=misc
There was a blog called
SlateStarCodex (aka SSC), written by a guy (Scott) who now writes a blog
called AstralCodexTen. Scott also wrote
some fiction.
For the first time in a while, somebody asked me what I
thought of him, and I happened to feel like thinking back on trivial
personal history things and writing a rambling post.
Scott wrote long
essays that I'd see links to on twitter and blogs. He had a combination of
characteristics that didn't usually go together:
- "anti-woke" but
also anti-Trump
- pro-technology, but also concerned about AI risk
-
utilitarian/egalitarian enough to donate a kidney to a stranger, but also
probably a jewish racial supremacist
- got negative coverage in the New
York Times but expected it to be positive
Did I like Scott's essays?
Meh. He wrote some things I thought were obvious but other people found
insightful, some things I wasn't interested in, and some things I disliked,
but nothing I felt like I really learned something from.
## my
connection to SSC
There were 2 main ways I could be considered
connected to SSC:
### meetups
What made me interested in
SSC was that it seemed to be a Schelling
point that might
be a way to find interesting people. I'm always on the lookout for those.
It's kind of how I didn't particularly like Paul Graham's essays, but I
still made a reddit account when I saw them. Or, like how I don't
particularly like Touhou games but there's some good Touhou
music - my apologies to ZUN
for putting him together with someone responsible for Sam Altman.
So,
I became an organizer for some SSC-related meetups. SSC wasn't unique in
that regard; I'd organized a few meetups for different things, for similar
reasons and with equally minimal effort.
Did I feel like I finally
found "my people"? No, not particularly. As a way to find interesting
people, SSC...wasn't a complete failure, but it also wasn't much better than
just trying to pick out smart people at restaurants, which does make the net
benefit a bit questionable. And of course, as a way to meet women I'd rate
that at -3.6/10.
(No, I'm not a fan of Altman. I tried to tell people
about him, but nooo. First he was good because Helion's stupid designs were
gonna make fusion power practical any day
now. Then he was the
thoughtful leader who actually understood AI risks and would do things
safely. Then he was the e/acc champion fighting those evil EA luddites. Then
he was the brilliant businessman who's going to make OpenAI more successful
than anyone else could. Every time he changes his posture he finds a new
batch of suckers. I can't stand his voice either - you might think all that
vocal fry came from sucking Thiel and Graham's dicks too hard, since that's
how he got his big breaks, but no, he *tried* to sound like that, kind of
like Elizabeth Holmes.)
### mention
I once got a brief mention on SSC. That was
because I won a bet with Scott where I was more optimistic about economic
performance during Trump's first term than Scott was, and my prize was a
link to a battery chemistry I designed. And that sort of thing probably
appealed to Scott a little bit.
That might give you the impression
that I'm strongly pro-Trump, so I should note that my view of Trump is:
-
TrumpCoin amounted to outright corruption
- the erratic tariff
announcements were largely for the sake of insider trading
- he was
complicit in Jeffrey Epstein's activities, and Barr deliberately facilitated
Epstein's death
- official reports and plans are
being
AI-generated
now
- he's starting to go
senile
Pretty crazy how
that's all within the Overton window now, huh? I didn't have to be a big fan
of Trump to think that US economic performance during his first term
wouldn't be bad, and it wasn't.
As for that battery thing, what
happened with that was:
- I made some improvements to the design.
- I
got myself a chemistry-related short-term remote job, to see if I could do
something with that. (To do that, I had to go with a doomed
chemistry-related startup. I wasn't nominally qualified for the job, but
obviously me keeping up with the chemistry PhDs was a non-issue.)
- I
went from that to a university lab where I could work on my own thing for a
while. (To do that, I had to go with a place that temporarily just wanted to
fill out their headcount. Before doing anything there I wrote up a
provisional patent application, what with university IP policies.)
- I
validated the design as best I could with what I had available.
- I got a
patent granted for that, on my own.
- I eventually gave up on doing more
with it because the prospects weren't looking good and I had other stuff
going on.
