Now this conversation is in the top ten out of all 266 podcast so thank you very much for that Mr. Carroll, you are also one of the very-very best scientist /teacher of the world so we audience are just lucky as hell I think! 😉👋
Your information content was a half million bits at the moment of conception. After that there is a slew of new information influencing the body from the environment. Not only stored in neurons but in the entire chemical/metabolic state of the body.
Yup. I think the answer to the "open" question is in fact rather simple: since Dr. Adami relates a growth of information content with an increased fitness one can conclude that if the current version of Dr. Carroll is fitter than it (that is, his version) used to be when he was born/conceived (for instance, now he knows how to use YT and that alone increases the fitness of an individual in the 2024), then his information content is increased as well. Similarly, the information content of the computer with the "BIOS" only is smaller than the content of the same computer with the popular software installed and working.
I really liked this approach to understanding Maxwell's demon. It makes so much more sense than the classic explanations, although it is basically the same answer through a different lens. Also it's amazing that information theory can turn maxwell's demon into an analogy for evolution
If Maxwell's demon could predict which molecules were hot and which were cold, it could open and close the window in anticipation of their arrival. For example, if those molecules came at predictable intervals (hot, cold, hot, cold, etc...), then the demon could set up a simple mechanism to open and close the window at the correct intervals. The microscopic world does not work that way, but the macroscopic world does. Scale the demon up enough and change its task from sorting molecules to sorting objects and the demon resembles no longer a fiction, but an organism in the real world. The miracle of our world is how these stochastic molecular systems become predictable simply through scale.
One reason could be that these stochastic molecular systems becomes more complex with increase in scale in turn giving rise to the possibility of it being the predictor
You are correct. A system with one bit of entropy, like a coin toss, can be in one of two possible states. Therefore, a system with 2 bits of entropy can be in one of four possible states, a system with 3 bits of entropy can be in one of eight possible states, and so on. Thus, a system with half a million bits of entropy can be in any of (2^{500,000}) possible states.
A simple way of thinking of the information in a message is what "new" stuff is conveyed that wasn't already known. For a yes/no question, the response typically contains 1 bit of information (e.g. 1=yes, 0=no), but that is only true if "yes" and "no" are each expected to occur 50% of the time. If the question in a certain context is answered as "no", 99.9% of the time, the answer has much less information, since you are expecting a "no" almost every time.
… as a German Biologist - In First Semester Biology Our Math Professor claimed: “We take a COW - as a Point Symmetric Mass…” Since this Second - I abandoned anything “Mathematical” And have a DEEP Allergy against “Mathematicians” …A “Mathematician” - is a pointless Mass Fits to me...
I am not so sure about this view on evolution, though I enjoyed this conversation. In particular, it seems to me not really true to say „information is that, which evolves“. It is not true that organisms tend to accumulate some sort of predictive power over time, thereby increasing their fitness. To me, this really confounds fitness with prediction power. Many animals with high predictive powers have died out (i.e. have proven to be not „fitting“ into their environment) while others, with much less predictive powers have survived (i.e. have proven to be fitter). For instance, the fittest, i.e. longest survived animal on earth is Hormiphora californensis, a kind of jelly fish that inhabits earth for over 700 Million years (i.e. since long before mammals came into being). Humans, by comparison, have survived maybe 2 or 3 Million years only and face extinction already. I don’t want to put forward a pessimistic agenda but just say that fitness does not equal information or predictive powers. Ofcourse, Hormiphora californensis has taken advantage of a habitat which changes only slowly and little; which translates into: there’s nothing much to know about once you’ve got the basics of underwater survival. Still. Here is a task: Suppose, some kind of turtle evolves which feeds on Hormiphora californensis. It manages to eat basically all the jelly fish before it can adapt; Hormiphora californensis dies out. Consequently, our turtle dies out, too, because of lack of prey. Now: Who of the two was fitter and who had more predictive powers? The jelly fish, which survived 700 Million years; or the turtle, which survived, say, 2 Million years only, but managed to eat all the jelly fish?
