to write grant proposals?
Perhaps they need to hire the slime mold.
Mod parent UP. (Score:2)
Excellent example.
Yes (Score:2)
Re: Yes (Score:3)
Re: Yes (Score:5, Insightful)
False dichotomy.
They are doing Philosophy of Science.
Which is philosophy, not science.
Re: (Score:2)
LLMs are not concious (Score:2)
For a start conciousness requires something to be going on in the waking mind all the time even if just self reflection or pondering the day. Precisely nothing happens in an LLM unless it is attempting to provide output to some input whether it be a text question, image or some other data. When someone can show an LLM thinking or dreaming (whatever the definitions may be) when idle then I might change mind. Until then...
Re: (Score:3)
Attempting to provide output to some input isn't such a bad definition of "waking."
The big thing that defines the last couple iterations of "LLMs" is feeding back their output internally. You might call it "reflection."
I'm not saying LLMs are conscious. I'm saying your definition is as vague a
Re: (Score:2)
Computers are useless. They can only give you answers. - Pablo Picasso
Re: LLMs are not concious (Score:2)
That feedback only happens when they're processing input. No input- no processing. Of any kind. That's not the case with the human brain unless you think deaf and blind people have no inner life.
Re: (Score:2)
Nonsense. You can stimulate an LLM or any other neural network however you want. Generative models in particular generate random noise internally. That's what makes them generative. You can certainly do that in the absence of input conditioning.
The reasoning models I mentioned start with input but feed it back repeatedly and reprocess it in the absence of input.
Re:LLMs are not concious (Score:5, Interesting)
The word "conscious" (and the family of related words) is sloppily defined. This is not a defect. In fact, it is a powerful feature of our language (and our mental abilities that allow us to process language) that we can operate really well with sloppily-defined concepts. It allows for very speedy information exchange on very practical matters (especially useful when in combat or other emergency situations).
But this same feature makes in-depth analysis difficult, especially when it presents logical traps (fallacies) that we can innocently fall into.
Our commonsense understanding of consciousness is rooted in our very practical need to quickly divide up the world of our experience into the categories of "conscious" and "not-conscious." Rocks, clouds, the wind, shadows....all not conscious. People, wild animals, divine beings (to the extent that one believes in them) all conscious. The point here is that the sloppiness of the definition is rooted in a stark practical reality for us: we interact differently with conscious beings than with inert matter, so we need to be able to make very quick snap-judgments about which is which.
For most of the history of our existence, this was enough. We just tossed plants over in the "not conscious" group and ran with it. Computers, too, went right into the "not conscious" group, and that was good enough.
This commonsense idea of consciousness is not very helpful when we dive deeply into the edge cases, especially the ones that are new in the history of our species. As AI becomes more sophisticated, we wind up with something that has elements common to both categories (its a metalic/plastic construct, so generally not conscious, but it can engage with us in lucid dialog and solve engineering problems and so on, so generally conscious).
We aren't going to be able to resolve this dilemma with what we have on-hand. Our basic intuition about what consciousness is does not give us a clear answer (and it will just become more fuzzy as the tech improves), and further scientific research is hard to do properly since such research must begin with clear and unambiguous definitions (which we don't have, for the reasons given above).
So, for now, it is still easier to toss these things in the "not conscious" bucket and move on, but if the hopes and dreams of interested parties come true, it will become a lot more difficult to do so in the near future.
All the same applies to the word "life" incidentally. And though we can clearly have things that are alive but not conscious (such as a human in a coma), AI raises the interesting question of whether we can ever have something that is conscious but not alive. At this point though, such a discussion is entirely semantics without substance. However, in the future, if the tech actually does make the leaps we hope fore, that discussion will become more poigniant.
Re: (Score:2)
For a start conciousness requires something to be going on in the waking mind all the time even if just self reflection or pondering the day. Precisely nothing happens in an LLM unless it is attempting to provide output to some input whether it be a text question, image or some other data. When someone can show an LLM thinking or dreaming (whatever the definitions may be) when idle then I might change mind. Until then...
This is easily solved by ignoring EOS tokens. 100% guaranteed to make any LLM lose its shit.
Re: (Score:2)
You are correct. LLMs are conscious in exactly the same way that movie characters on a screen are conscious. Both display attributes that convincingly portray consciousness, but both are 100% illusions, just patterns (of pixels or language tokens).
Re: (Score:2)
That is a good point, but doesn't cover the issue in full. First thing is, that a modern neural network has many layers. Maybe we can call something that emerges at layer 5 life, which develops until layer 40, before it dies with the output?
