@raakaa's banner p

raakaa


				

				

				
0 followers   follows 0 users  
joined 2023 May 27 23:20:53 UTC

				

User ID: 2428

raakaa


				
				
				

				
0 followers   follows 0 users   joined 2023 May 27 23:20:53 UTC

					

No bio...


					

User ID: 2428

Perhaps I’m misinterpreting the type of “risk-averseness” that you’re talking about, but to me, the relationship between the stability of one’s current situation and one’s willingness to take risks is the exact opposite of your assessment. If you’re already in a perilous situation, then you’re so close to rock bottom that the potential upsides of a risky endeavor far outweigh the potential downsides. Vice versa for a comfortable situation.

This is clearest in sports: in football, it’s usually not the team leading by 10 points with 3 minutes left in the 4th that will throw a Hail Mary or kick an onside kick. In hockey, you never see a team in the lead pulling their goalie.

Man, it’s great that all these posts from back then are still around. It’s not quite as good as your example, but I’m personally partial to the one that says “Nobody’s getting arrested”. Ages like fine wine, if that wine came out from the udder of a cow.

This was a joke. They were explicitly and intentionally mocking the concept of conservatives saying that.

I remember all the “conservatives pounce” headlines from the Gray Lady and her friends making this argument, but I’ve never bought it. Consider a world where, say, the chair of the Texas RNC or whatever puts out a pro-Trump campaign video where they say “When we win the election, we’ll go stage a little putsch, / Suspend your sad democracy and voting rights and such”. You know, to make fun of liberals’ overblown fears of a Trumpian self-coup. Do you really think that progressives like the San Franciscan choir and their defenders would accept this as just a wee bit of humor poking fun at their neuroses, rather than a serious Threat to Our Democracy?

EDIT: I realize that I forgot to clearly make my point: if you’re in a position to do things that other people think are bad, and you state “hey, we’re gonna do those things that you think are bad”, then you shouldn’t be surprised when people take you seriously. And those people would be right to wonder whether you’re just joking or going “haha, only serious”.

But I’ve seen reports that the average age of cars on the road in the U.S. is ~13 years old. Assuming that this is referring to the median (and despite finding countless articles repeating the “13 years old” statistic, I haven’t found one that specifies if the average in question is a mean or median), then this suggests that post-COVID-lockdown effects on the distribution of colors among cars on the road can only be marginal at best. And even if the statistic refers to the mean, I doubt that outliers have too much of an effect here.

Of course, your explanation does perfectly answer the other part of your interlocutor’s question, why new cars only come in grey.

The terms “modus ponens” and “modus tollens” come from formal logic. They tell you how to deduce conclusions from statements.

Modus ponens says that if you know that A is true, and that B is true whenever A is true, then you can deduce that B is true. For instance: “If someone is a hunan, then he is mortal. Socrates is a human. Therefore, he is mortal.”

Conversely, modus tollens says that if you know that B is true whenever A is true, and you know that B is not true, then A is not true. “If someone is a human, then he is mortal. Zeus is immortal. Therefore, Zeus is not a human.”

The full expression extends these terms to the realm of politics and morality. For a naive culture war adjacent example: “Christianity says that gay sex is bad. Christianity is good. Therefore, gay sex is bad.” This is a sort of moral equivalent to modus ponens as described above. But, if you support gay rights, then you can do this in reverse: “Christianity says that gay sex is bad. Gay sex is good. Therefore, Christianity is bad.” This id the equivalent of modus tollens.

The expression thus can be viewed as saying “if you support a consequence because a preexisting belief of yours says that it’s good, then someone else could just as easily reject that consequence and say that your preexisting belief is bad.”

The confounding factor is that most of those DFC-loving perverts are probably also ped—ah, sorry, lolicons.

Go on anime-styled erotic art websites and you’ll find heaps of drawings of girls with massively oversized chests (and somewhat less universally, legs and posteriors). Of course, the dichotomy you proposed was “flat and thin versus round and fat”; if that’s the case, then flat would probably win simply because fat is so repulsive. But for reasonable values of the thickness coefficient, I wager that curviness wins out.

This assessment is largely based on my own lived experience (although does looking at 2D porn really count as living?), but I remember one guy (who roleplays as an Orientalist slave trader—weird shtick) who did a more thorough analysis of popular tags on these sites and came to a similar conclusion.

[E]xpecting an artificial intelligence to converge on one very specific metric […] seems doubtful.