While you are correct from a certain point of view, you are missing something. That is, "what information". The information that accumulates as a result of evolution is the information that enhances offspring production and survival, not necessarily "general" information. In the case of humans, we appear to have given up on some directly inherited information in exchange for a brain that is enhanced to form associations - that is, we are good at learning. Learning is sort of meta information - it's generally good under all conditions. However, as you point out, environments can change and since the rate at which evolution can generate change is limited to the rate at which beneficial mutations can be discovered and spread though populations, it's really most effective when change is slow. Unfortunately for us humans (and all the other species that we are impacting) our ability to change the environment is outpacing the rate at which evolution can generate solutions. That we may be changing the state of the environment so rapidly that evolution can not identify and localize the information needed to survive in these modified environments quickly enough to avoid extinction events. To address the turtle and jellyfish question. From the perspective of information in evolution, the jellyfish is well suited to its niche (the particular way that it goes about "making a living") and this niche has been stable for millions and millions of years. If a turtle arises that eats that jellyfish then the niche that the jellyfish occupied would suddenly change, and either something would need to cause it to change back (the turtles hunting the jellyfish to near extinction, resulting in the turtles' extinction), the jellyfish would need to evolve to survive in their new modified niche (that contains predator turtles) or they would go extinct. Just as humans may be changing the world on a global scale at a rate that is faster than evolution, the turtle in your question may result in a change that is too rapid to allow evolution to rescue the jellyfish. By the way, Dr. Adami is my grad advisor, so it may be that I have thought about these sorts of things too much. There is more that could be said, but I'll leave it here.
Consider this: The environment for the jellyfish changed when the turtle was introduced. And the environment for the turtle changed after they had eaten all the jellyfish. Neither was able to adapt fast enough to their changing environment.
Sean, first let me say that my interpretation of quantum mecanics is that at the planck scale we start to look at a event that is at an infinitesimal small point in the future. i know the problematic of defining the 'now' or present moment, but I think space-time is a continuum, so the probabilistic state of quantum physics is all about the future in my mind. Second, have you ever heard of Ian Stevenson? The guy research on reincarnation and found out that the phenomenom of kids remembering past lives was a factual phenomena, as a philosopher and as a scientist, I guess we have to face this fact and try to come up with a better explanation than reincarntion, knowing that reincarnation and the transcending nature of consciousness is a plausible hipotesys. Pardon my english, i'm not used to write in this language, just listen. Hug!
I thought Voynich is written in soul language 39:17 but if you were already capable of doing that then the antidote floods the system? Kind of like suicide if the previous version was not fit enough to put in order the new uptick of information in the environment?
I haven't the knowledge in biology to know, but against all odds I guess that could go with the work hypothesis that your body is solely defined by DNA when investigating anything but the brain connectome. Wouldn't DNA try to keep an "information homeostatis" so it can continue being useful as a control mechanism? I dont know how this mindset helps biologists, which is what models and mindsets and all about, but ok. Whatever suits them. I quite enjoyed the lack of subtlety of the interviewed, when he did not recognize tongue-in-cheek in Sean's tone. He must have been really in geeman serious mode, and a bit nervous. Bless his heart.
Very unimpressed with this guest. Keeps going back to biology while seemingly lacking even basic knowledge of it. The first example with the dna completely ignores the huge part that the environment plays in a person's development, including their body, from conception and potentially even earlier, and other aspects of how dna gets expressed. Sean poked a very obvious hole in this, and then he has to "go away and think about it."
I think the thrust of his argument still holds. That "Darwin's demon" in a sense works to encode information in the genome of living organisms relative to the fitness environment they happen to exist in. But when he said "I can tell you how much ordered state there is. It's about half a million bits" the qualifiers he put on that statement were insufficient. It's not enough information to fully describe how far the molecules of a person are "from an equilibrium state". At most it is an approximate description of that persons evolutionary fitness. Those 500,000 bits also don't encode for the invention of writing.
With respect, near the end of the episode believe that this was clarified -- 8% of the human genome codes to proteins with about 3GB of raw nucleotide to start. The awkward back-and-forth was the gentle suggestion that this is closer to 240MB, not KB. Agree with you that our minds and epigenetic factors are important: they make us individuals; but to quote Prof Adami "it is our genetic code that keeps us alive" and is the focus of his statistical work. My understanding of the homework excise was to research how much of the brain and bodily development is emergent from DNA. My homework is to buy and read his book.
The tension between genetic influence and environmental influence is a really important point. However, the degree to which environment can have an effect is mediated by the fact that the genome is crucial for the development and operation of an organism. So even if it were the case that the form and behavior of an organism was 99% determined by environment, that 1% that is determined by genetics would still be critical to lay the groundwork. And this is where evolution acts to encode information - even if that information is mostly about how to interact with one's environment.
Now this conversation is in the top ten out of all 266 podcast so thank you very much for that Mr. Carroll, you are also one of the very-very best scientist /teacher of the world so we audience are just lucky as hell I think! 😉👋
Liked it very much! The guest was very fun to listen to. Cheers !
Your information content was a half million bits at the moment of conception. After that there is a slew of new information influencing the body from the environment. Not only stored in neurons but in the entire chemical/metabolic state of the body.