(What SSC linked to used the real concept but not the real
chemical choices, because I figured any people interested enough to try it
would comment or try to contact me first. Well, it wasn't very successful,
but that had zero to do with whether the design worked or not, and a lot to
do with how many lying battery startups there are. It's not that *nobody*
was interested, it's that the kinds of people who were interested because of
that weren't relevant for that sort of thing. Well, you know how it is,
investors need to focus on more
credible
## schizophrenia & alzheimer's
This reminds me...I
never did post that schizophrenia theory I tried to explain to Scott.
I met Scott once at an in-person meetup. I was the first person to
notice him, and took that opportunity to try to explain my theory about
schizophrenia to him, and the issues I had with Scott's nascent theory of
it. I didn't get through much explanation before he went to talk to someone
he knew, but it didn't matter anyway because I could tell he didn't
understand.
I suppose I could write a post about that now...but I did
post this about
Alzheimer's, and isn't one enough? Arguably one was too many. Anyway, my
schizophrenia thing was just some speculation based on
- finding
analogies between the effect of some recreational drugs and elements of
schizophrenia, to find some relevant receptors
- using some molecular
biology logic about turnover rates to guess some relationships between
receptor types for molecules with different types of modifications
- some
guesses about how the neural subsystems are connected based on how you'd try
to replicate brain functionality with artificial neural network structures
- put together into some guesses about how to produce an opposite effect by
targeting different receptor variants
Pretty speculative stuff
obviously, more so than that Alzheimer's post, but also more useful for
treatment if correct.
I think Scott eventually gave up on his silly
schizophrenia theory, but at this time, Scott has just recently put up a
guest post about Alzheimer's. I don't even want to link to it because it's
very dumb, but here it
is.
Anyway, it just cherry-picks studies that the post author doesn't even
understand, and the core thesis is that:
- amyloid plaques spread because
they're basically prions (no they're not)
- then they cause Tau protein
abnormalities (how? why? who knows)
Their main "evidence" for this
is:
There were people in the past whose obsolete medical treatment
(growth hormone from dead brains) led to them getting CJD, a prion disease.
It turns out old samples of that serum also contain beta-amyloid. (So what?
That's normal.) And then, mice genetically engineered to have abnormal
amyloid protein that wants to form plaques, if you inject this serum that
has amyloid proteins that formed some clusters, they act as seems for the
amyloid plaques. But that means nothing! The plaques form because the
amyloid is abnormal! Normally there's an equilibrium between soluble and
insoluble forms of amyloid protein. It's like saying, "the reason your glass
of water froze is because your water was exposed to dust that seeded the
ice". No, the reason your glass of water froze is because the water was
below freezing. Yes, CJD prions are like ice-9, but amyloid is not; people
would've actually noticed plaques suddenly forming from **normal** amyloid
once seeded - like they noticed CJD prions replicating - if it was. It's
just...so tiresome.
Oh, and also there were a few of those people
with a rare disorder who got various medical treatments (including serum
maybe containing CJD prions, radiation treatment, etc) who got (what
appeared to be) Alzheimer's at an early age. I'm not sure what that's
supposed to prove.
Yes, those papers cited were *trying* to make a
link to causes of Alzheimer's, and to the amyloid theory in particular.
That's just what some people do in papers to get published. You get used to
that, but not if you don't have the experience to see it happening,
because...
> I am David Schneider-Joseph, an engineer formerly with
SpaceX and Google, now working in AI safety. Alzheimer’s isn’t my field, but
I got very interested in it, spent six months studying the literature
Clearly he didn't study the literature enough, because these:
>
Amyloid deposits are a necessary (i.e. but-for) cause in all instances of
Alzheimer dementia
> ...
> Severe enough amyloid pathology is a
sufficient cause of Alzheimer dementia in almost all brains
...are
wrong.
Well, unless you define Alzheimer's as requiring amyloid deposits, but then
that's just circular logic.
So, that post's author, and Scott who
posted it...let me try to phrase this in a charitable way...maybe they could
apply to be science writers for the NYT, since they're *almost* up to that
level.
