Arithmetic coding permits one to squeeze an infinite number of finite precision parameters into a single real number. Nor is this mere pedantry in the information age. Statistics has a veritable zoo of so-called information criteria for model selection. But what do they mean by information? How do you quantify it?
Some say bits.
But how do you quantify the number of bits in a model?
Ultimately you want to compute predictions to test your hypotheses. No one has a universal turing machine because it has an infinite tape. All we finite beings of the computer age have are computers.
This conveniently provides us with a way to quantify the number of bits in a model including the residual errors made in predicting observations that are under consideration in the model:
The size of the executable archive of the data under consideration.
This is a principled the way of comparing models of the same data set.
People don't want to do this for various reasons. One reason is that it's more difficult to come up with an executable archive of your data than it is to heaven more or less informal description of the model and statistical characterizations of the errors. Another more nefarious reason is that people don't want to be held to a gold standard rigorous accounting of how bad their models are. These two factors work hand in hand to deprive us of a fundamental advance in the scientific method that could bridge the gap between the grotesque centralization of wealth and power on the one hand and those of us out here in the Heather trying to come up with explanations for what's going wrong with society.
All the well intentioned people with all that centralized wealth of power need to do, since they are so far removed from the actual theories to be able to Judge who should be funded and who shouldn't be funded, is get all the so-called experts to agree on a set of data that they all agree to compete on and then provide funding to those who best compress it.
The way you separate the sheep from the goat some experts is to see which of them most strongly resents this.
This makes me think about the big bang and age of the universe questions. Part of my journey into philosophy/history of science was due to young-earth creationism. The issues with young-earth scientific theories notwithstanding, their criticisms of the standard scientific account seem valid and remind me of many of the issues discussed in this post. For example, my understanding is that the Big Bang model is "wrong" because there is substantially less mass in the universe than necessary in order for the Big Bang to produce our universe. Solution? Add "dark matter" and "dark energy" (nb: epicycles) to make the calculations come out correctly.
In many ways, I wonder if the "scientific method" as popularly promulgated is actually responsible for a lot of these issues. Start with a hypothesis, run an experiment, and then either confirm or falsify the hypothesis. But in many ways this is backwards. This seems like starting with the answer already in mind. Running lots of experiments just to see what will happen and then trying to generalize the results seems much more reliable.
Oh awesome. Thanks for sharing! I really appreciate what you’re doing even though a lot of it is over my head. Ultimately I think the skepticism of modern philosophy means “science” has replaced philosophy proper, and so a lot of these theories end up being lots of brilliant math and experiments cobbled together with really horrible philosophy. Sometimes one gets the sense that many of these scientists resent the demand that their theories actually make sense.
If your theory is all about exciting oscillators(standard model), and all your experiments involve exciting oscillators (particle accelerator experiments) what are you really learning?
The reasoning is somewhat like this: QED gave us the nuclear age, therefore it is the science of the future, and we need enormous particle accelerators to see the fundamental processes of reality. However, Einstein won the Nobel prize, not for relativity, but for the photoelectric effect. The world looks and feels as it does because of electromagnetism, and the greatest advancements of communication, transportation, and computation are in the realm of electromagnetics and photonics.
Feynman was a disgusting sexist, a rapist and a misogynist. Read any of his autobiographies in which he makes full confessions that would make the 1990s pickup artist movement squeal in joy.
At no time does he get consent from his rape victims, he assumes from how they were dressed that they "wanted it".
It's disgusting that you're slandering a man decades after his death for incidents that you didn't witness. You'll notice that none of his so-called "victims" seemed to complain about it afterwards. And women do, in fact, dress to attract sexual attention. Or not, as the case may be. The fact that you apparently do not understand this very simple and easily observed trait of human behavior shows that all may henceforth ignore anything you say.
A brief comment about parameter counting:
Arithmetic coding permits one to squeeze an infinite number of finite precision parameters into a single real number. Nor is this mere pedantry in the information age. Statistics has a veritable zoo of so-called information criteria for model selection. But what do they mean by information? How do you quantify it?
