Career, Physics and Goals (was: Artificial General Intelligence Speculations)

I have a similar intuition about philosophy: no one part is all that hard and people should be able to learn it if they’d just actually try. Part of me thinks that. It’s something I’ve thought about before.

did they get stuck on problems that are so hard and so individualized that no one could figure out how to solve them

Being stuck on individualized problems that require a bunch of introspection is one of the reasons that people stop wanting to learn it or evade. I don’t know how to cleanly separate the causes between evasion/dishonesty, choosing not to, or other causes.

Lots of evasion is because they got stuck in some way first. Why evade if it’s going well? Or maybe they evaded something small (and maybe didn’t even mention it in any way to me), then got stuck, then did a bunch more evasion that was more visible. It’s hard to know.

With philosophy, a fair amount of people start off looking kinda promising and make some fairly rapid progress. They learn some parts of philosophy that they don’t get stuck on. But then once they find a hard part (for them) and get stuck, they stay stuck and start becoming dishonest/evasive/etc. It’s common that people never get past the first major difficulty.

One issue with philosophy is I think the current educational materials are less complete than other fields. That is, even if you learned all the stuff I’ve written down, I doubt you’d be able to do what I do. I think I’d need to figure out some more stuff that I do to write down (or you’d have to figure out some stuff yourself). It’s hard to guess how much stuff I’m doing that I haven’t recognized as important to teach and written about. That seems like a significant concern if you want to be a physics “genius” like Feynman or Einstein – no doubt they were doing some things that are not explained in any textbooks or taught in schools – but it seems like you can at least reach a functional, effective level of physics skill based on stuff that is currently taught. It’s unclear how much and which parts of what I do could be omitted and still reach a highly effective level of philosophy skill.

Incompleteness of educational materials is an issue in all fields, but not as big of an issue.

Also, other fields have more semi-redundant material which helps people get unstuck. E.g. if you’re having trouble with one math textbook, you can try a different one that explains the same concept in a similar but different way.

With philosophy, there is sometimes only a small number of main educational sources for an idea. Some philosophy concepts only have a single source. Some of what Ayn Rand or I wrote is not available anywhere else. (This is one reason I write about the same topic multiple times. I’m trying to give more robust options and explanations for it instead of just one.)

Also, with philosophy it’s sometimes important to use educational materials from outside of your field. That is less of an issue with physics – you can kinda just stick to math and physics stuff (and maybe programming). (This assumes you already have widespread early education stuff like learning to read and write.) With philosophy, it could be important to learn some stuff from grammar educational materials, evolution, math, logic, trees, some programming, whatever category people put Goldratt in, etc. And learning some history or economics seems useful even if not required. Part of what’s going on here is the current understanding of the field of philosophy, and boundaries put around it, is wrong. Or it could be called “traditional philosophy” or something. It’s a meaningful bundle of stuff that I wouldn’t want to have no term for. But if your goal is to seek the truth and get really good at critical thinking, learning, judging ideas, etc., then the boundaries of your field are pretty different than traditional/academic philosophy and include parts of a bunch of things that are considered their own autonomous field.

There are many examples of people failing/quitting at philosophy in discussion archives. A lot of that is in emails. Here’s one:

Here’s another:

This was my reply to a new poster who responded:

Thanks Elliot for the excellent and thoughtful feedback. I’ll try to come up with a new version that improves the problems of the existing version.

I joined the other group. I hope to be an active participant.

He never followed up on this epistemology stuff. He posted in some other threads then quit. He is now a CritRat leader with a blog who hates me, is involved in the harassment campaign and, according to a source, has broken libel, slander, or defamation laws regarding me. In some sense, he wasn’t uninterested. He blogs about epistemology now but wouldn’t actually engage in discussion about it. Maybe he didn’t like being challenged or corrected, and thought it would be easy to fix his ideas to take into account the points I made, but then he wasn’t able to do it when he tried – maybe he got stuck and found it hard because his goal was just to make superficial changes. Maybe he just wants to posture as smart and social climb. Or maybe something else. It’s hard to know. After not posting for months, I recall he later denied that he’d quit (IIRC I brought it up and pinged him) and he seemed offended that I didn’t regard him as a current, active, involved philosopher … but after that he still didn’t follow up and actually continue the initial discussion of epistemology that he’d praised.