## meetups
So, I organized a few meetups at parks
or restaurants I liked. That went fine, but the number of people gradually
declined. That was fine with me, because interesting people can usually find
things without having to promote them. But after a while, I figured I'd done
enough meetup organizing, and passed management of things over to 2 people,
let's call them Bob and Joe. Mostly because they were the only people
interested in doing that.
Joe started promoting the meetup group
various places online, which brought in more people, and Bob ran a Discord
server.
### Joe
Some time later, Joe tried to help some
rich guy who made his money from sketchy stuff - I think including
multi-level marketing - steal my battery thing. The first step was me
agreeing to transfer an unspecified % of ownership in exchange for his
personal lawyer drafting a *provisional* patent application. Also, Joe
insisted on being the sole point of contact for everything instead of just
introducing me. Anyway, I noped right out of that, and after that Joe turned
more negative towards me. In the end, well, obviously I didn't actually need
help to get a patent. I'd rather do nothing than work with guys like that,
so I certainly have no regrets.
### Bob
After the above,
I'd still sometimes go by meetups briefly and leave early, just to check if
somebody interesting showed up. But then, in 2023, I disagreed with Bob
about the Al-Ahli Arab Hospital explosion
.
The
analysis by Forensic Architecture is basically what my position was. (I
still think it's a good analysis, but this is kind of a moot point now,
because everyone's already come to a conclusion about the nature of what
Israel's been doing in Gaza.) Bob took issue with that. You see,
>
the US intelligence community had determined with "high confidence" that the
cause of the explosion was not an Israeli airstrike
and to Bob, that
was the final word. He insisted I therefore must renounce all future claims
to credibility on that topic, and I went "lol no". I figured he might ban me
from the Discord after that, so I started a script to delete all my messages
because that's just general good practice, and indeed, without me saying
anything else, he then banned me from that Discord.
So, I fixed some
green tea, put on some classical
music, and that's when I
started reading Super
Supportive, which
is a pretty good story so I guess things worked out.
Bob's job
involved writing software for a defense contractor, to help the US
government monitor US cell phone messages, and he considered himself an
honorary member of the US "Intelligence Community" (IC). So, he felt that
challenges to the credibility of the IC were an attack on his identity. I
understand, I do, I just can't really respect people who operate at that
level of thinking.
Yes, politicians lie about intelligence stuff all
the time, but in Bob's view, the "rules" were something like:
- if
it's a politician talking it doesn't count
- if it's a retired IC person
talking, also doesn't count
- it has to be from an active IC person
acting in an official role, or an official statement from an IC agency as a
whole
- bonus credibility if it says "with medium confidence" or "with
high confidence"
Then, if those rules are followed, it's 100%
trustworthy and never lying, because that means IC credibility is being put
on the line...according to Bob, that is. Me, I'm not so sure. For a recent
example of just how trustworthy statements are when they follow the rules,
here's a FBI
statement on Epstein's
death, saying:
> The conclusion that Epstein died by suicide is
further supported by video footage from the
common area of the Special
Housing Unit (SHU) where Epstein was housed at the time of his
death. As
DOJ’s Inspector General explained in 2023, anyone entering or attempting to
enter the
tier where Epstein’s cell was located from the SHU common area
would have been captured by
this footage. The FBI’s independent review of
this footage confirmed that from the time Epstein
was locked in his cell
at around 10:40 pm on August 9, 2019, until around 6:30 am the next
morning, nobody entered any of the tiers in the SHU.
But the thing
is...
- the released video shows several people coming and going
-
Epstein's cell can clearly be reached without being seen by that camera
-
there's missing time segments and other artifacts
- metadata indicated it
was edited in Adobe
Premier
from multiple source clips, despite claims it was raw and unedited
That was all explained afterwards, right? Well, Pam Bondi said:
>
"There was a minute that was off that counter, and what we learned from
Bureau of Prisons was every year, every night, they redo that video," Bondi
said July 8, noting that the system was old. "Every night is reset, so every
night should have that same missing minute. So we're looking for that video
as well, to show it's missing every night."