Some say bits.
But how do you quantify the number of bits in a model?
Ultimately you want to compute predictions to test your hypotheses. No one has a universal turing machine because it has an infinite tape. All we finite beings of the computer age have are computers.
This conveniently provides us with a way to quantify the number of bits in a model including the residual errors made in predicting observations that are under consideration in the model:
The size of the executable archive of the data under consideration.
This is a principled the way of comparing models of the same data set.
People don't want to do this for various reasons. One reason is that it's more difficult to come up with an executable archive of your data than it is to heaven more or less informal description of the model and statistical characterizations of the errors. Another more nefarious reason is that people don't want to be held to a gold standard rigorous accounting of how bad their models are. These two factors work hand in hand to deprive us of a fundamental advance in the scientific method that could bridge the gap between the grotesque centralization of wealth and power on the one hand and those of us out here in the Heather trying to come up with explanations for what's going wrong with society.
All the well intentioned people with all that centralized wealth of power need to do, since they are so far removed from the actual theories to be able to Judge who should be funded and who shouldn't be funded, is get all the so-called experts to agree on a set of data that they all agree to compete on and then provide funding to those who best compress it.
The way you separate the sheep from the goat some experts is to see which of them most strongly resents this.
This makes me think about the big bang and age of the universe questions. Part of my journey into philosophy/history of science was due to young-earth creationism. The issues with young-earth scientific theories notwithstanding, their criticisms of the standard scientific account seem valid and remind me of many of the issues discussed in this post. For example, my understanding is that the Big Bang model is "wrong" because there is substantially less mass in the universe than necessary in order for the Big Bang to produce our universe. Solution? Add "dark matter" and "dark energy" (nb: epicycles) to make the calculations come out correctly.
In many ways, I wonder if the "scientific method" as popularly promulgated is actually responsible for a lot of these issues. Start with a hypothesis, run an experiment, and then either confirm or falsify the hypothesis. But in many ways this is backwards. This seems like starting with the answer already in mind. Running lots of experiments just to see what will happen and then trying to generalize the results seems much more reliable.
Exactly. I spoke about "dark matter" specifically in this book review a while back. https://aetherczar.substack.com/p/quantized-inertia-from-anomalies
Oh awesome. Thanks for sharing! I really appreciate what you’re doing even though a lot of it is over my head. Ultimately I think the skepticism of modern philosophy means “science” has replaced philosophy proper, and so a lot of these theories end up being lots of brilliant math and experiments cobbled together with really horrible philosophy. Sometimes one gets the sense that many of these scientists resent the demand that their theories actually make sense.
Copenhagen interpretation delenda est!
If your theory is all about exciting oscillators(standard model), and all your experiments involve exciting oscillators (particle accelerator experiments) what are you really learning?
Great talk, thank you.
The reasoning is somewhat like this: QED gave us the nuclear age, therefore it is the science of the future, and we need enormous particle accelerators to see the fundamental processes of reality. However, Einstein won the Nobel prize, not for relativity, but for the photoelectric effect. The world looks and feels as it does because of electromagnetism, and the greatest advancements of communication, transportation, and computation are in the realm of electromagnetics and photonics.
But someday, someday!
Physics history is more interesting -- and seemingly often more informative -- than the physics.
My brain shut down while taking QED once that renormalization nonsense came. It was as bad as trying to master JavaScript many years later.
Bad ideas don't deserve expending that much brainpower.
Feynman was a disgusting sexist, a rapist and a misogynist. Read any of his autobiographies in which he makes full confessions that would make the 1990s pickup artist movement squeal in joy.
At no time does he get consent from his rape victims, he assumes from how they were dressed that they "wanted it".
It's disgusting that you're slandering a man decades after his death for incidents that you didn't witness. You'll notice that none of his so-called "victims" seemed to complain about it afterwards. And women do, in fact, dress to attract sexual attention. Or not, as the case may be. The fact that you apparently do not understand this very simple and easily observed trait of human behavior shows that all may henceforth ignore anything you say.