saulmunn.com; brasstacks.blog
Saul Munn
Karma: 1,004
Thanks for posting this!
in chronological order
Why not in the order that you got value from them? This would have made it easier for me to navigate the links, though I can imagine this would be a bit more annoying on your end.
Do you still use Ommwriter?
This post is useful for providing the handles “wizard power” (knowing concrete/grounded/gears-level models on how the world works) and “king power” (leading, managing, allocating, communicating, etc). I also think that the post insightfully points out how social gradients can often harmfully and myopically push people toward king power.
But I think johnwentsworth is wrong in saying that king power is “fake,” and that people should only aim for wizard power: in particular, it sure seems like coordination is an incredibly taught resource (as johnwentsworth himself insightfully points out in his “gears that turn the world” sequence), and that (good, actually-powerful) kings can be exceptionally valuable for coordination.
I also think that there is clearly not a binary; that people can have more or less of each kind of power without necessarily having more or less of the other — indeed, it seems to me that having more wizard power causes one to be a better king (and probably vice-versa, though I’m less confident and this would probably be to a lesser extent). I’d love to hear more analysis of interplay between growing both wizard and king power intraperson and intraorganization.
Another experiment along these lines: blog-a-thon!
Attendees posted three times in one day (!!), once every three hours — starting at 10am, and posting once before each of 1pm, 4pm, and 7pm.
Attendees who published on-schedule were provided dinner for free; everyone else had to buy their own. This was the replacement for kicking people out if they didn’t post.
The output was pretty good, but it was fairly stressful and lasted quite a long duration. I think the blog-a-thon ended up being mostly useful for pushing mostly-finished-drafts into actually-published-posts, rather than getting people started on drafts in the first place — which is I think is great, tbc.
Some changes I’d make to future blog-a-thons:
Fewer quantity of posts per day. Probably two?
...but maybe spread out over two days instead of one! Especially if it’s on a long-weekend, and we can get a bunch of mattresses or cots or something, then make a sleepover out of it.
More financially sustainable for me — I don’t think I could buy everyone lunch & dinner every time. Some ideas: have fewer people come (~5 people instead of ~25); don’t buy lunches, just dinners; if this runs regularly, quietly ask 1-2 regulars to help split the cost; etc.
Some explicit structure for people to give feedback on each others’ posts.
Overall quite happy with how the blog-a-thon went! :)
I enjoyed reading this; thank you for writing it! (Though as some data, this much detail is definitely not important for my continued donations.)
In the ‘future plans’ section of your 2024 fundraising post, you briefly mentioned slowly building out an FHI-of-the-west as one of the things for which you wish you had the time & funding. I didn’t notice such a project in the same section of this post — curious what happened to your plans for this? (Have you given up on it? Or is it just not in your top priorities of what you’d do with extra funding? Or something else?)
I said in my first comment that:
It would be great if, without needing to read many paragraphs or click through many links, I knew what I knew what this post was announcing/describing.
I would really like to know what at all you’re talking about without needing to click through to other links or figure out which sections to check! Like, that’s an annoying amount of labor just to answer extremely basic questions about the program. I’m much more likely to just click away (which I’d prefer not to do!).
And of course, my point in leaving these comments is to provide you data about how your post may interact with its readers. You should feel absolutely free to ignore that data, it’s your God-given right to do so, but arguing against my experience of your post seems to me like a somewhat odd thing to do.
You could build something like this into the interface — e.g. a button that reads “Make this post pop back into my feed at increasing intervals over time” or “Email me about this post in 6 months”
This post clearly & succinctly facilitated a better decision-making process to a question that I (& many others) have: Should I cut & bulk?
The answer is not straightforwardly given in the literature, but I nevertheless found the post helpful in figuring out what the right cruxes I should be focusing on are.
Thanks!
What’s the Freeman’s Mind joke?
This was the first piece of short-fiction I’ve written! I’m keen to hear feedback, especially from folks who’ve written lots of this style of short-fiction before (speculative, playing-around-with-the-format, etc). Thanks :)
I learned some about Chinese history in the late 1900s, and so added ~10 cards about the relevant dates (“when did Mao die?”; “how long was Hua Guofeng in power?”; “when did Deng Xiaoping first come to be paramount leader?” etc.). Introspection is an unreliable narrator, but it seems like the stuff I learned stuck better, and conversations I’ve had about it since then have been easier to navigate.