Second, neural networks can have a continuity in time. People often parrot the "stochastic parrots" metaphor like a LLM would be a better dice roll for the next token, but in fact it has something called "context" that serves as a memory of what happened so far and is co
Not exactly new (Score:1)
This idea among physicists isn't exactly new or revolutionary or even controversial. The post misstates the proposition. It's not that the systems are more than the sum of their parts. They are just the sum of their parts and if we knew everything about all of the parts we could predict the system exactly. It's that the Uncertainty Principle says that beyond a certain limit we can't know everything about the individual parts and so our "predictions" become more and more just random guesses and the system ap
Re: (Score:2)
Various scientists who are slipping off the deep end love to conflate the phenomenon of weak emergence with strong. You describe weak emergence and we see it all the time. Strong emergence claims that some of those emergent phenomenon are actually magic.
Of course, they have to use things we don't fully understand as examples. Physicists have not yet put together an air tight theory of things we can't observe, most of them not even in principle? Maybe it's magic! Consciousness? Magic! Life? Magic!
It's like t
Re: (Score:2)
No, the particle has a momentum and it's knowable. It's just trying to measure both at the same time that fails: the more accurately you measure one, the less accurately you can measure the other at the same time. No matter what you do, beyond a certain limit your measurements have an error in them you can't reduce which throws any predictions based on the state of the system off.
Ignoring the AI spiel (Score:2)
Re: (Score:2)
Tell that to the anaerobic bacteria.
Re: Ignoring the AI spiel (Score:2)
Re: (Score:2)
Define "respiration". If your definition includes something along the lines of "perform gas exchange with the environment" - then there are many organisms alive that don't do that.
Re: Ignoring the AI spiel (Score:2)
Re: Ignoring the AI spiel (Score:2)
Re: (Score:2)
Let's try something else (Score:2)
How about we start with something a bit more simple than what is life? Why not what is love [youtube.com]?
Re: (Score:2)
No (Score:2)
Why are 'elementary' particle as they are ? (Score:1)
Re: Why are 'elementary' particle as they are ? (Score:1)
Gonna get into theology pretty soon that way, if you're not there already.
Good science isn't about "why" in the sense of "why are the elementary particles the way they are" but rather about "how" and "what" in the sense of "how do they behave" and "what can be predicted about the future given measurements in the past."
The "how" question is a predicate to the "what" question, and the "why" question is sometimes a source of helpful suggestions for ways to proceed and sometimes a philosophical and theological
Re: (Score:2)
Instead of these philosophical musings, I think physics should be questioning WHY the standard model is as it is, what physics could be under the model to give us its "elementary" particles.
The problem is the universe does not owe us an explanation, and Godels theorem proves there are truths within a system of rules that can’t be explained by or derived from that system of rules. Each second, new light from the edge of the visible universe horizon reaches us, but even if we traveled back along that line at the speed of light wouldn’t ever be able to reach that point due to the expansion of the universe so it’s clear there are vast regions hidden from view and kept from expe
What is love? (Score:2)
Don't hurt me no more
Baby, don't hurt me
Don't hurt me no more
Re: (Score:2)
Life is life! Nana-nanana!
When we all give the power
We all give the best
Every minute of an hour
Don't think about the rest
And you all get the power
You all get the best
When everyone gets everything
And every song everybody sings
Life is life!
Should biologists study: What is dark matter? (Score:3)
No. Or at least, if they want to, switch to the field of science which studies it.
This seems like the academic equivalent of, "Sales of sugary drinks are down. Should Coca-Cola start building self-driving cars?"
No, and the question is silly. (Score:5, Interesting)
Re: (Score:2)
Life is not definable in terms of mass and electromagnetic fields... and yet, that is what physics is currently. In other words, there is more to the Universe than mass and electromagnetic fields... so how does one find that 'more'? We have evidence that our Universe is far more interesting than the two axes of mass/electromagnetism and space/time.
Physicists should study the physical world (Score:2)
This guy is just playing word games (Score:2)
Calling something that's perfectly valid but not physics "physics" and then saying "physicists" should do it.
It's not physics. It's applied physics which is what all the differently named sciences are already. Chemistry, for example, is just applied physics. And so is whatever this guy is trying to rename to physics.
BIOLOGISTS. (Score:2)
Don't you just love it when a specialist decides that hey, they are the geniuses and starts telling scientists in an entirely different field how to do their job?
You know what, Biologists should start telling Physicists how stars work.
Yes, I understand the problem is life as we do not know it and therefore do not understand or know much about.
But just as the physicists want to lead this discussion, it is the BIOLOGISTS that should be leading it. They are the ones that actually understand the real basics of
Distant worlds? (Score:1)
Re: (Score:2)
LOL. There is zero evidence life happened here more than once in 4 billion years
The strange structure/composition of the ribosome is definitely evidence that some other kind of life existed at one time. It's not conclusive by any means, but it's evidence.
Yes But Not Discover--Rather Define. (Score:2)
The argument of whether or not a virus is alive or a paramecium, for example, is absurd.
It's not something that is or isn't, it's how do you want to use the word?