Anyway, I think how and why people get stuck is hard, subtle, nuanced, and there’s more to know than “evasion”. Even if it is evasion at one level of explanation/abstraction, it could still be studied on other levels, e.g. one could look at detailed types and causes of evasion.

Also, @ingracke has a lot of experiencing doing math tutoring (high school level or lower), and has some experience with how/why people get stuck re math. One result is she knows a lot of common errors, gaps in understanding, etc., and how to check for them. And she has talked to people who are, allegedly, really good at math, have advanced credentials, etc., and found that some of them are actually pretty bad at math, don’t understand basic stuff, etc. One thing that happens with math education is people in university get some fragile knowledge to be able to do certain calculations, but they never actually have a good conceptual understanding of math. I think it’s the same with physicists. I think you’re overestimating the understanding of many of your peers and professors, and possibly yourself. Or in other words, I think the failure rate in your fields is a lot higher than you think, if defined in a reasonable way. You can be bad at physics and get a job, do some bad work, get grad students to do some work for you and take credit, do a bunch of social climbing and office politics, etc. People do that. I’d count that as failure but not everyone would since it meets some criteria of success like getting a desirable, competitive job. (The risk of failing like that, and not really knowing you’re a failure (though maybe many of them know more than zero about what they are, like Guy Francon in The Fountainhead did), is one of the risks of not knowing enough philosophy. It’s an example of a major systemic problem in a field that members of the field broadly don’t recognize and talk about.)

Feynman wrote about fragile knowledge of physics students in Brazil: Richard Feynman on education in Brazil | Rob’s Posts

I think you’d be disturbed by how much similar issues apply in the U.S., Europe, etc. It can be hard to find out without knowing the right ways to poke people (e.g. questions to ask), because they try to hide their ignorance, inability, etc. And most people won’t just answer questions from most other people in most contexts. They don’t do Paths Forward and they try to limit and control what they say that might be wrong, embarrassing, etc.

Related is Feyman’s experience with math textbooks in California.

Broadly, in conversations about everything, most people constantly pretend to understand more than they do. People misunderstand each other all the time and they both hide the issues, gloss them over, and act like everything is going well (a lot of this is automatized, so they don’t consciously know they’re doing lots of it). I think most people are fooled most of the time. It’s similar to reactions I get to my articles: people say they like it and think to themselves that they understood it OK, but they don’t understand it well. They don’t know what understanding it well would be like or what they’re missing, so they don’t recognize themselves as missing it. That happens all the time. I challenge and question my fans a lot more than other creators, but I still ignore it most of the time even when I see signs that people are being dishonest. People often take it at face value when someone says they liked your article and learned a lot, and maybe say a few more things that prima facie appear to show some understanding.

Do you see a conflict between these two statements? If not, what do you mean by “extremely difficult to learn”?

I definitely do not see a conflict.

By saying it was “extremely difficult to learn” I had in mind facts like:

  • String theory stuff is not taught in courses, so it requires a lot of motivation and learning from books / papers.
  • Another reason why significant independent study is needed is that if you want to do string theory, you need to be a few years ahead of where you’re “supposed” to be. E.g. before you can really start learning string theory you need to know a lot about quantum field theory, which is notoriously hard to learn (it requires solid knowledge of field theory and quantum mechanics, it’s computationally difficult, and it’s conceptually difficult because understanding the concepts requires doing the computations), and it’s something that non- string theorists usually don’t learn until grad school.
  • With a lot of the stuff I was learning, I had to answer my own questions because no one at my undergraduate institution knew the answer (or in the cases where someone did, they were too busy).
  • A lot of the math I know is not stuff that physicists typically know, and so I had to learn it on my own, or by taking work-intensive advanced math courses on top of the physics courses and the independent study of string theory.

In summary, all I’m really saying is that the stuff I know took a long time to learn, a lot of motivation, and a lot of hard work. I still believe that anyone else could do what I did over the course of a few years if they were motivated enough and didn’t make bad choices.

It’s difficult to assess the professors, because even if their understanding is in some sense bad or not where it should be, they certainly know a hell of a lot more than me. It happens quite often that I read something in a paper and think “this guy has to have made a mistake,” but then later I realize that he was actually correct for some subtle reasons.

I will say that I am consistently disappointed by the level of understanding of my peers. My undergraduate institution was mediocre, so I attributed it to that, but I have been at world-class institutions for my master’s and PhD. I don’t know if it’s cultural decline, or if it has always been like that.

I think I’m very honest with myself about what I know and what I don’t know.

