Just this guy, you know?
Dagon
Karma: 11,652
With some further thinking, I’m realizing one of my concerns is that this is an unfortunate direction for LW. I LIKE that LW is text-heavy. I LIKE that the norm is to use complete sentences, and to leave actual comments when there are questions or clarifications to the main point.
I don’t think adding more mechanisms for low-effort low-information-content (what, 5 bits per react?) will make LW better, and in fact could make it much worse, if it substitutes for some amount of comments. Cutesy icon debates are fun, but don’t actually add value.
[ note: partly posted to see if I can get some downvotes or negative reacts. but also because it’s a real concern, though I don’t think I understand the dynamics well enough to know if this is actually a problem for this feature. ]
Upvoted for being an important idea, but I actually disagree with the advice. The relationship of ideas to action is exceedingly complex, and I strongly doubt (but do not know how to test) that the idea simply hadn’t occurred to someone who wanted attention through harm.
I find it much more likely that there’s a large uncertainty in the effectiveness (in terms of attention to be had) of uncommon attacks, and when it’s not already in the public eye, it’s known but not considered as a reasonable mechanism. Much like cryogenics is weird and uncertain, even for people who would like to be revived, poisoning medicine (dangerous idea: why only medicine, not other foods?) was weird and uncertain until it had been shown to work.
I suspect the dangerous information is that it has succeeded at least once, and gotten a lot of press attention. This information is much harder (and less desirable) to suppress.
In the software world, ideas are rampant and cheap. Execution of the correct idea is the path to success. I expect it’s similar as a terrorist, except there are way fewer people to help you choose, refine, and change your ideas, so you only get one shot (as it were).
I also note a similarity to the disclosure debate about computer vulnerabilities—there’s a tension between publishing so that potential victims can protect themselves or watch for attacks vs keeping quiet so vendors can fix underlying bugs before very many attackers know of it. There are a LOT of factors that go into these decisions, it’s not as simple as “don’t spread harmful information”.
Another example, which I don’t know if it supports my position or yours: Tom Clancy published Debt of Honor in 1994, which included a near-decapitation of the US government by a pilot-turned-terrorist flying his 747 into the capital building. Only 7 years later, real-life terrorists did something very similar. We immediately instituted systems to prevent repeats (and a bunch of systems that added irritation and did not protect anything), and there have been no copycats for 17 years.
Unfortunately, there’s no way to publicly examine, measure, or discuss group differences in a way that doesn’t disproportionately attract those who’d misinterpret and misuse this against the group(s) in question, and therefore without those groups legitimately feeling attacked by those giving visibility to the topic.
It’s the core of politics (how to characterize groups and their relation to each other), and therefore rationality in hard-mode. There are very few forums where it won’t cause more harm than good.
In fact, we don’t have good epistemic practices for studying or thinking about groups, as distinct from individuals. I suspect there is a lot of reality in the self-organized sortation by visible traits, and this does cause group-norms to differ by distribution of invisible traits. But I don’t know of many studies that break it down that way.
I don’t like that this conflates upvoting someone’s writing with trusting their voting judgement. The VAST majority of upvotes on most systems comes from people who don’t post (or at least don’t post much), and a lot of posters who I like their posts, I disagree pretty strongly with their liking on other topics.
More importantly, I think this puts too much weight on a pretty lightweight mechanism, effectively accelerating the goodhart cycle by making karma important enough to be worth gaming.
I’d like to support the “do nothing” proposal.
Karma is destined to be imperfect—there’s no way to motivate any particular use of it when there’s no mechanism to limit invalid use, and no actual utility of the points. It’s current very simple implementation provides a little bit of benefit in guiding attention to popular posts, and that’s enough.Anything more complex and it will distract people from the content of the site. Or at least, it’ll distract and annoy me, and I’d prefer not to add unnecessary complexity to a site I enjoy.
At this point, if we can’t undo mass-voting, it would be better to just disable downvotes or voting entirely.
It’s not the only cause of LW’s decline, but it does contribute.
Self-sacrifice is a scarce resource.
I frame it a little differently. “Self” is the scarce resource. Self-sacrifice can be evaluated just like spending/losing (sacrificing) any other scarce and valuable resource. Is the benefit/impact greater than the next-best thing you could do with that resource?
As you point out in your examples, the answer is mostly “no”. You’re usually better off accumulating more self (becoming stronger), and then leveraging that to get more result with less sacrifice. The balance may change as you age, and the future rewards of self-preservation get smaller as your expected future self-hours decrease. But even toward end-of-life, the things often visible as self-sacrifice remain low-impact and don’t rise above the alternate uses of self.
