roystgnr
4mo ago
Thematically I wouldn't say Lovecraft emphasizes alien power so much as human powerlessness; it's just that those are two sides of the same coin. His is a universe where humans are alive not because we're strong enough to survive so much as because we're weak enough to go unnoticed.
The first Gateway book is probably a good answer in the same sense as the Chernobyl miniseries; I'd say the final book in the series ("The Annals of the Heechee") is a good answer in other ways. I don't think I can say much more about the distinction without spoilers for both, though.
The rest of the Gateway series isn't as good as the first book, though - bigger ideas but without the same depth of characterization.
roystgnr
4mo ago
I was expecting a link to this. "I'm gonna kick some ass with my own pipe wrench..."
roystgnr
4mo ago
One of the ideas I liked from LessWrong was that of "anti-inductive systems": systems where a proposition about them can become false via the process of discovering it was true. At first glance the idea sounds paradoxical; then at second glance you can come up with a couple cases where it's so obviously correct you don't need to think about it, but there are cases in-between where consciously worrying about the problem is helpful, and I think "don't take your own or your group's own virtue for granted" encompasses a whole category of them. "I'm too rational to fall for simple human biases" is like "we can trust the priests/feminist-allies/cops/non-profits/Crusaders/etc. completely" - even if such a proposition is actually true at the moment you conclude it, as soon as you come to believe it you're likely to drop your guard too far and eventually be betrayed by your own overconfidence.
I think that argument just isn't supported by the evidence of 40 years of computers more or less working just fine.
Just looking at the last 40 days of computers is enough to support "less" rather than "more". My favorite from quickly skimming reports from the last month or two would probably be Rsync contains six vulnerabilities: "When combined, the first two vulnerabilities (heap buffer overflow and information leak) allow a client to execute arbitrary code on a device that has an Rsync server running. The client requires only anonymous read-access to the server, such as public mirrors. Additionally, attackers can take control of a malicious server and read/write arbitrary files of any connected client. Sensitive data, such as SSH keys, can be extracted, and malicious code can be executed by overwriting files such as .bashrc or .popt"
The vulnerabilities are "present within versions 3.3.0 and below" of software that has been heavily used for nearly 30 years now. I'll agree with you that this is "thanks to the internet", but can we really call it "suddenly" now that the internet is four decades old, especially for cases like this where the software was specifically written to make use of the internet? I'm sympathetic with (and a perpetrator of) mistakes of the form "Joe writes a program that reads a file, and expects it to be used by very clever people reading only files that they and their personally-known friends/coworkers created, but then it becomes more popular and now every nescient email user who double-clicks strange attachments is one JoesInstaller away from putting a backdoor on their computer." But for web browsers, mass-market software, servers listening for arbitrary TCP connections, etc., surely there's a better solution to e.g. heap buffer overflows than expecting every software author to finally Git Gud.
roystgnr
5mo ago
Beat Saber seems to be the most popular, and I definitely had fun with it. I prefer Supernatural, but that does have a recurring fee.
roystgnr
5mo ago
The Night Court theme is asked the question of "how can you possibly give an appropriate introduction to such an awesome show", and yet manages to answer with a confident "don't worry, I got this".
roystgnr
6mo ago
I think math just inherently requires more structure
Khan Academy has pretty much solved this problem for pre-college math.
and pushing
But here ... yeah, it really depends on the kid. Letting our kids work independently (along with a charter school that is very flexible), my son got 5 or 6 years ahead in math because he loves math, his older sister got a couple years ahead because she loves the idea of getting into a good college, and his younger sister just does what she's asked to because she loves her mommy, which is going to put our "don't be pushy parents" philosophy to the test in the coming years.
roystgnr
6mo ago
How is that not a threat? It's just "If you do that when I'm able to kill you, I will" but in subjunctive rather than future tense. Either way, after the criminal's release the victim has to choose between avoiding the threatened speech or getting killed.
roystgnr
7mo ago
I'm not even sure if that's a federal crime. The Secret Service generally won't commit crimes, and they aren't generally treated as co-conspirators or anything if they witness crimes and decline to intervene, although in theory they're at least required to report on it after the fact and testify about it if subpoenaed. This doesn't come up much for obvious reasons, but IIRC when one of GW Bush's daughters was drinking underage there was a Secret Service guard witnessing it without stopping it, and the justification was basically "if we interfere she's just going to decline or ditch protection and then she's neither sober nor safe".