No explanation for the
Adobe Premier thing, or the other issues. And also Bondi was lying about
that missing minute being normal for cameras in that prison:
> But a
high-level government source familiar with the investigation told CBS News
that the FBI, the Bureau of Prisons and the Department of Justice's Office
of Inspector General are in possession of full unedited copies of the video,
and those copies do not have a missing minute. Why Bondi said that the video
resets is not clear.
I could go on, of course, but this has already
been discussed pretty thoroughly.
### why do I find these people
I don't want to give the impression that I think about Joe and Bob all
the time here. They did make it onto my list of "people I met that I'm
reminded of sometimes", but that's a long list, including such characters
as:
- the girl I briefly dated who later became a trans man and
founded an AI startup
- the ocaml programmer who named his kid Kyon
before becoming a trans woman and dating someone who looked like their twin
- the programmer/chemist who synthesized their own experimental ADHD
medication and deliberately gave themself a split personality
- the CEO
who liked Lesswrong & Project Orion & math, who was posting about colonizing
Mars while his company was in crisis
- the guy who proudly announced in a
classroom that his favorite anime was Himegoto
- the mom who sent 1 kid
away to CPS and punished the other kid for not doing homework by sending
away the family dog
- the brilliant physicist who later became a notable
anti-(COVID lab leak theory) person
- the grad student who committed
suicide after retaliation for pushing back on fake data
...dang it,
now I'm sad again.
It might seem like the common factor there is me,
but it's not my fault I keep running into weird people. I really want to
stress just how boring I am.
The weird advice I give people is stuff like
"maybe drink less soda" or "maybe eat oatmeal
sometimes". My idea of
excitement is a slow 2-hour hike on a mountain or trying a new restaurant.
When I see, say, special forces doing some crazy training, or med
students doing 80 hours a week, my response is "let's just avoid that then".
I have a blog, sure, but is it really that weird to like learning about some
science stuff and sometimes tell people about what you learned? I asked my
friend for comments just now, and the only "interesting" thing about me he
could think of was that I got hit by a train once - and it's not like I
chose to get hit by a train, plus that wasn't even very important for me,
let alone other people.
So...why, then? My working theory is that I
was cursed by an ancient totem I touched as a child, but I'm open to
suggestions.
### public intellectuals
It makes sense
that people who liked SSC would be interested in other "public
intellectuals". The problem is, they were trash-tier.
#### Jordan
Peterson
Some of them liked Jordan Peterson. One of the meetups
involved watching him debate
Zizek. (Don't watch that, it's
not important.) If you put Travis Scott in front of people and say he's a
good musician, some of them will buy it; I guess the same went for Jordan
Peterson.
He's one of those people whose words have the volume and
the solidity of styrofoam. There are lots of vague statements, but if you
pick one and try to pin down what it means, either they can't defend it, or
it's not relevant, or it doesn't mean anything at all. So such people try to
avoid having the meaning of anything they say pinned down. So if eg Jordan
Peterson can't avoid getting the meaning of something pinned down, they end
up looking rather silly.
Anyway, me taking someone like Peterson seriously would be about as
inefficient as demolishing a styrofoam wall using APFSDS.
These days,
we have AI bands trending on Spotify, and I wonder what effect LLMs are
going to have on the state of such public intellectuals...maybe the
popularity of people like Peterson will fall off now that you can get
**customized** word slop with LLMs, on demand. By the way, if I ever just
quote something from an AI in my response to someone, it's my way of saying:
"The level of this debate is such that, as you can see here, AI output is at
least as good, therefore it's entirely pointless."
####
Richard Hanania
Bob and Joe were both fans of Hanania. Joe had some
personal connection to Hanania, and they picked some book of his for a book
group.
Well, I looked up his twitter, looked up that book, and it was
pretty obviously not worth buying. The last I saw of Hanania, he was just a
generally dumb and bad person who optimizes takes on a monthly basis for
maximum social media clicks, someone that paying attention to will only make
your life worse.
The only semi-interesting thing about Hanania was
that he was a project by some group of people including Tyler Cowen to
establish an associated social media figure, though I wasn't clear on why
they picked Hanania in particular. Well, he *certainly* wasn't getting links
from bloggers early on because he was already influential or because his
ideas were that interesting, so logically there was a different reason. As
soon as I realized that, I could see what Cowen et al were doing.