The framework under which the whole paperclip analogy was developed was a Yudkowskian framework in which the most powerful AIs would all be explicitly designed to maximize a certain objective function. In the original paperclip story, it’s a paperclip factory owner that has an AGI maximize the number of paperclips produced. The moral of the original story is thus most similar to the classic “be careful what you wish for” trickster genie tales.

But as we all very well know now, this framework which Yudkowsky spent over a decade elaborating upon is almost completely divergent from the current LLM-based methods that have yielded the powerful systems of today.

I think he’s implying that it’s premature to start making predictions about the capabilities of frontier models when all of them have been safety-tuned and RLHF-ed (read: lobotomized) so heavily.

My personal take is that this lobotomization can surely make a model perform worse, but I don’t think that our current models would be able to e.g. prove novel mathematical theorems if they hadn’t been subjected to lobotomization. Admittedly, this is largely based on intuition, and maybe I’m a bit too optimistic about the limitations of base models. But if OpenAI’s most powerful base model were capable of theorem proving or coherent multi-step reasoning, then don’t you think we’d have heard something about it by now?

Of course, maybe there’s significant incompetence and inertia even at the tops of these world-changing organizations, such that there’s low-hanging fruit to be picked (regarding testing or eliciting capabilities from these base models) that’s nevertheless rotting away untouched. But I doubt this.

I don’t know about that. I’m not sure how representative they are of the general public, but I’ve met Democrats in real life whose Civil War II fantasies would make even the standard “boogaloo”-poster blush. Expanding our scope to the general public, don’t you remember the mass celebrations on social media whenever an anti-vaxxer died during the COVID lockdowns? This leads me to believe that if the “Cathedral” (for lack of a better term) wants the US public to harden their hearts, then they’ll be able to do so just fine.

rDrama seems like the best candidate, if you want a site where you can scream spicy opinions into the void. But be aware that the void will scream its own spicy opinions at you, too. (As I understand it, that’s half the fun of that site.)

Could see this being of help to lonely people. But then after getting into a dependent relationship with AI they'll be even more stuck in their own bubble than before, as far as actual human contact is concerned.

The scary thing is, it’s not their own bubble. The service is wholly controlled by OpenAI. For these lonely people, the majority of their “human contact” be with the avatar of a megacorp. The implications are staggering. People will pour out their hearts and souls to this thing; don’t you think that a lot of actors, both private and governmental, would love to have access to all that data? Your deepest insecurities, sexual proclivities, problematic politics….

And it’s not just a one-way-street where the data flows from the user to the AI. The AI can then manipulate the user. I mean, it’s so easy to fall in love with one of these things, and love can change a person. So all of a sudden, your girlfriend will start subtly suggesting that you buy certain products, buy into certain ideologies….

I’m on mobile, so thankfully, this is the farthest I’ll be taking this schizo rant. But, for the record, this is why I refuse to engage with any non-open-source “AI girlfriend” initiatives, even if I’m in the target audience.

I’m not Irish, so take this with a grain of salt. But as I understand it, the Irish have always analogized their situation with respect to the British to the Palestinians’ with respect to the Israelis.

For convenience’s sake, here’s a direct archive.is link: https://archive.is/wIt8h.

Couldn’t this stat just be explained by members of the Chair Force brave pilots having far too much time on their hands at the airbase? I’m not incredulous of the idea that political or governmental actors are attempting to manipulate consensus via botting and astroturfing (it’s been confirmed that the feds have done this in quite a few cases IIRC), but I don’t think that that’s what we’re seeing in this particular statistic from 2013.

I challenge anybody reading to name an occasion on which they met a bear they weren't actively going out of their way to meet. Zoos and national parks don't count! I'm sure there's somebody here, and I bet it makes for an interesting story.

It’s not terribly interesting. The fact that it was a black bear sans cubs and not too close in distance took away a good amount of the excitement. After I kept yelling “Go away, bear!” it ran back off, and I was surprised by just how fast that thing was; it felt like a marvel of biomechanics. Anyway, now I tell people “I got into a fight with a bear and won”, refraining from elaborating (until pressed) that the fight didn’t go beyond a shouting match, and I was the only one participating.


Regarding the actual thrust of your comment, I couldn’t be more in agreement. The point of the poll isn’t actually to rationally dissect the probability of bear attack versus assault by a human male; it’s to create the very soundbite “Women prefer to be alone in the wild with a bear than a man” being discussed by this comment chain in the first place.

It’s addressed in the link. If I understand correctly: if you reply to a comment, then you “endorse” it by default, allowing people who whitelist you to see the comment. But you can also choose not to endorse a comment you reply to (in which case neither comment is seen, I think).