But even if they couldn't turn the situation into some "manipulating the Secret Service for X" charge before and definitely can't do so now, it still can't be thrown out of civil court due to a criminal pardon, and a civil judgement could still be enforced.
roystgnr
7mo ago
Thanks for the feedback! "Be a jerk more"; check!
"surely, they're not going to try to land this thing on a coin toss?"
The only alternative I can see piles conspiracy theory on top of conspiracy theory: start with "The White House pushes to get FAA approval before the election so Elon can publicly embarrass himself" but then checkmate it with "Elon knows they would pull something like that if they saw the chance and so expresses grave doubts even though he knows they'll succeed".
Better a false-positive-driven crash into the ocean, than a false-negative-driven crash into the spaceport.
Definitely. This is what they do with Return To Landing Site booster landings for Falcon 9 too, and those are just aimed at slabs of concrete, not expensive ground support equipment. They had one splash down back in 2018, when a grid fin actuator failed and it didn't/couldn't turn toward the landing pad. SuperHeavy is designed to be much easier to land than F9, but it'll also be a much bigger kaboom when a catch fails.
The big risk they're taking is with upper stage catches. With a booster return, it's flying over ocean all the way, and if it can't fly any more than it just goes in the water like any other rocket company's booster would. Landing the upper stage on an east-coast launch tower, on the other hand, requires it to reenter from the west over land. When that works it should work fine. IFT-5 managed to drop right on top of camera buoys pre-positioned to film the action. But IFT-4 just had one mostly-lost flap and ended up about 5 miles off target, and back in its day the shuttle Columbia scattered debris over a 250-mile-long swath of Texas. Starship is a hardier design than the Shuttle was, and any debris would probably include some scarily huge chunks. Just this year SpaceX started exclusively bringing Dragon capsules back to the Pacific rather than also to the Atlantic, because the discarded "trunk" rings that were supposed to be flimsy enough to burn up on reentry turned out to have too-big chunks of debris reaching the ground too frequently. SpaceX really can't afford to fail an upper stage RTLS Starship reentry, not until they've got a west-coast or island-based or ocean-going catch tower to practice with afterwards, and they have no near-term plans to build any of those.
roystgnr
8mo ago
Oh, I didn't mean to imply that he was faking.
IMHO Trump's only clear advantage over most Republican candidates is that, despite his very loose relationship with honesty in general, he will occasionally turn on "I'm going to piss off everyone who disagrees with me" mode as a costly signal of sincerity. Kamala's trying to backpedal on her previous statements on things like policing and border security in very weasel-worded ways that are obviously intended to barely placate moderates while being easy to un-backpedal from later. Vance's "Trump would have won except for big tech" phrasing is a great way to say something technically true without either actually agreeing with Republicans who think it was vote fraud or openly disagreeing and pissing them off. But when Trump goes full on anti-illegal-immigration or 2020-was-fraud or whatever he doesn't use his charismatic-real-estate-negotiator language, he brings out scare-the-normies level language that he's clearly not going to back down from. He might be mistaken but he's not lying.
My point is just that relying on him alone to fix any problems is strictly inferior to relying on both him and on grassroots-level efforts too. Maybe some of the evidence that convinced him was spurious (the reason I'm not convinced is that I waded through enough of that) and he'll go after those red herrings and never get around to other real problems, so the only way to get real problems fixed is to publish the evidence. Maybe he won't get reelected because that's just not happening again, and fixing any vulnerabilities will have to be done by others, so the only way to point out what to fix is to publish the evidence. Maybe he would have been reelected if he made a strong case for voter fraud evidence, but he didn't, so there are people who would have voted for the anti-voter-fraud Trump but will vote against the weird-fake-electors-thing Trump. Maybe he will be reelected and will go after real problems but will be thwarted by federalism or another branch or the deep state or whatever and fixes will have to come in at the local level. There's just so many ways that Trump making a public case could make things better. It seems like the one big risk here for him is that putting everything out in the open might reveal that none of it is convincing, but that's also a situation where "good for Trump" vs "good for the country" diverge, and I'd be on the "good for the country" side in that case.
