3 ms·

Well, here I am talking shit about Wolfram, but I'm a physicist. There are thousands of us spending our lives on incredibly hard problems. The only difference b

by knzhou 7y ago

Well, here I am talking shit about Wolfram, but I'm a physicist. There are thousands of us spending our lives on incredibly hard problems. The only difference between Wolfram and Weinstein and the rest of us is that they're declaring victory in an incredibly premature, flashy way. When the rest of us have an idea, we try hard to prove it wrong (which is fundamental to how science works), not search for a lay audience to promote it to.

natalyarostova 7y ago
Fair enough. I'm not a physicist, so I'll defer to you on this. Do you think Wolfram's strange style is worth trying, as a moon-shot? (Ignoring for a second his personality).
- knzhou 7y ago
  Physics research being the way it is, everything is worth trying if somebody really believes in it, because everything is a moonshot. I'm personally working on one right now, so I'm too busy to work on Wolfram's, but more power to him if he wants to continue! My personal intuition is that a new language is only useful if it has enough "meat" to constrain things. For example, most physicists know almost nothing about logic, because it seems so far upstream of everything else that changes in it have no effect. (Indeed, logic has changed a lot in the past 100 years, and nothing happened to us!) But almost everybody in physics agrees that the language of differential forms is awesome, because with some minimal assumptions, they say that there's essentially one way to write down the theory of electromagnetism -- and it turns out to be the right way. Similarly, it looks like there's little promise in applying computability theory to physics, because it is grounded in what happens "at infinity" (and hence does not lend itself to predicting anything in our inherently finite experiments), but real promise in applying information theory and computational complexity theory, which can tell us about asymptotics. So that was why my first reaction is that Wolfram's exceedingly general language wouldn't be very helpful. However, attitudes can and have changed. If Wolfram and co. come up with a sharp success, where they derive something important without directly putting what they want to get into their starting assumptions, people would pay attention. That's precisely why, e.g. special and general relativity, quantum field theory and string theory are so important today. They all started this way.

by knzhou 7y ago

natalyarostova 7y ago

Fair enough. I'm not a physicist, so I'll defer to you on this. Do you think Wolfram's strange style is worth trying, as a moon-shot? (Ignoring for a second his personality).

knzhou 7y ago
Physics research being the way it is, everything is worth trying if somebody really believes in it, because everything is a moonshot. I'm personally working on one right now, so I'm too busy to work on Wolfram's, but more power to him if he wants to continue! My personal intuition is that a new language is only useful if it has enough "meat" to constrain things. For example, most physicists know almost nothing about logic, because it seems so far upstream of everything else that changes in it have no effect. (Indeed, logic has changed a lot in the past 100 years, and nothing happened to us!) But almost everybody in physics agrees that the language of differential forms is awesome, because with some minimal assumptions, they say that there's essentially one way to write down the theory of electromagnetism -- and it turns out to be the right way. Similarly, it looks like there's little promise in applying computability theory to physics, because it is grounded in what happens "at infinity" (and hence does not lend itself to predicting anything in our inherently finite experiments), but real promise in applying information theory and computational complexity theory, which can tell us about asymptotics. So that was why my first reaction is that Wolfram's exceedingly general language wouldn't be very helpful. However, attitudes can and have changed. If Wolfram and co. come up with a sharp success, where they derive something important without directly putting what they want to get into their starting assumptions, people would pay attention. That's precisely why, e.g. special and general relativity, quantum field theory and string theory are so important today. They all started this way.