Comments on “What they don’t teach you at STEM school”

Thank you. I’ve been building and pushing meta rationality for a dozen years or so…but self constructing it, without the academic reading you’ve got, and then trying to sell. I was missing 2-3 pieces from my ability to explain it. Youve provided clarity on those, and i owe you thanks. Or even a beverage if we cross paths

It’s not quite addressing the same problem, but I found Hamming’s “You and your research” to be particularly helpful in developing “meta-systematic skills”. It’s dense - I’ve reread it a few times several months apart and seemed to pick up something new each time.

journal articles don’t explain the thinking process.

I’ve toyed with the idea of a series of vignettes about the (hypothetical) thought process that went into cool mathematical/scientific discoveries.

One of the best lectures I’ve ever attended as an undergrad was when my professor started by very informally discussing how one might go about interpolating a series of points. “We could just draw a line straight from x_1 to x_2 then x_2 to x_3… that’s an interpolation! But that seems unsatisfactory. Maybe we decide to fit a n-1 degree polynomial for all the points but… we have a problem called ‘Runge’s phenomena’. … Now if we use this cubic spline approach, we don’t have enough constraints, so we could try [description of ‘not a knot’]. Or maybe we try [description of a natural cubic spline].”

It really stuck with me because it was the first time I’d thought of mathematical discovery as an inherently creative, unpredictable process. I’d love to see a whole book along these lines, but a major difficulty in trying to write something like it is that I’m not good enough at introspecting on my own thought process yet - and writing it for other people is going to be an order of magnitude more difficult still. Plus, the intended audience is pretty small.

I really like this project. Hopefully, as I develop more of these skills on my own, I can contribute in a better way then nodding my head and saying, “Hmm… good idea” a lot.

I realize this isn’t really fair (because of the Buddhist/Tantric angle), but this whole post reads remarkably like one gearhead’s attempt to understand how normal people think. I’m pretty sure that most reasonably intelligent people outside of academia are already stage 5 (and not stage 3 as you may have claimed earlier), and that being stuck in stage 4 is basically a neurological disorder.

Furthermore, outside of a conscious understanding of systems and their limitations, I expect it would be futile to try to improve one’s “meta-rational aptitude”, because it’s just a stand-in for general intelligence.

Any ideas why they’re all management consultants? It seems you would have one.

Since there never was any ground, you always were walking on clouds—and that worked pretty well! Your eternalistic belief in systems was mistaken, but your activity was relatively effective nonetheless. In other words: because we do understand and act effectively, therefore we can. The remaining work, from 4.6 to 5.0, is learning more about how that can be, and how to do it better.

At first I read the first two sentences and felt relief at what seems surprisingly obvious. Then I reflected and realized that I don’t have good criteria for evaluating whether I understand and act effectively. Clearly everyone reading this page is fed, alive, breathing, not-suicided. Is this good enough?

To me this seems to relate to FOMO – perhaps there’s a better system just around the Internet-corner that I’d miss if I called off the search. Even if I don’t believe there’s an Ultimate System, I can’t rule out that there’s an infinite sequence of better systems that I should be chasing (instead of taking other action). Obviously this is exacerbated by the Internet, etc, etc.

It seems ironic to me that the Holy Grail of rationality – math – essentially relies on meta-systematic discovery. Do rigid STEM people experience cognitive dissonance for this? If math discovery were systematic then someone would have already followed the rules to solve one of the remaining Millenium Problems for fame, glory, money.

For the critical feedback: this article reads to me very vague, wishywashy, though I do tend to believe most of what it says. Yeah, I know, nebulosity.

Also your taste in memes seems about 5+ years behind (to this <25 Millenial) (FWIW).

Great post, in general your recent (this year) spate of increased post density has been very solid.