Yup. I think the answer to the "open" question is in fact rather simple: since Dr. Adami relates a growth of information content with an increased fitness one can conclude that if the current version of Dr. Carroll is fitter than it (that is, his version) used to be when he was born/conceived (for instance, now he knows how to use YT and that alone increases the fitness of an individual in the 2024), then his information content is increased as well.
Similarly, the information content of the computer with the "BIOS" only is smaller than the content of the same computer with the popular software installed and working.
I really liked this approach to understanding Maxwell's demon. It makes so much more sense than the classic explanations, although it is basically the same answer through a different lens.
Also it's amazing that information theory can turn maxwell's demon into an analogy for evolution
Great talk - thank you!
I really needed this conversation!
If Maxwell's demon could predict which molecules were hot and which were cold, it could open and close the window in anticipation of their arrival. For example, if those molecules came at predictable intervals (hot, cold, hot, cold, etc...), then the demon could set up a simple mechanism to open and close the window at the correct intervals. The microscopic world does not work that way, but the macroscopic world does. Scale the demon up enough and change its task from sorting molecules to sorting objects and the demon resembles no longer a fiction, but an organism in the real world.
The miracle of our world is how these stochastic molecular systems become predictable simply through scale.
One reason could be that these stochastic molecular systems becomes more complex with increase in scale in turn giving rise to the possibility of it being the predictor
Thank you! This deserves a lot more views!
perfect episode
I love Bauman! I used to live in NYC and in Vegas so I've been to both. Never expected that to get a shoutout here.
12:45 half a million bits, *interacting with each other*, so half a million is the exponent, not the base...? that could if would explain that.
You are correct. A system with one bit of entropy, like a coin toss, can be in one of two possible states. Therefore, a system with 2 bits of entropy can be in one of four possible states, a system with 3 bits of entropy can be in one of eight possible states, and so on. Thus, a system with half a million bits of entropy can be in any of (2^{500,000}) possible states.
A simple way of thinking of the information in a message is what "new" stuff is conveyed that wasn't already known. For a yes/no question, the response typically contains 1 bit of information (e.g. 1=yes, 0=no), but that is only true if "yes" and "no" are each expected to occur 50% of the time. If the question in a certain context is answered as "no", 99.9% of the time, the answer has much less information, since you are expecting a "no" almost every time.
… as a German Biologist -
In First Semester Biology
Our Math Professor claimed:
“We take a COW -
as a Point Symmetric Mass…”
Since this Second -
I abandoned anything
“Mathematical”
And have a DEEP Allergy against
“Mathematicians”
…A “Mathematician” -
is a pointless Mass
Fits to me...
I am not so sure about this view on evolution, though I enjoyed this conversation. In particular, it seems to me not really true to say „information is that, which evolves“. It is not true that organisms tend to accumulate some sort of predictive power over time, thereby increasing their fitness. To me, this really confounds fitness with prediction power. Many animals with high predictive powers have died out (i.e. have proven to be not „fitting“ into their environment) while others, with much less predictive powers have survived (i.e. have proven to be fitter). For instance, the fittest, i.e. longest survived animal on earth is Hormiphora californensis, a kind of jelly fish that inhabits earth for over 700 Million years (i.e. since long before mammals came into being). Humans, by comparison, have survived maybe 2 or 3 Million years only and face extinction already. I don’t want to put forward a pessimistic agenda but just say that fitness does not equal information or predictive powers. Ofcourse, Hormiphora californensis has taken advantage of a habitat which changes only slowly and little; which translates into: there’s nothing much to know about once you’ve got the basics of underwater survival. Still.
Here is a task: Suppose, some kind of turtle evolves which feeds on Hormiphora californensis. It manages to eat basically all the jelly fish before it can adapt; Hormiphora californensis dies out. Consequently, our turtle dies out, too, because of lack of prey. Now: Who of the two was fitter and who had more predictive powers? The jelly fish, which survived 700 Million years; or the turtle, which survived, say, 2 Million years only, but managed to eat all the jelly fish?
While you are correct from a certain point of view, you are missing something. That is, "what information". The information that accumulates as a result of evolution is the information that enhances offspring production and survival, not necessarily "general" information. In the case of humans, we appear to have given up on some directly inherited information in exchange for a brain that is enhanced to form associations - that is, we are good at learning. Learning is sort of meta information - it's generally good under all conditions. However, as you point out, environments can change and since the rate at which evolution can generate change is limited to the rate at which beneficial mutations can be discovered and spread though populations, it's really most effective when change is slow. Unfortunately for us humans (and all the other species that we are impacting) our ability to change the environment is outpacing the rate at which evolution can generate solutions. That we may be changing the state of the environment so rapidly that evolution can not identify and localize the information needed to survive in these modified environments quickly enough to avoid extinction events.