#### third example
Then, there was some dumb professor who wrote
a book just to say:
> People have culture, and when they immigrate to
another country, sometimes they bring some of their culture with them, which
could potentially affect institutions.
> This is my original theory,
never considered before now, because I am Very Smart and not many people can
handle this kind of advanced theory.
Yeah, congrats on that dude. He
was also just generally a smug BS-generator that annoyed me; it disgusts me
to see work like his put on a pedestal any higher than the average reddit
post gets. Personally, I think that professors with his level of unjustified
arrogance and lack of understanding of existing thinking have negative value
as professors.
#### my suggestions
What public
intellectuals do I recommend, then?
##### touch grass
To
anyone looking for a better Substack to read, my top suggestion would be
this one called "literally just get off the
internet for a while". This
doesn't mean "you'll be instantly amazed at how awesome trees are". What it
means is that mass media often has negative value but fools people into
feeling otherwise.
There are news addicts who spend 6+ hours a day
reading the New York Times / etc. This does not make their lives better, or
even mean they're well-informed about the world. Reading text isn't
inherently bad, and in theory newspapers and Substack posts can be good for
people. The root problem, I think, is the quality of US mass media, in
particular due to the influence of money. Newspapers get sold, they sell
ads, they run articles from PR firms, etc etc. And then the people who work
there are the people who are OK with all that, which drives away a lot of
the people you'd really want to read.
Ironically, I've sometimes
ended up socializing online instead of IRL because I had to go online to
find people that were less terminally online.
So, you can avoid
national-scale mass media, but there are still 2 problems.
The first
problem is that as people have looked for alternative media online,
influence campaigns have spread out. There are bots on twitter, reddit
posters working for social media companies, and notable bloggers might
become shills.
The second problem is that US mass media permeates
into other areas. Were you unhappy with, say, "Trump or Biden" as your
choice in an election? Were you hoping for Iron Man but you got Ironed
Man? Mass media was
responsible for establishing what the options for people to choose were. If
you go to a hipster restaurant, there's a good chance they burn their food
or will feed you quinoa because supposedly it's healthy (it isn't), and
that's also downstream of mass media. It's not like people just
spontaneously decided to start eat quinoa.
Science & tech media isn't
great either. Apart from getting people to buy tech that isn't good or that
they don't need (don't trust Wirecutter, obviously!), it's common for big
science/engineering youtubers to do misleading videos shilling startups. I
don't have a lot of good examples because I tend to hit "don't recommend
channel" if I see a single dumb video, but I do remember:
-
Kurzgesagt not understanding things and sometimes being crackpot
- Real
Engineering being a shill for startups
and as for better ones...
- Mentour Now (aircraft) videos look like clickbait but actually aren't
- Technology Connections (household & obsolete tech details) isn't
particularly important info but at least the guy is sane and usually correct
- Captain Disillusion knows his special effects
- Asianometry has decent
introductions to semiconductor stuff
Is it *important* that you watch
any of those? No, not really, but at least they won't leave you misinformed
and overconfident, which is good enough for me.
There's a fundamental
problem with youtube, which is that good content takes effort and most
people only have a limited amount of interesting stuff to say, but youtube
rewards uploading multiple times a week for years. So in general the most
popular channels aren't going to have the best content.
Sure, some
amount of news/etc is useful. My suggestion, then, is to just mentally
adjust downward the value you feel like you're getting from it in the
moment. If you're browsing youtube shorts on your phone, or reading a news
article that feels just marginally worth reading, try to realize that it's
likely "intellectual pornography" that's specially designed to *appear* to
be worth paying attention to, and then just go do something else. The
internet is full of pied
pipers and you don't want to
end up like the slot machine zombies.
Another suggestion I have is
that, if the culture you live in seems to have some problems, to try to get
some context from outside of it. You could read some old books, or online
fiction from other countries, or maybe discuss strategy games with small
German or Japanese or Russian streamers. That can give you some more
perspective, and you might even learn a language or two.