I seem to be feeling my way through the same cluster (vajrayana, adult development models, the good-parts continental philosophy flyover, the interesting bits of anthropology sub-fields, etc). I’ve been pointing some people to How to Think Real Good out of lack of anything better (at least that is a post and not 5 books), so I’m happy to see you mark it out as something worthy of improvement.

Some things that have been fruitful for engaging more with nebulosity as a recovering 4.5 abyss gazer: creativity research (direct engagement with meaning making as play) as exemplified by John Cleese’s Open Mode talk, physical practices (why does this work? lol no one has any idea), design pattern stuff (Sarah Perry’s reading list). Oh, this paper is great for certain types: http://www.math.cornell.edu/~mathclub/Media/thurston-proof-and-progress.pdf

Would be interested in discussing the indispensability argument with you some time via higher bandwidth channels.

Great post! Getting to 5 seems a lot more involved than I would’ve thought. A couple questions:

1. What are some examples of “legendary feats of meta-rationality,” out of curiosity?
2. What might the steps for an English major, or anyone beginning from a non-rational / true pomo background look like? (Using the same notation or whichever)

Interested in your thoughts!

Regarding rational arguments against nihilism, some stuff in the essay linked below clicked for me. It’s actually possible I found it through you, I don’t recall! https://humaniterations.net/2016/05/10/nihilism-a-lie-in-service-to-the-existing/

Have you read Anathem? It strikes me that a fictional account of going from 4-5 could be more effective in that it rolls in the unblending from beliefs that need to change step since it is happening to a character and not the reader. Also just generally more accessible. Siddartha by Hesse is similar.

For me this is the single best thing you’ve written on this subject. Thank you!

Strong points:

1. Framing it as a curriculum makes it easier for my practical, systematic mind to make sense of it. Tying it down to “what you need to do to get such-and-such a result” keeps it from wandering off into suspiciously vague abstractions—and unfortunately “suspiciously vague abstractions” is a fair description of how I read most of the book. I found your pieces on nutrition science and gambling especially clear for similar reasons.

2. Supplementing all the stubs with recommended reading is really helpful. You’ve successfully convinced me that you’re talking about something important, so it’s depressing to know that the most important parts are not likely to be finished anytime soon. But this way, even if reading the book is 100x as hard as whatever you’re planning, we have something to fall back on while we wait.

On a different note, is there a shortish answer as to what “Certainty” means in “Beacon of Certainty”? On the face of it, “certainty” sounds like the sort of eternalist ideal you’re criticizing.

Hi, big fan - I enjoy your writing a lot (much of it is revelatory), but I keep finding one thing confusing among all the abstraction: when you talk about systems not being able to provide “solid ground” or lacking ultimate justification, what exactly is required that isn’t delivered?

Would such a system need to provide satisfactory answers to any question you might have? Give you an explicit procedure for aquiring certain knowledge or making optimal decisions at all times?

I find it obvious that such systems do not exist. Is that all your critique of systems are supposed to mean or are you making a stronger claim? In other words, are we talking “there is no truly ‘correct’ map of the territory” or “there is no territory at all, only maps”?

what exactly is required that isn’t delivered?

Meaning that is context free, in the same way that science is ostensibly the search for truth that is context free.

Unfortunately, this is a type error, meaning is inextricably bound up in context. The strongest meaning structures have the richest contexts. Observe the trope of ‘world building’ in fiction and how it interacts with the meaning people assign to the stories told within those worlds.

The scarier part (for stem folk) is the harder to discern fact that even our most context free scientific invariants are not fully context free. Quine is on about exactly this in his famous ontological commitments argument (indispensibility arguments in math).

This discussion reminds me of when I flunked philosophy of science. We learned about Hume and the problem of induction, but I didn’t really see the problem of it. It seemed to me that it could be treated as any other approximation problem; you could try different induction and meta-induction schemes and go with the one that looks like it’ll work the best. The weakness of this scheme is that it already assumes that the future will be like the past, but it seems like the best assumption to start from, as it is a zeroth order polynomial fit. It will be wrong sometimes, but whenever you look at the past it will have been probably approximately correct.