People in the US are nowhere near as bad as the Brazilians in Feynman’s story. There might be a more complicated way in which peoples’ knowledge is still fragile, but if that’s true, it’s much more subtle than the sort of thing that Feynman describes.

This is true. Even as a grad student, I find that it’s hard af to actually talk to professors: they are bad about responding to emails, and all the grad students are kind of afraid to talk to them. The outward-facing, PC explanation for this is that the profs are extremely busy, and grad students don’t want to look like idiots and don’t want to waste the professor’s time. But maybe this culture is somehow a result of profs are faking something and avoiding criticism. It’s hard to know.

Knowing stuff isn’t a single dimension or thing where they know more. They know more in some noticeable ways, but that doesn’t tell you how thorough, consistent or reliable they are, or whether they have an appropriate knowledge base to actually discover anything rather than merely shut up and calculate.

Also, professors are used to dealing with student questions/issues. They have lots of extra practice at that, and lots of incentive to figure out ways to deal with it which work well for them. If you brought stuff up in different categories, and also lacked some common traits of students that the professors might be relying on, then you might get different results.

E.g. I imagine some of those professors deny MWI, and would do poorly in a debate with me about MWI, and doing that would reveal a bunch of bad things.

I don’t think the past was better overall, but in the past a lot fewer people went to university or became academics, so it was more selective. There could easily be more people today who are any good at stuff (a larger percentage increase than the increase in population), but they’re a significantly lower proportion of the people involved. And in the past, I think there was more self-selection – a lot of people didn’t see it as high status and didn’t want it. For example, DD told me roughly that physics was low status until physicists enabled nuking stuff in WWII, and then the elites started caring a lot about physics, raised its status, started controlling it and destroying it, etc.

It’s also possible that, in the past, worse communications and travel technology (and downstream effects less like centralized government control over schooling) made society less homogeneous which helped with having some big positive outliers.

Having looked into various things (e.g. reading some papers from many fields), I’m confident tons of them are faking in at least some fields. And there’s large public evidence like the replication crisis.

It’s not just a matter of honesty. People often don’t know they don’t know. You need methods to check.

What have you done to check your work and find out your knowledge quality? What if you don’t know what you’re missing or what to look for? What enables an objective assessment? And what context is that assessment relative to, and what limitations does that context have which might be bad?

The usual things include checking some calculations against what a computer says, getting approval from peers or people above you in the field’s social hierarchy, or taking tests. But people’s opinions are unreliable and tests are an inadequate metric. Tests are bad at testing lots of types of explanatory understanding.

I totally believe that social science fields are faking in a major way, and that they pretend to know things that they don’t (e.g. through explanation-less “science”). I know for a fact that there is some degree of faking happening in theoretical physics, but there’s also a lot of real knowledge, and it’s hard to separate the wheat from the chaff.

The objective methods I use to check my work are:

  • Checking calculations against “physical intuition” (what this really means is like, our physical theories—besides just being math—say some loose/imprecise things about how things ought to behave, and I can check my precise mathematical answers to see if they are consistent with that expectation)
  • Checking the number/result I found against what others found, against computer simulations, etc
  • Rigorous mathematical proofs in some places

I definitely think I’ve been guilty of using non-objective methods, too (as I was getting at in this post).

Right. In any endeavor, it’s always going to be the case to one extent or another that a person is missing stuff and is unsure where his errors are.

I think that all the methods I listed above are objective assessments, but they only work in a relatively local context. They are only capable of exposing some errors. Like, using the methods I have listed it would be hard to tell if e.g. the whole idea of string theory is wrong/bad. Or, even if it’s true that string theory is a good thing to try, I don’t know how to objectively assess if the specific project I’m working on is the right one.

Yeah so methods like those (plus tests or practice problems with answer keys) find certain categories of errors but not other categories. They aren’t very thorough.

It’s not easy to do better. There isn’t a quick fix. But it’s an important problem.

One way to approach it is:

Do you routinely score below 100% on tests?

Schools have a culture of treating sub-100 scores as good or high test scores, normalizing them, and viewing errors as something that just happen, somehow at some rate. Some errors in answers are expected and are not seen as indications of errors in one’s learning/knowledge.

It’d be better to post mortem errors and try to understand (and solve) root causes. If you don’t understand the cause of an error, you don’t know how big a deal it is or what its consequences are.