To me, this calls for either multiple words or one word with types thereof or with parameters (my preference).
Biologists might say life is an autonomous entity that replicates itself with a small number of mutations.
This makes sense in terms of evolutionary life, of course.
Still, you have to ask what constitutes an "entity"--perhaps a set of spatia
So... chemistry (Score:2)
Re: (Score:2)
Most of modern "chemistry" is quantum mechanics and either does QM directly or some adaptations of the modeling methods developed in the 1950 and 1960s for nuclear physics, e.g. the Hartree-Fock approximation of the Schrodinger equation. Moreover, with the development of femtochemistry, it is becoming possible to use classical electrodynamics to describe the chemical processes.
In other words, the old, 18th century boundaries between the sciences have long ago vanished.
Why stop with physicists? (Score:2)
Physicists are just exploring a mix of applied and theoretical mathematics. Get closer to the source and just ask mathematicians to define life. They're just as likely to understand chemistry and biology, and in a much better position to create and explore novel system models.
Biology is an emergent phenomena of chemistry. Chemistry is an emergent phenomena of physics. The study of where biology emerges from chemistry is best pursued by biologists and chemists, not physicists. Their area of expertise is
Re: (Score:2)
Physicists are just exploring a mix of applied and theoretical mathematics.
Umm, no, physics isn't based in math, it is using math as an expressive and clear language, but is still concerned with what happens in the real world. It may turn out it is pure math, but it hasn't so far.
There are no such "arbitrary" limitations for the mathematician, so their work isn't really physics.
There is "Metaphysics" for those... (Score:2)
The term "life" can NEVER be defined (Score:2)
Re: (Score:2)
You could not be more wrong. The definition of "life" is so banal that even the US constitution gets it right - it is the incessable pursuit of happiness manifested by an appropriate agent even when personified by an 'orrible cunt.
Re: (Score:2)
Hi - did you read my article?
What is "happiness"?
Be careful - the criteria might be met by a AI bot...
Re: (Score:2)
No, I didn't. Can't you summarize it succinctly in one short sentence so that even a child can understand the point, per that fake Einstein quote?
Happiness is the state where you want the time to stop so that the moment will go on forever. As the poet called it,
Then, to the Moment I’d dare say:
‘Stay a while! You are so lovely!’
Every living thing know what I'm talking about, the chatbot has no idea.
Re: (Score:2)
"Every living thing know what I'm talking about, the chatbot has no idea."
You are referring to sentient. That is not how scientists try to define life. Not all life is sentient - e.g. a bacterium is not sentient.
Re: (Score:2)
You are referring to sentient.
Every life is "sentient" in some way or other. You can easily tell if a plant is happy or not, for example, if you understand happiness. It signals plenty.
That is not how scientists try to define life.
And why would I care about how a "scientist" defines life?
Moreover, how do you define a "scientist"?
I have an anecdotal definition. A scientist decides to study the frog. He puts the frog on a table, and tells it, "Jump!". The frog jumps, the scientist dutifully records: "Frog, 4 legs, jumps 1 meter". Then he cuts one leg off, tells the frog "Jump!". It j
Re: (Score:2)
Re: (Score:2)
Sorry, but you clearly do not understand what science is.
Sorry, but the opinion of a nobody on slashdot isn't a measure whether someone "understands" what science is or not, especially coming from a rather ignorant troll hobbyist with nothing to show for in science. Understanding of science is judged by other means, and while I won't flash badges, I'm quite confident I'm a bit ahead of you there :)
And the claim that "every life is sentient" is circular because you are saying that all life is sentient and sentience is life.
Which life isn't sentient, darling?
Finally, there is no evidence that primitive life, such as a bacterium, is sentient - any such claim is an absurd conjecture.
Aaaah... Is that so? Sentience is the ability to have feelings and sensations. There is plenty of evidence that bacteria have some se
Re: (Score:2)
"Sorry, but the opinion of a nobody on slashdot isn't a measure whether someone "understands" what science is or not"
I am a scientist, with masters degrees from an Ivy League school in both nuclear science and operations research. Your comment that science is just a hobby is dismissive, arrogant, and ignorant. If not for science, we would not have cellphones, cars, or any other modern conveniences. Science is based on a combination of theory and experimental proof.
Science is performed by humans, who are fla
Easy... (Score:2)
Prior Debate Art (Score:1)
C2 wiki used to have long arguments over this. [c2.com]
Understanding intangibles is difficult (Score:2)
There's the Newtonian world, which we can experience and intuitively understand with our five senses.
There's the quantum world, which defies intuition [yale.edu], and we cannot directly experience.
There's the information world, which we do experience, but with information being intangible, have difficulty thinking about.
Light. Time. Gravity. We experience them but they are intangible. Information is the same way.
"How to make a bundt cake" is information. Does it exist only in the context of life? Did it always exist,
We Know What Life Is (Score:2)