Looking back I think the point of the problem of induction was that there is no solid logical ground for predicting anything you haven’t yet seen. You cannot prove that your system contains it because there is nothing in logic that FORCES the next instant to be drawn from the same source as all the previous instants.

I couldn’t understand why this was a problem at the time because the system I had chosen for the world was based on metaphorical evolution. The motto was: if you stop doing all the things that didn’t work, you’ll be left with something that usually works. (Which made Penrose’s “The Emperor’s New Mind” seem very silly)

When I was young I really enjoyed trying to figure out what the full set of parameters you would need to generate all the different bowls cups plates and everything in between (leaving off decorative patterns), or all the intermediates between knives forks and spoons. Or how a desk, table, chair, couch and bed could be transformed into each other.

Thanks for taking the time to answer so thoroughly, it’s appreciated.

I’m totally on board with everything you said about nebulosity, and the constant inability for this insight to penetrate popular discourse drives me nuts. I’m so sick of the insistence to define categories by drawing lines along their fuzzy boundaries instead of just articulating the imperfect pattern that forms the center of the category (looking at you there, philosophy of science).

The “definite answers” bit is to me what needs unpacking. I find that there are many different kinds of answers and questions and they can’t be lumped into a category of “general questions”. There is the fact-value distinction for once (including composites), and then there is nebulosity, then vague social concepts like “meaning”, and some questions are just nonsensical like “what’s the color of tuesdays” (I include questions like “do people have free will?” in this category). So when I hear “answers to questions” I don’t know what it refers to. I mean, did logical positivism think it could answer all questions? Didn’t it dismiss whole classes of questions and statements as nonsensical? That doesn’t sound like eternalism as described.

The rejection of certain knowledge or fully correct maps is often treated as a much bigger deal than what’s warranted, I feel. We can’t have certainty or perfection, but it still seems obvious that some maps are more accurate than others - truth and falsity are nebulous too and don’t cease to exist because they are not sharply delineated categories. That we can’t have infinite certainty and correctness doesn’t necessarily mean that we can’t become more and more certain and correct - even if we can’t build an infinitely tall skyscraper there is still no hard limit to how tall one can be.

I also seems to me that nebulosity decreases or possibly even vanishes as we go from the macroscopic to the microscopic, I’d be much less willing call a concept like “methane molecule” nebulous, and that’s why physics and chemistry are different (and easier) than sociology. One could argue that it’s just another model (á la po-mo) but I think that’s making it a little too easy.

This is probably just a matter of differing interests and priorities, but to me stating that there is a real territory and therefore a correct map (even if it’s completely impractical and inaccessible) is important on principle - because that, combined with the preceding two paragraphs, counters the idea that our models are completely untethered from any real reality by providing some (hypothetical) real but intractable system that can be used as a theoretical reference point against which our various approximatory models can be evaluated (but never perfectly).

I don’t think you mean our models are completely untethered, but many people sound like they do. And it does appear contradictory when you say both that “no system can, even in principle, have a fully adequate description vocabulary” and then “If we had a complete unified physical theory, it might be a ‘truly correct map’, but (as in one of Borges’ stories) it would be useless for most purposes because it would be the same size as the territory”. Practically useless but philosophically valuable, I say (somewhat like an existence proof).

Crap, I just wrote 600 words in one go. I’m rambling. Forgive me if this is unclear, I’m just sorting out my thoughts on this and I’ll probably have a better model (sic!) soon.

P.S I don’t find the problem of induction compelling either.

The current discussion here reminded me of this quote from The Logic of Scientific Discovery:

Time and again an entirely new philosophical movement arises which finally unmasks the old philosophical problems as pseudo-problems, and which confronts the wicked nonsense of philosophy with the good sense of meaningful, positive, empirical science. And time and again do the despised defenders of 'traditional philosophy' try to explain to the leaders of the latest positivistic assault that the main problem of philosophy is the critical analysis of the appeal to the authority of 'experience' - precisely that 'experience' which every latest discoverer of positivism is, as ever, artlessly taking for granted.