There's also
no real reason to choose content because it's the most popular. Some guy on
Twitch with 10k followers is probably more sane and entertaining than a
Hollywood actress who got picked because she was in the right city and
attractive and had a good performance on the casting couch. Often, the only
difference between pretty normal people and celebrities is an opportunity.
For example, I remember seeing some girl on Twitch play Counter-Strike for
25 viewers, and stream awkwardly walking to a Japanese restaurant and eating
alone, talking about how she was having trouble finding a job but she was
cooking for her family to contribute something. Then she got picked up by
some vtuber company and got a million+ subscribers. Or, take me - I've gone
back and forth between: unemployed loser with no real credentials,
productive programmer, crackpot inventor, patent writer for corporations,
skilled material scientist, and more. Completely different ways that society
related to me, but I didn't actually change very much. Or take Hunter Biden,
going from "random drug addict" to a board of directors; it all just goes to
show that a lot of credentials are fake and don't matter.
I think
tech media is also somewhat poisoning the intellectual well, smothering the
actually good ideas by promoting memes. People read some posts or watch some
videos, and somehow what they get out of it is something like "the MOST
IMPORTANT BEST THING WE SHOULD BE DOING is molten salt thorium reactors!!" -
which...
1. No.
2. Even if that was a good idea, it wouldn't
be as important as they think.
3. They have negligible influence over
whether that will happen, because the people with the relevant power don't
care about their opinion, partly because...
4. The only useful thing
about studying niche technologies like that is developing an underlying
understanding of the technologies and physics it relies on, but did they
learn anything about mechanisms of metal corrosions by metal salts? About
the economics of supercritical CO2 cycles vs steam turbines? About T-S
diagrams, or thermal pinch analysis? No.
5. It's completely
irrelevant to their life.
This reminds me; there's a blogger called
Zvi, a reasonably smart guy who decided to
try pushing for some slightly-positive government policy thing that wasn't a
right/left thing and nobody would care about. He picked the Jones
Act which says
US-to-US ship transport has to use US-made ships. I told Zvi that wasn't
"pulling the rope sideways" as much as he thought, and wasn't going anywhere
as an actual policy, but he disagreed. And recently, I've seen comments on
eg Hacker News, saying things like: "the Jones Act is what's responsible for
the decline of US shipbuilding, that's why we need to repeal it". Not
exactly accurate, and it's still not actually leading to a policy change,
but...congratulations, I guess. That's what you wanted, right?
##### The proposed scope is too broad.
People have limited
areas of expertise, and you need to be able to judge the scope, beyond just
general themes of blogs. Take the good, leave the bad. Single role models,
copying someone's views as a whole, these are things that came from ancient
times with villages and hunter-gatherer bands. These days we have the
internet, an unlimited supply of writing and video from billions of people.
As they say, woe to the thinker who is not the gardener but only the soil of
the plants that grow in him.
Yes, it's true, back when I was in 2nd
grade, I wanted to be like Feynman, but then I grew up. Now, what do I mean
by that?
- The character in Feynman books isn't exactly who actually
existed.
- Universities changed since Feynman's time, and I eventually
realized I didn't want to get a PhD or try to become a professor.
- The
state of theoretical physics had also changed, and around 6th grade I
decided that there was something suspicious about string theory and maybe I
didn't actually want to be a physicist.
- There's no ongoing Manhattan
Project, which was essential to his career path.
To the extent that
there are modern equivalents of the Manhattan Project - which I suppose
would be things like modern AI, manufacturing techniques for military UAVs,
and bioweapons - I don't trust current US government or corporate leadership
all that much. And in fact I did once turn down an AI-related job for
ethical reasons. The Manhattan Project made nukes to drop on Japan, but I'd
probably rather make, say, Taiwan-deterrent military UAVs for Japan than the
US, and where does that put me as a US citizen? (Not that Japan would want
to hire me for that. Plus the constitutional issues.)
##### start
with empirical things
If you need some expertise in a field you can't
evaluate, my advice is:
1. Start with people who are interested in
the topic and good at something you can evaluate empirically, even if it's
not directly related to the topic. Maybe some material science, or
chemistry, or reaching grandmaster in Backpack Battles, or getting a
70-streak in a difficult roguelike game, or making their own video game.