To address the turtle and jellyfish question. From the perspective of information in evolution, the jellyfish is well suited to its niche (the particular way that it goes about "making a living") and this niche has been stable for millions and millions of years. If a turtle arises that eats that jellyfish then the niche that the jellyfish occupied would suddenly change, and either something would need to cause it to change back (the turtles hunting the jellyfish to near extinction, resulting in the turtles' extinction), the jellyfish would need to evolve to survive in their new modified niche (that contains predator turtles) or they would go extinct. Just as humans may be changing the world on a global scale at a rate that is faster than evolution, the turtle in your question may result in a change that is too rapid to allow evolution to rescue the jellyfish.
By the way, Dr. Adami is my grad advisor, so it may be that I have thought about these sorts of things too much. There is more that could be said, but I'll leave it here.
Consider this: The environment for the jellyfish changed when the turtle was introduced. And the environment for the turtle changed after they had eaten all the jellyfish. Neither was able to adapt fast enough to their changing environment.
Caught it on Spotify!! ❤️❤️
An info dense episode 👍👍👍
Bom episodio!
Gostei muito.
Sean, first let me say that my interpretation of quantum mecanics is that at the planck scale we start to look at a event that is at an infinitesimal small point in the future. i know the problematic of defining the 'now' or present moment, but I think space-time is a continuum, so the probabilistic state of quantum physics is all about the future in my mind. Second, have you ever heard of Ian Stevenson? The guy research on reincarnation and found out that the phenomenom of kids remembering past lives was a factual phenomena, as a philosopher and as a scientist, I guess we have to face this fact and try to come up with a better explanation than reincarntion, knowing that reincarnation and the transcending nature of consciousness is a plausible hipotesys. Pardon my english, i'm not used to write in this language, just listen. Hug!
I thought Voynich is written in soul language 39:17 but if you were already capable of doing that then the antidote floods the system? Kind of like suicide if the previous version was not fit enough to put in order the new uptick of information in the environment?
MORE INFORMATION IS STORED WITHIN THE ELECTRICAL FIELDS WITHIN CELLS... (;
I haven't the knowledge in biology to know, but against all odds I guess that could go with the work hypothesis that your body is solely defined by DNA when investigating anything but the brain connectome. Wouldn't DNA try to keep an "information homeostatis" so it can continue being useful as a control mechanism?
I dont know how this mindset helps biologists, which is what models and mindsets and all about, but ok. Whatever suits them.
I quite enjoyed the lack of subtlety of the interviewed, when he did not recognize tongue-in-cheek in Sean's tone. He must have been really in geeman serious mode, and a bit nervous. Bless his heart.
memes
Entanglement is geometry is information is gravity is thermodynamics.
Very unimpressed with this guest. Keeps going back to biology while seemingly lacking even basic knowledge of it. The first example with the dna completely ignores the huge part that the environment plays in a person's development, including their body, from conception and potentially even earlier, and other aspects of how dna gets expressed. Sean poked a very obvious hole in this, and then he has to "go away and think about it."
I think the thrust of his argument still holds. That "Darwin's demon" in a sense works to encode information in the genome of living organisms relative to the fitness environment they happen to exist in.
But when he said "I can tell you how much ordered state there is. It's about half a million bits" the qualifiers he put on that statement were insufficient.
It's not enough information to fully describe how far the molecules of a person are "from an equilibrium state". At most it is an approximate description of that persons evolutionary fitness.
Those 500,000 bits also don't encode for the invention of writing.
With respect, near the end of the episode believe that this was clarified -- 8% of the human genome codes to proteins with about 3GB of raw nucleotide to start. The awkward back-and-forth was the gentle suggestion that this is closer to 240MB, not KB. Agree with you that our minds and epigenetic factors are important: they make us individuals; but to quote Prof Adami "it is our genetic code that keeps us alive" and is the focus of his statistical work. My understanding of the homework excise was to research how much of the brain and bodily development is emergent from DNA. My homework is to buy and read his book.
The tension between genetic influence and environmental influence is a really important point. However, the degree to which environment can have an effect is mediated by the fact that the genome is crucial for the development and operation of an organism. So even if it were the case that the form and behavior of an organism was 99% determined by environment, that 1% that is determined by genetics would still be critical to lay the groundwork. And this is where evolution acts to encode information - even if that information is mostly about how to interact with one's environment.
The guy just uses new terminology for a known stuff and says it's new science lmao