roystgnr
8mo ago
Who said anything about court? Your theory that judges are more likely to dismiss people who publish more evidence is an interesting one, but there is a reason why I said "publish", not "file". As I admitted, if the judges dismiss you then you still lose 2020, but if the voters don't then your team wins 2022 and 2024 and a lot of opportunities to prevent whatever fraud you detected from happening again. Maybe it takes time to go through evidence in the moment and make sure you're not hurting your credibility by putting credence in bad evidence too, but after three or four years have passed Vance shouldn't be dancing around when asked if Trump lost, he shouldn't be pointing to social media censorship (or whatever "big tech rigged the election" meant) as reasons why Trump morally won, he should be advertising "trumpwon2020.com" or whatever URL they picked to host all the evidence they have that Trump actually won.
roystgnr
8mo ago
Start with a standard Manhattan recipe, but not only do you make sure not to forget the Maraschino cherry, you also add a dash of Maraschino liqueur and a dash of Cherry Heering.
Weeski: 1 1/2 oz Irish whiskey, 1/2 oz Cointreau, 1 dash orange bitters, 1 oz Lillet Blanc, orange peel (candied since I'm lazy) garnish.
Mint julep, but always with granulated sugar instead of syrup (to muddle the mint leaves better), and about twice as many leaves as the standard recipe calls for.
roystgnr
8mo ago
The orange peel garnish actually seems to make a difference to an Old Fashioned for me, but I'm too lazy to keep fresh oranges on hand and carve peels for cocktails, so I've recently got a jar of candied orange peels for the purpose, and that seems to add 50% of the value for 5% of the work.
roystgnr
9mo ago
There's lots of high-end options out there these days. We tried delivery and curbside pickup in 2020, and the former wasn't worth the price but the latter is still how we do our regular grocery shopping. A few bucks extra, and we can't pick produce ourselves, but saving 40 minutes per trip is usually more than worth it.
roystgnr
9mo ago
It would be even more true of a workers' coop when considering the market for the provision of goods. The place workers' coops fails is, ironically, when considering the market for the provision of workers' compensation. Putting your investments in a single business instead of diversifying is in general a bad idea; making that business be the same business you rely on for a salary makes it an even worse idea. Maybe not for the customers, because from their perspective the workers' incentives are now just about as well-aligned as they possibly could be, but when considering the workers or even just considering both together it's hard to beat keeping most wealth in index funds rather than a "pray to God this one company doesn't go under" fund.
roystgnr
9mo ago
I'd also like to know the answer to that question.
IIRC the likely-better short-term alternative to sulfate aerosol is calcite aerosol (so the main side effect is to reduce rather than increase ozone depletion), and the likely-better long-term alternative is enhanced rock weathering (to actually get excess CO2 out of the atmosphere rather than just papering over a few of the problems it causes), but IIRC they're even further back in the theoretical/experimental stages.
roystgnr
9mo ago
It's clearly not close to the norm in the sense of "normal distribution", but it's closer than it should be in the sense of "normative". Back when the pregnant woman got run over you could find de jure support for the victim in pre-existing Arkansas state publications, and the state police settled out of court later, but if the cop who ran her down wasn't fired then in some de facto sense wasn't any norm against that superseded by a "let the vehicular assailant get away with it if they're a cop" norm they consider more important?
roystgnr
10mo ago
it seems like you would if you could.
Forgive me for only skimming this discussion, but is there some other comment that gave you this impression? I don't get it from what you quoted.
It's hard to talk about threat in the capabilities sense without any subtext of threat in the intentions sense, but I don't perceive any deliberate subtext.
roystgnr
10mo ago
Reaching back 60 years to find an example seems like strong (if unintended) support for both "certainly possible" and "it doesn't happen often".
roystgnr
10mo ago
To be fair, it's not like the flair itself has a hyperlink. In context it's a weird cross between a justifiable brag and a plausibly deniable cry for help, but out of context it does look hostile.
roystgnr
1yr ago
We were discussing ability bias vs human capital - did anyone bring up signaling before I did? It seemed a very weird thing to leave out of the conversation, so I thought it was surely worth mentioning that it could be nearly as large an effect as ability bias despite falling on the opposite side of the "should I go to a cheap college" question.