I agree with the claim that “compressing skill acquisition into extremely intense, short-duration periods (‘explosive skill acquisition’) can be much more effective than extending small chunks of skill acquisition over long-duration periods (‘incremental skill acquisition’).”
I also disagree with the claim that “explosive skill acquisition {Pareto dominates, is generally more effective than} incremental skill acquisition.” I think that — if you do incremental skill acquisition right[1] — it can be pretty effective, sometimes (?often?) more effective than explosive skill acquisition.
So with that in mind, some healthy pushback to each of your points:
Overlapping forgetting curves — the answer to “shoot I might forget stuff” isn’t “bunch reminders as close as possible together.” This is super inefficient, so will requires significantly more total time than optimally spacing your reminders.
Richness of context — memory systems let you ‘remember your rabbit holes’. If I spend a couple hours improving on some skill while encoding it in a web of flashcards, I can pick it up a few months later just where I left off.
Discontinuous practice opportunities — yeah, I think generally explosive skill acquisition is better here.
Self-signaling/top idea — I do think that being the top idea in your mind is a real thing and can be shockingly powerful; same re: self-signalling. However, incremental skill acquisition can have sort-of analogue for each:
Re: top idea, spaced repetition systems can be used to program attention. When you review units of a memory system (be they flashcards, extracts, blips, etc), you bring them back into salience, where they collide with whatever else is on your mind.
Re: self-signalling, spaced repetition memory systems make memory a choice. Too often, people treat their memory systems like an inbox and subscribe to any email list they think they ought to like — then get overwhelmed with bullshit in the ensuing weeks. If it is instead treated like a mental home, then one feels more inclined to decorate it with only the most sacred, beautiful, valuable pieces. After all, the wall-space is limited.
Quantity — IDK man, ceteris paribus (including holding skill-level constant), I’d really rather do fewer reps. For some skills (e.g. writing) ceteris doesn’t end up being paribus, but for others (e.g. remembering a vocab word in another language) it totally is.
Some more reasons against explosive skill acquisition:
It’s costly. In high school, I had my summer breaks cut in ~half so that I could spend eight-twelve hours a day bouncing between lectures and drills and practice debates and research on policy and philosophy and critical theory. I became a vastly
more competentless incompetent debater; I missed out on a few family vacations. Totally worth it, but still quite costly.It’s both costly in terms of opportunity cost (I missed out on family vacations), but also in terms of direct costs (those were in ~5th percentile most stressful weeks of my life).
It’s (often) not durable. I think this is more the case for some skills and less for others, but my impression is that people underestimate how quickly the skill they just learned will be forgotten. I think that incremental skill acquisition (done well[1]) solves this.
You don’t get enough contact with reality to know what parts are important. When you spread skill acquisition over time, it’s easier to notice “hey wait I’m learning this sub-skill, which seemed important at first, but it sure looks like nobody in practice ever actually needs it? maybe I can just skip it?” or “hmm interesting this other sub-skill which wasn’t in the textbook seems pretty clearly foundational to all of the actual stuff, maybe I should focus a bit more on that.”
But I think most of the above is basically moot relative to the fact that most people do skill acquisition way way way way way less effectively than they could (cf “The MathAcademy Way” & more of Justin Skycak’s stuff; “How Learning Happens”).
Thanks for writing this, Ben!
- ^
Which itself can require quite a bit of skill/effort!
...I only ever even saw a small fraction. … I mostly just click on things on the front page.
Did you browse inkhaven.blog, or just the front page of LessWrong? The Inkhaven website has a few neat features that curate some of the better posts (e.g. encouraging authors to highlight their best posts), and includes a number of posts that authors seemed to not post to LessWrong.
I think as a reader I’d have liked the results better if participants had to publish every other day instead.
I think there are a number of potential solutions to the problem you’re outlining. This is one, but there are many others (e.g. encouraging some of the days to be more focused on editing past pieces, or only publishing the best of every three days’ posts, etc).