Preferably multiple things.
2. See who they recommend as experts on
topics they're interested in, and check out those people.
### just rationalist things
Some misc things I noticed about
rationalist people in general:
#### candles
A lot of them
liked burning candles indoors. Now, this isn't uncommon in America, it's
just that spending money to make smoke in your house with no real benefit
isn't very...**rational**, is it now? Some of them even did some silly
ritual thing involving lightning candles for "Petrov
Day".
These days, if somebody's house has
candles burning in it, I'm turning around and leaving. People dumb enough to
do that just aren't worth putting up with their air pollution.
#### betting on predictions
A lot of rationalist-associated people
seem to want to track who said what for later scoring. I'm not generally a
fan of that because:
1. The purpose of discussion isn't, in general,
to establish an intellectual pecking order. Me, I only take arguing
seriously to the extent that one of these is true:
- the other people
are bringing serious arguments
- the outcome of the discussion actually
matters, and I don't mean reddit karma
2. If you must score and rank
people, it'd be better to use either Metaculus predictions or something more
empirical like science or video games.
But one problem with rating
people on Metaculus-style predictions is: What top Metaculus predictors are
really good at is **mixing**. Even if you're pretty good at guessing things,
for any given question, the truth is generally going to be somewhere between
your guess and an average of other guessers. So you want to average together
multiple people's predictions, but the thing is, if those people are already
giving you pre-averaged guesses, that's less helpful: you then need to
compensate for an unknown amount of mixing with unknown other positions, and
you don't know if multiple people are copying from the same source or
different sources.
#### detachment from the practical
This
article
came out a few days ago, and I'd like to quote some of it.
> Here is
a sampling of answers from people in and close to dysfunctional groups: “We
spent all our time talking about philosophy and psychology and human social
dynamics, often within the group.” “Really tense ten-hour conversations
about whether, when you ate the last chip, that was a signal that you were
intending to let down your comrades in selfish ways in the future.” “Like
somebody's going through a real bad breakup and you need to, like, hold
their hair out of the toilet over text, tell them that it's going to be okay
and help them put their life back together, except for years.”
>
Conversely, when I asked interviewees in high-demand groups that wound up
basically functional, like Benton House, they gave answers like: “I learned
to drive, which was really important. I got my first programming job. We
played calibration games and a rationalist version of Family Feud called
Schelling Point.” "I wrote LessWrong posts and started a rationalist fiction
novel, helped someone immigrate, and once scrambled like thirty eggs in a
wok."
...
> Rationalist groups tend to be functional to the
extent that their activities involve learning to program, writing papers for
general publication, building a giant dome on the playa, or otherwise
interacting with the real world. Rationalist groups tend to be dysfunctional
to the extent that their activities involve very long conversations about
human psychology and social dynamics, especially dynamics within the group
itself. Relatedly, the clearest-cut cases of rationalists being right have
involved external events in the world and not the nature of human beings.
Maybe you could say something similar about university departments.
Anyway, I try to push discussions towards more practical things, towards
actually doing something, but it seems like a lot of "rationalists" never
learned how to have a normal intellectual conversation, so I thought I'd
give a quick example here.
CORRECT:
Alice: Hello, I'm Alice,
some people know me for [stuff I made, eg a blog or a video game]. I'm
interested in [range of activities] here.
Bob: [comment showing
understanding some implications] [project representing Bob's interests]
[related project of Bob's friend]
Alice: [allusions to potential
collaborations at a later stage] [indication that the introduction phase is
complete]
WRONG:
Alice: Hello, I'm Alice, some people know me
for [stuff I made, eg a blog or a video game]. I'm interested in [range of
activities] here.
Bob: Cool, sounds like you'll have something to
contribute to our discussion of my Substack subscriptions.
Alice: No,
I don't want to hear about your fucking Substack subscriptions.
## conclusion
In retrospect, I probably should've just
stopped organizing meetups and done nothing instead of actively passing
meetup stuff over. My bad.