He was presumably grumpy about all the positivists converging on Vienna. Ironically the one thing everyone now remembers about Popper is the boiled-down Science! claim that theories must be falsifiable, which is a nice first step but doesn’t capture much of what he was interested in.

Late response, sorry.

When I was critiquing your meme taste I had the “how many layers of irony r u on” meme in mind. Later that day I think I saw you reference the same meme on Twitter. What the fuck David?

Doesn’t Venketash Rao also do management consulting of some kind? Doesn’t the GTD guy also? Seems like a popping gig for people who know everything.

I think I’ve asked a question on another thread and was also directed to the Purpose section. I’ll go there now to expend some angst in the comment section.

Of course, mentioning GTD because of the comparison in the Hacker News thread about you

Well, the project was to explain how Science could answer all meaningful questions.

My non-expert sense is that they more or less defined “meaningfulness” as “being answerable by science”, which makes it fairly easy to construct a “perfect” system…

Rationalist eternalism is a pattern of emotional thinking that craves definite answers and pretends to believe that they must somehow be available through application of rationality.

Ok, maybe I’m the odd one out here but I never got the feeling that this is very common - almost a straw man. When people talk about things being optimal, rational and such I don’t see it as being context-free or metaphysical, more like practical. To me, the fact that I can’t have infinite certainty isn’t much more relevant than the fact that I can’t make infinitely much money. I could be assuming too much about others, though.

Well, a unified physical theory wouldn't answer questions like "was George Washington really the first American President?"

By complete system I was thinking more along the lines of “complete history of the locations and interactions of every particle in the universe” rather than just unified physics. It’d have difficulty with such questions too, but that’s because of the limitations of the (different) conceptual system embodied in ordinary language, not the hypothetical complete system itself. In theory, all questions posed in terms of the constituents of the system themselves (i.e particles) could be answered correctly.

When I read your writing I try to work out whether you’re saying that I’m wrong about something or whether you’re saying things I agree with but coming from a different direction with a different emphasis. This exchange has made things clearer (looks more like the second), so thanks, and I’m looking forward to reading the full book when it gets finished.

What’s Gödel woo?

I don’t think I’ve come across that. As a computer scientist, learning about Gödel’s proof, and also the halting problem, was helpful in leading me to see the inherent limits of systems, and start to wonder what other ways of engaging with the world there might be.

That sounds like a pretty straightforward conclusion from an elegant proof.

Is this the woo? Or is it attributing some sort of mystical meaning to incompleteness… perhaps something like “normal systems are limited by Gödel’s incompleteness proof, but the human brain isn’t because magic?”

Note, though, that this (by itself) has almost no implications for any of the things most people mostly care about.

Oh, absolutely! The “things we care about” often don’t have true answers at all (and those that do still have conceptual nebulosity and epistemic uncertainty). I just find a hypothetical “absolute correctness” useful as a sort of navigational aid and standard against which to ground and judge the merits of approximations, even if we can’t reach it. A philosophical North Star, if you will.

we can find the rules by video recording people eating breakfast, and watching carefully, over and over.

That sounds very much like how machine learning works.

No?

STEM education has been on the rise in recent years. It is a field that covers many different disciplines, but one that is increasingly important for future job prospects.

STEM-related fields have always been a big draw for students, but the demand for STEM-related jobs has grown significantly as well. This is due to the fact that many of these jobs are high paying and offer opportunities to work with cutting-edge technology.

STEM education teaches students how to use their analytical and creative skills in order to solve problems and tackle challenges that arise in everyday life. It also helps them develop the skills needed to work with new technologies and innovations - which will be crucial in the future.