Tests aren’t a great way of finding problems, but they do find some problems which could then be investigated much more than they typically are.

BTW, there also exist cases where the test is wrong – that is, the cause of a wrong answer is not actually worth fixing. The test was asking you to have some knowledge that isn’t worth the effort. One cause of this, besides bad test writing, is reusing the same test with multiple students who have different problem situations.

Here are some common types of dishonesty. Are these the kinds of things you have in mind for what you avoid doing?

For these examples I’m thinking of math calculation problems as a context.

  • Read how to do something, think you know it, move on.
  • Do a practice problem, get it wrong, think the mistake was bad luck (or some other kind of unimportant error) and you actually do know how to do it, move on.
  • Review how to do something, then do a very similar problem immediately, get it right, then think you’re done learning it.
  • Do 10 problems of the same type, get 5 right, focus selective attention on the ones you got right, think you’re good at it, move on.
  • Do a problem, get it wrong, read the answer key, say “ah I see what I did wrong; I get it now”. Be done.
1 Like

Yeah, definitely. Everyone does.

I agree with your point that “Tests are bad at testing lots of types of explanatory understanding.” I’ve met people who seem like they don’t really get it but know how to consistently obtain high marks. I have been given bad marks for tests where I think I understood the material very well, and I’ve received excellent scores on tests that I thought deserved to fail.

I don’t consider it problematic at all when I don’t get near 100% on a test, because I think that the causes of my errors are typically not worth fixing. Here are some of the most common causes for my errors on tests that I think are completely reasonable:

  • I didn’t study something that was on the test, because the person who wrote the test wants me to know things that I am not interested in learning, and so I put in the minimum amount of effort that I think will be sufficient to get a solid A after the curve.
  • Sometimes the reason I make errors on tests is that tests are timed, and the person who wrote the test is testing me on things that I already know but learned a long time ago (and I didn’t want to review it because it would be boring), and so it takes me a bit longer to work through them than the people who just learned it and did 10 similar problems the night before last.
  • Sometimes the reason I make errors on tests is that I have to do a long calculation and I get unlucky and make an arithmetic mistake on line 2, and the error doesn’t cause any easily identifiable problems until later on in the calculation when it’s too late to fix (because tests are timed).
  • It’s common for tests to ask questions that can only really be solved if the student remembers some unimportant esoteric trick from some old homework problem (in fact, I think questions like that are more common than questions that test conceptual knowledge), and sometimes I didn’t happen to review that homework problem.

Admittedly, sometimes there have been bad reasons why I make errors, like I panicked about something, or rushed through the problem description and misunderstood what the instructor wrote, or I didn’t realize something would be on the test, or when studying I tricked myself into thinking I understood something when I didn’t. But I think I have gotten better about stuff like that with time.


Tangent: This sounds really interesting to me. Do you have any idea where DD learned that?

  • I didn’t study something that was on the test, because the person who wrote the test wants me to know things that I am not interested in learning, and so I put in the minimum amount of effort that I think will be sufficient to get a solid A after the curve.

This is blaming the test.

  • Sometimes the reason I make errors on tests is that tests are timed, and the person who wrote the test is testing me on things that I already know but learned a long time ago (and I didn’t want to review it because it would be boring), and so it takes me a bit longer to work through them than the people who just learned it and did 10 similar problems the night before last.

This is blaming the test.

  • Sometimes the reason I make errors on tests is that I have to do a long calculation and I get unlucky and make an arithmetic mistake on line 2, and the error doesn’t cause any easily identifiable problems until later on in the calculation when it’s too late to fix (because tests are timed).

This is blaming the test (for time constraints) and bad luck.

Why do you regard arithmetic errors as bad luck? Have you tried identifying causes (e.g. categorizing the errors and trying to pinpoint where in your process they happen) or changing methods?

  • It’s common for tests to ask questions that can only really be solved if the student remembers some unimportant esoteric trick from some old homework problem (in fact, I think questions like that are more common than questions that test conceptual knowledge), and sometimes I didn’t happen to review that homework problem.

This is blaming the test.

So I think you’re saying that most your test errors are due to flaws in the tests, and you would typically score 100% if only the tests were better, which is why you don’t care about getting lower scores. Is that correct?

Have you gone through and categorized test errors after a test is graded? Do you have any metrics or documentation over time? Have you compared notes with any other students (or anyone) to see if they make the same errors or different ones, or to see if they agree about your categorizations of your errors?