But as long as I'm bringing up weird things to leave out of the conversation - what's your source for "Once you control for that, the differences are pretty slight"? I was providing what seemed to be a relevant counterexample to an assertion not yet tied to data, but if you do have more relevant data then that's a trump card - just go ahead and play it?
As a parent:
For "mathy":
The best thing for my son was Khan Academy, which basically covers everything you'd want in grade school, up to a little basic undergraduate-engineering-major-level math. When Covid hit and all the other kids' brains started atrophying, he instead was thrilled to discover he could now go as fast as he wanted (or occasionally as slow as he needed to?) He started squeezing it in to the sad little "online school" schedule that had been hectically thrown together, then he eventually got permission from a teacher to let his little sister do his "real" math homework so he could spend more time studying way ahead, and he got far ahead before testing out of a few physical classes and joining others. (or "auditing" others; they have to call his Calculus class "independent study in math" due to some age/grade restriction, but fortunately there's no restriction on the AP tests).
When he got a bit older he really started getting into math competitions, so instead of just racing ahead he's spent a lot of time getting better at the sorts of questions that a typical kid his age knows enough math to understand but still can't necessarily solve. These have also been helpful at avoiding ego inflation, putting him up against the sharpest math students in the city or the country rather than just a few classmates and a standardized curriculum.
We've done a lot of talking together about the basics of things like set theory, boolean algebra, group theory, linear algebra. There's a lot of math that's understandable to very small kids but that doesn't get covered in a standard curriculum. This is reasonable of the standard curriculum, since most non-scientist non-engineers will never need to know e.g. what a power set is, and even most scientists and engineers can get away with believing theorems without picking up the groundwork to prove them themselves ... but if you've got a kid who's interested in math, then she may be interested in math enough not to worry too hard about which of the things she's learning have what future applications. Being able to learn multiple things at a time can also be helpful if one doesn't "click"; even professional mathematicians often just dive into one subfield they really enjoy, or end up mostly on one side of a divide like the "algebraist-vs-analyst" rift.
For "daughter":
I'm actually not sure? I might have screwed this one up badly somehow! My oldest daughter has a great talent for and a great dislike for math. She's taking Calculus at 15 and doing a couple math competitions, but solely to spruce up her college applications and get ahead on engineering major requirements. Watch from 4:55 through 6:10 of NewsRadio "Houses of the Holy", but imagine a girl doing math instead of a man doing magic tricks.
My youngest doesn't really dislike math, but she doesn't have any of the interest her brother does. She's joined him in a math club, but whereas he's there for the "math" part she's there for the "club" part. Your daughter might not have the stereotypical people-vs-things gender difference, but for any child having a community to work with and (to a limited extent) compete with can be strongly motivating.
As a former mathy child:
I really wish I'd grown up with more resources like Khan Academy, but also Wikipedia and especially math YouTube. The adults around me all tried to help me hit my full potential, to great effect, but there are interesting subfields that I didn't even know existed until I stumbled across references to them somewhere else.
I wish I'd somehow challenged myself more in high school. I took a few night courses at the local university after exhausting the available high school math classes, but still had a sudden shock when I got into a selective university and could no longer just ace every math test without serious study first.
I ended up in applied math, after nearly just going full engineering, and in hindsight this was more of a good idea than a mere diversion. The state of the art in pure math is so far ahead of us lowly applied mathematicians, and I think it's good for my morale to come up with ideas that I can then immediately implement and use, rather than ideas that for all I know might be foundational to 22nd century physics but that might more likely then be nothing more than footnotes in papers nobody reads. And despite what I said above about proper math brains not caring about future applications, it's still easier for me to remember new ideas I learn if I can immediately see a few ways to apply them to something connected to reality rather than if they just feel like a neat self-contained game.
More options
Context Copy link