I’m old, and my memories of being young, poor, and having multiple roommates I don’t know and love are rather distant. I never saw co-living as particularly attractive, and I’ve long been in the “good fences make good neighbors” camp, so I’m likely not the target of this warning.
But I think there are some good, general, pieces of advice that could be extracted from this—I’d frame them as recommendations, rather than as warnings. They apply to ANY sharing of significant parts of your life, not just Bay-area group houses. I wish I’d read it before going into business with some acquaintances.
Don’t do it for the money. It’s fine to save money when possible, but that should never be the primary reason for a living arrangement. Do it because you expect to enjoy and be satisfied with the daily experiences.
Culture and “fit” matters more than you think. Don’t go into it without knowing what’s important to you and what’s important to your partner(s).
Yes, “partner(s)”. Anyone you’re living that closely with on a day-to-day basis is more than an asset-share arrangement. You have shared goals and habits that affect each other’s happiness.
Have an exit plan (preferably multiple). Know what you will do if it becomes too painful.
Have exit triggers. Know what “too painful” means. Decide in advance how bad it’ll get before you have to live in your car (or whatever your worst-case exit plan is).
Have a way to track positive and negative experiences. Don’t trust your memory to know what happened last week. This is key to staying somewhat objective about the exit triggers. You don’t want to be too objective (and you really don’t want to use this as evidence to beat your housemates up), as this is all about your emotional evaluation of your living situation. But you also don’t want to overreact to temporary problems.
upvoted for interesting ideas and personal experience on the topic. If I could strong-disagree, I would. I do not recommend this to anyone.
Mostly my reasoning is “not safe”. You’re correct that historically, the IRS doesn’t come at small non-payers very hard. You’re incorrect to extend that to “never” or to “that won’t change without warning due to technology, or legal/political environment”. You’re also correct that, at current interest rates, it’s about double at ten years. You’re incorrect, though, to think that’s the biggest risk. If they decide there’s a pattern that shows intent, penalties become MUCH higher. Unlikely to include jail time, but in terms of net expectation over all possible universes, it’s zero or somewhat negative for almost everybody. Admittedly, weighted toward “most get away with small amounts, a few have a VERY BAD experience”. But you could recreate that in Vegas with a simple 3- to 6- step martingale.
I also warn about lifestyle-distortion effects. Unless you reverse course and pay up (which is probably possible “just” by paying the back taxes and penalties, until some IRS investigator decides to start seriously documenting your intent), you can’t take a good W-2 job, can’t invest in real-estate, and need to keep a low enough profile that the IRS doesn’t decide to actually go after you.
Thinking a bit more about the times when I nitpick a technicality rather than digging into a real disagreement:
Both of your examples are cases where there’s a LOT of subtext (friend implying status and/or attempting to influence my purchase decisions, grandma asserting authority and enforcing her preferences by implying universality), and much of the time I won’t want to risk the relationships by directly contradicting or correcting them, _NOR_ by walking away and never seeing them again. And yet, it grates on me to take the option you didn’t call out: just meekly accept the underlying disagreement. I totally get that the initial-response mechanism of pointing out that they’ve used a word in a way that doesn’t exactly match a dictionary does not further a rational discussion, but I’m not sure a rational discussion is what these examples are.
Pedantry like this _is_ a way to assert a little bit of independence/disagreement (or, less justifiably, dominance), and to open the concept of disagreement in a way that’s deniable, and start a subtle, unacknowledged negotiation which can be de-escalated easily if either party decides it’s not worth pursuing.
Do keep in mind that if a friend actually follows through, you’ve significantly raised the stakes of saying “no” later.
[note: I don’t consider myself Utilitarian and sometimes apply True Scotsman to argue that no human can be, but that’s mostly trolling and not my intent here. I’m not an EA in any but the most big-tent form (I try to be effective in things I do, and I am somewhat altruistic in many of my preferences). ]
I think Alice is confused about how status and group participation works. Which is fine, we all are—it’s insanely complicated. But she’s not even aware how confused she is, and she’s making a huge typical mind fallacy in telling Bob that he can’t use her preferred label “Utilitarian”.
I think she’s also VERY confused about sizes and structures of organization. Neither “the Effective Altruist movement” nor “rationalist community” are coherent structures in the sense she’s talking about. Different sites, group homes, companies, and other specific groups CAN make decisions on who is invited and what behaviors are encouraged or discouraged. If she’d said “Bob, I won’t hire you for my lab working on X because you don’t seem to be serious about Y”, there would be ZERO controversy. This is a useful and clear communication. When she says “I don’t think you should call yourself Utilitarian”, she’s just showing herself as insecure and controlling.