I don’t think he had a specific source, just general knowledge of the field, its history, and of many individual physicists from various time periods. Besides reading stuff, he’s talked with lots of physicists of different ages. @alanforr might be able to comment on this.

I found the conversation because it’s relevant in a different way.

Part of an offlist DD email from 2005 (link is dead now but I think it’s the 5th Solvay Conference – @alanforr also received this email so there’s a chance he remembers):

Historically interesting movie. 1927. In those days there were such things as great physicists. Several of them met on that famous occasion.

They all look so drab. No casual observer could tell that many of them were smarter than Deep Thought on steroids and that underneath those hats there was such beauty.

DD did not reply to my email asking “how do u mean there were great phys but aren’t now?” so I asked it in IMs:

curi42 (6:51:11 PM): how do u mean there were great phys but aren’t now?
oxfordphysicist (6:51:30 PM): Literally.
curi42 (6:51:44 PM): no1 is good now?
curi42 (6:51:54 PM): why? how did this happen?
curi42 (6:52:00 PM): u r not great?
oxfordphysicist (6:54:17 PM): I am arguably one of the best currently working. A handful such as Martin Rees are arguably a bit better than me. All of us are third rate. Possibly low second rate, if one is charitable, on a scale where at that conference there were at least six first rate physicists.
oxfordphysicist (6:54:54 PM): (I have qualities in addition to being a good physicist, though.)
curi42 (6:55:02 PM): you sure do! :)
curi42 (6:55:15 PM): but how did this happen?
curi42 (6:55:24 PM): how did they become great, and why does that process not work anymore?
oxfordphysicist (7:02:05 PM): Feynman said it’s because of the system of science education.
curi42 (7:02:21 PM): what did the system used to be like?
curi42 (7:02:24 PM): i know it is bad now
oxfordphysicist (7:02:50 PM): Basically, neglect. Other things were considered important. No one knew physics was anything until after WW2.
oxfordphysicist (7:03:17 PM): I have been told then when Einstein came to America, the newspapers called him a “famous mathematician”.
curi42 (7:03:28 PM): haha
curi42 (7:03:36 PM): and now that the schools are trying
curi42 (7:03:38 PM): it’s a mess
curi42 (7:03:39 PM): lovely
oxfordphysicist (7:03:46 PM): Whether that’s so or not, it was the attitude.
curi42 (7:03:55 PM): :(
curi42 (7:04:00 PM): do we need great physicists?
oxfordphysicist (7:05:10 PM): Well, I have a separate theory – just a wild guess really – that in the long run there won’t be geniuses. That being a genius is like being one of those calculating prodigies.
oxfordphysicist (7:05:41 PM): No, we, society, doesn’t need them. It is enough if we have creative people who love what they do.
curi42 (7:07:01 PM): i’m glad i didn’t try to become great at chess
curi42 (7:15:25 PM): i think ppl will get good enough at things that, at least by present day standards, they will be geniuses
curi42 (7:15:30 PM): sometimes at multiple things at once
oxfordphysicist (7:16:08 PM): That would be nice.
curi42 (7:16:33 PM): why not?
curi42 (7:16:44 PM): you are very good at a variety of things. so am i, and i’m young.
oxfordphysicist (7:17:29 PM): You could be right

Also, in 2010 DD told me:

17:53:15 oxfordphysicist: No one is a Feynman, but Penrose is pretty high up among the geniuses of our age.
17:53:21 curidotus: i don’t agree
17:53:47 curidotus: i went to his talk. i thought it was magical thinking.
17:55:45 oxfordphysicist: Yes that doesn’t mean he’s not a genius. Gödel was flat-out bonkers. But his proofs were still genius…

Note how DD agrees with me that Penrose’s talk (that I attended IRL) was magical thinking, but still thinks Penrose qualifies as a genius by current standards. He sees Penrose as having huge flaws but thinks Penrose also had some very high points. Related, DD said mean things about Penrose to me, including that he has “silly ideas about quantum theory” – but DD still considers him “one of the great minds of the day”.

Also from the 2010 IM conversation:

17:49:52 oxfordphysicist: BTW Al Schild one explained to me, when I was a grad student, his plan for improving science publications: (1) abolish the PhD degree, and (2) make all submissions to learned journals anonymous.
17:50:04 oxfordphysicist: So I eagerly went and told this idea to Dennis Sciama.
17:50:19 oxfordphysicist: And he persuaded me otherwise, as follows:
17:52:07 oxfordphysicist: If I’m a referee and I get a paper that defends an absurd position with poor arguments, I want to reject it after reading the first paragraph and so do the readers. But if Feynman wrote it, I want to read it whatever the hell it says, and so do the readers.