One maybe reasonable intuition pump here is: Imagine you had never heard of auto structures before. You have no idea what it is, or anything about it. What information would be relevant to know? Thanks for the response!
I wasn’t getting at “add the word volunteer” — and unfortunately no, the italics are not enough.
Some more-concrete questions that I basically can’t find:
When does this start? When does it end?
Is it online or in-person? If the latter, where?
How many other people do you expect to do this?
Concretely, day-to-day, what will a participant be doing/working on/contributing to?
Etc. Imagine you had never heard of auto structures before. You have no idea what it is, or anything about it. What information would be relevant to know?
Did you read “Reality is Plastic”? What’d you think?
It would be great if, without needing to read many paragraphs or click through many links, I knew what I knew what this post was announcing/describing.
What kind of thing is ‘Autostructures’? Is it a research scholarship? A fellowship for undergrads? A design-focused internship?
You describe stuff like what your goals are, what vibes you have/want to have, but… there’s no ‘here are the 5-10 words that clearly describe what this thing actually is.’
some updates:
(1) just spoke with a doctor[1] who’d looked into it a bit. he basically wasn’t worried about it at all — his take was (paraphrasing quite a bit):
causes of heart failure are numerous & complex — it’d be really confusing mechanistically if melatonin affected a sufficiently large subset of those causes sufficiently strongly to have this strong of an effect.
also, because they’re so numerous & complex, it seems pretty implausible that they could’ve done even close to a sane job of manually controlling for all of the right variables.
[this bullet added by saul, not explicitly said by the doctor but he sort-of implied it] in particular, badness of insomnia is comorbid with extent of lots of other bad stuff. i.e. it’d be reasonable to expect having a prescription for melatonin selects for having very bad insomnia, which then seems like a straightforward causal factor for getting heart failures. it’s obviously possible they did some manual controlling for this, but since we don’t have the full paper, we can’t tell (and it seems like it’d be pretty tough to actually effectively control for this).
causes of heart failure typically take quite a while to build up; it’d be surprising if it only took melatonin 5 years to have such a significant effect.
we naturally produce some melatonin, and the amount one would exogenously supplement isn’t (typically, at least for prescribed patients) substantially higher than what you’d expect to see in one’s natural range of endogenous melatonin production.
if the american heart association was actually worried about this, they’d have done some public messaging independent from the paper abstract — it’s possible such a message is forthcoming, but we probably should have expected to see something about it by now.
(2) i’ve emailed the authors asking for the full paper. i’ll aim to keep this thread updated with thoughts if/when the paper arrives in my inbox!
- ^
doctor was a psychiatrist, not a cardiologist or internal medicine doctor — so though you should take it with a grain of salt, he did still get an MD. which implies way more medical knowledge than i have, at least!
The study only looked at patients who were prescribed melatonin (though they indeed do not detail what the dosages are).
[Edit: probably not as bad as it first looks. See comment for more thoughts.]
Long-term melatonin usage for insomniacs was associated with doubling of all-cause mortality*
New article from the American Heart Association seems pretty damning for long-term melatonin usage safety:
In a large, multinational real-world cohort rigorously matched on >40 baseline variables, long-term melatonin supplementation in insomnia was associated with an 89% higher hazard of incident heart failure, a three-fold increase in HF-related hospitalizations, and a doubling of all-cause mortality over 5 years.
*Caveats:
we just have the abstract, not the full article
observational study, not experimental
their sample is only of insomniacs, not of the general population
Responses to the caveats:
we’ll get the full article soon-ish (probably a month or so?)
it seems they did quite a bit of controlling? though we won’t know how good their controlling was until the full article comes out
I can’t imagine that the validity of their results for non-insomniacs is many orders of magnitude less than for insomniacs — like, maybe a factor of two or five, but that’d still be a huge effect size
All things considered — this seems like a crazily high effect size. Am I missing something?
Veritasium (science YouTube channel) recently published a 55-min video on EUV lithography: “The World’s Most Important Machine.” Having no background knowledge in semiconductor manufacturing, I found it an enjoyable & informative watch!