Honestly, the most effective people (note: distinct from “hardest working”) in sane organizations do have the most respect and influence. But that’s not a binary, and it’s not what everyone is capable of or seeks. MOST actual humans are members of multiple groups, and have many terms in their imputed utility function. How much of one’s effort to give to a given part of life is a pretty wide continuum.
I did a lot of interviewing and interview training for a former large employer, and an important rule (handed down through an oral tradition because it can’t really be written down and made legible) was “don’t hire jerks”. I’d rather work with Bob than Alice, and I’m sad that Alice probably won’t understand why.
Absolutely common. Most non-disparagement agreements are paired with non-disclosure agreements (or clauses in the non-disparagement wording) that prohibit talking about the agreement, as much as talking about the forbidden topics.
It’s pretty obvious to lawyers that “I would like to say this, but I have a legal agreement that I won’t” is equivalent, in many cases, to saying it outright.
I think the problem is that it’s not actually a good analogy, and EY made an error in using the current event to amplify his message. AFAIK, there’s never been anything that has to be perfect the very first time, and pointing out all the times we chose iteration over perfection isn’t evidence for that thesis.
The fact that there are ZERO good past analogies may be evidence that EY is wrong, or it may not be. But Matt Levine definitely has an advantage in communication that he can pick a new example (or at least a new aspect of it) every day for the 10 or so themes he repeats over and over. EY has no such source of repeated stories.
I draw a few more lessons from this (and from conversations with other survivors and escapees from horrific regimes):
6. Change is both gradual and terrifyingly fast—there is often months or years of buildup and warning, before weeks of crisis.
7. Terrifyingly fast is not instantaneous. It costs a lot, but one can get out if one actually believes the evidence in time.
If I make a comment, then the author deletes the post, is my comment lost as well? I’m pretty sure I don’t actually need a record of things I’ve said, and that my comments aren’t valuable enough to justify extra effort, but it kind of bugs me.
Has anyone written the scraper or script to save a copy of everything you write on LW, so it’s available even if it gets deleted later? If it archived the post and comment thread you’re responding to, it would be all right.
This is worth thinking about, but a reminder to participants that the space of possibilities is larger than we can imagine, and FAR larger than we can write compelling stories around.
I think my preferred story is NOT that it’s recognizable as Humanity vs AI. Unless the FOOM is worst-case (like many orders of magnitude literally overnight), it’s likely to look much more like AI-assisted humans vs other humans’ preferred cultural and legal norms. This will be a commercial AI getting so good at optimizing that company’s control of resources that it just seems like a powerful company pursuing goals orthogonal to customers/consumers/victims. Which then manages to lobby and advertise well enough that it gets MORE power rather than less when subgroups of humans try to fight it.I think it’s almost guaranteed that in the early days (or decades) of AI takeover, there will be lots of human Quislings who (wittingly or un-) cooperate with the AI in the belief that it benefits them (or even think it benefits humanity overall).
Think “FAANG team up to take over the world with more AI assistance than you expected” more than “AI takes over violently and overtly”. AI won’t care about status or being known to rule. It’ll only care about the goals it has (those it was created with plus those it’s learned/self-modified to want).Also, this may already be happening. It’s hard to distinguish a smart, driven AI from a successful corporate behemoth. In fact, the argument can be made that corporations ARE agents in the AI sense, distinct from the cells/humans who they comprise.
Should we show the agreement number as a ratio rather than a sum? regular votes can be summed, because “low total” doesn’t matter much whether it’s a mix of up and down, or just low engagement overall. But for agreement, I want to know how agreed it was among those who bothered to have an opinion. Not having an opinion is not a negative on agreement.
I think I’d either show total and number of votes (as 20 / 12), or just the ratio (1.66).
edit: I may get this in the current setup from looking at agreement compared to karma, once I get used to it. But that makes it worth aligning the default self-votes for the two, so comments don’t start out controversial.
Thanks for posting this—reports of experience are interesting and useful. I advise caution. That style of emotional belief is useful in motivation, and is a good hint toward areas to model more closely and prioritize in terms of actions. But it’s also over-general and under-nuanced, and lacks humility and acknowledgement that it might be incorrect.
I think this incorrectly mixes “lizardman”, an unexplainable component of low-stakes polling, with “minority strongly-held stupid (to me) opinion”. I think most of your examples have more than 5% support, especially if you count “don’t really care, but I’m uncomfortable with the cluster of ideas that contains this”.
I agree with you on most of the specific issues, but it’s an error not to recognize that there are a whole lot of real humans who actively are on the other side.