Note: That was a rebuttal about (2) but not about (1).

So, lmf, you seem to have a higher opinion of the physics education you’ve experienced than DD (or Feynman or I) does.

You probably think you are or will be better than a fourth rate physicist, but DD probably disagrees, and considers most or all of your professors fourth rate or lower.

Being fourth rate or lower means you have lots of errors in your physics and math knowledge. There are problems with how you understand stuff. Stuff is going wrong in your learning and thinking processes.

I think you also don’t think you have such big low points like a Godel or Penrose, which could be an indication of not seeing them. It could also be due to being early career, not fully formed yet, strengths and weaknesses still under development. It could also be related to being well rounded and fairly consistent with low variance, so nothing really great or awful, which is realistically the good outcome of university science educations now (so basically all medium, as against the other outcomes of being a mix of medium and bad or being all bad, since our universities don’t create greatness).

DD/Feynman/me think, basically, that our science education system is so bad that it can’t produce a first rate physicist. I think maybe you have a significantly higher opinion of it than they do.

1 Like

I didn’t find anything useful with for

Got some blank pages and this error page:

But I think I found what conference it was:


If by saying I’m “blaming the test” in those bullet points, you mean something like I’m attributing my failure to get 100% to errors that I consider to not be worth fixing, then you are correct that I am “blaming the test.”

I would never say verbatim that “I blame the test” or that all my errors are “due to flaws in the tests,” because I don’t think that any of the bullet points I talked about are failures on the part of the test designer that could reasonably be improved (with the possible exception of the last bullet point, though I think it’s hard to test conceptual knowledge so I’m not sure). Like, the prof who wrote the test has to set some time limit, the administration expects a B average so he has to curve, he has to choose some topics to test and isn’t always going to choose ones that I like, etc. That’s just how tests are; I don’t have any good general ideas for how the tests could be “better,” given the context.

There are some bad design choices that test-makers occasionally do make, like ambiguous questions, but I didn’t mention those because I don’t think they are a major cause of my error rate.

That’s a good point. My arithmetic errors aren’t bad luck in the sense that I believe I could reduce the error rate by a bit if put a lot of conscious effort into improving. However, I think that my error rate is already low enough that arithmetic errors not even close to being a bottleneck issue for me, so it’s not something that I worry about.

I’m not quite sure what you mean by “categorized” here, but I do go through my errors after a test is graded and try to figure out why I made them (unless I recognize an error during the test, like a question on something that I chose not to study).

My grades are one metric. I think I made more bad errors when I first started out (by which I mean the sorts of uncontrolled errors I listed below the place where I put the bullet points), and I think I improved on those. This belief is consistent with the fact that my grades improved over time (mixed Bs and As early on, straight As in the years after that).

I talk about tests with other students, but probably not in a way that’s as objective as what you have in mind.


Do you have a source of Feynman talking about how bad it all is? Is it in his books? I didn’t find that either of the Feynman articles you linked related very well to my own experience.

Maybe my primary school science education was really bad, but I don’t remember it.

Maybe there are some parallels to be drawn between Brazilian schools in the 60s and Anglosphere schools in the present, but I don’t see them clearly. The obviously bad things for which Feynman was criticizing the Brazilian schools simply don’t apply to the physics classes taught in the Anglosphere schools I’ve attended. Like, we don’t memorize sentences.

edit: I just noticed this

IIRC Deutsch knew Feynman. I definitely believe that Feynman said something like that, but it would be good to know Feynman’s reasons. Did it apply to the entire system? How bad is it? What specifically is so bad about the system? What did first-rate physicists used to be like, and how do they differ from today’s uni professors? etc.

To some extent I could answer these questions for myself if I looked into it, but if there’s a specific thing you think I should read, let me know

How should I go about identifying them?

Do you think I was on the right track with my own attempt, or do you think (or suspect) that I am completely missing the mark?

From a private DD email from 2011, about when he had a short conversation with Feynman IRL:

he [Feynman] then went on to declare that existing educational practices, mainly in universities, will ensure that “there will be no more great physicists”

1 Like

btw, are you as pessimistic about academic math as you are about academic physics?