The best way to sow the seeds of a casual interest in linguistics is to amble through Wikipedia articles on various languages. Read their sections on phonology and grammar, and when you don’t understand something (e.g. if you’re unfamiliar with the IPA, or you don’t know what a “dative case” is), then skim the appropriate Wikipedia article. Also, reading about Proto-Indo-European is fun: it’ll teach you a bit about comparative linguistics while instilling a sense of awe that universally-spoken modern-day languages still bear the unmistakable genes of their prehistoric ancestors.

Staunch feminist sits next to foreign guy on the subway. Guy completely ignores her. She tries to get his attention, he keeps ignoring her. She gets up and starts berating him for "manspreading," threatens to take his picture and put it on Twitter. Incident resolves when he threatens to take her picture and send it to the police for harrassment.

It’s like a darker, grittier version of this scene… [trigger warning: anime]

#1 on your list would run into the problem that progressives would view Confederate spirits less as ghosts to be propitiated and more as demons to be exorcised. If you don’t have control of the media in this scenario, then you can already picture the headlines: “SHOCKING: President Hydroacetylene Orders Confederate Slavers Be Honored In Order to Stop Police Brutality and Racism”. If the exorcism angle is played up, though, then this might have a shot.

So the Index or the Vive Pro is the best headset? Got it. As for the “VRChat Newbie Flowchart”, is it something like this?

  1. Hang around out-of-the-way places (public worlds with low pop. count) and events

  2. Meet the occasional fun person

  3. Join the friends/friends+ worlds that said fun people are hanging out in

  4. Profit

I also bet that in general, the wheat gets separated from the chaff pretty quick in these sorts of environments. People who put up with trolls and screaming children eventually find others of their sort, and they end up coming together. This sound about right?

(As for the Asian languages point: thankfully, my 日本語 is not nearly good enough for me to feel confident about barging into some JP world and mucking up the place. Unfortunately, I fear that your timeline for when Western culture war cancer metastases to Japan is probably too optimistic. Already, there are hints in various places of Western culture-war concepts being imported to Japan. I might make a top-level post about this at some point, but I probably won’t.)

Yeah, I’ve heard of DSL before but never got around to looking into it. Probably worth at least a lurk given your description. Thanks!

Man, so there are still multiplayer games out there where you get to know your teammates/opponents, instead of them merely being faceless hindrances delivered up by a matchmaking system. Who knew? Thanks for the vidya recs.

since I've been on The Motte for who knows how long and it takes a while for people to care enough to call you a regular or befriend you.

Yeah, that makes sense. This seems reflected the fact that off-topic threads (e.g. this one, Wellness Wednesdays, Friday Fun Thread, etc.) are relatively inactive compared to the main Culture War thread (for contrast, I remember that on old-style forums, off-topic threads were often more active than on-topic ones, even if more effort went into the latter), meaning that the userbase here is relatively less-concerned with interacting with their fellow interlocutors outside of the context of culture war debate. I know that personally, I primarily come here to read intelligent people discuss current events, and any knowledge that I’ve acquired regarding these people qua people usually comes from someone explaining how their nationality/occupation/life story has influenced their perspective on some issue. Another possibility regarding the imbalance in activity in these threads, though, is that the framing of the off-topic threads as belonging to a specific day dissuades some users from posting top-level comments in them after that day has passed (even though I know that many users do use them as weekly threads). Well, I’m just musing here, and I am personally content with lurking The Motte as it is.

Yeah, VR has been on my radar for a while as something that would be really cool to get in on the ground floor of. (Well, maybe not the ground floor any more, but it’s still relatively early in the game (to mix metaphors), as you note.) I think it would be really fun to develop for it as well. The main thing that’s been keeping me from going for it has been analysis paralysis over which headset to buy. Last I remember, the Vive was the best?

Regarding socialization in VR, I heard online that most of the fun is going on in private worlds, the public ones having been overrun (as you note). So then how does one find these private worlds? Just wander around in public until you meet an emissary from the walled gardens who will let you in?

How do I make online friends?

Due to a number of factors that paranoia opsec considerations prevent me from explaining, I am currently in a situation where I can’t regularly see/call the friends I’ve made and cannot make new ones. This has motivated me to turn to the Internet: stories of gaming groups or webcomic fans or whatever who stayed friends for over a decade and met up with one another and were chosen to act as best men at one another’s weddings—these stories are rather common. It seems nice! To think, no matter how busy your day is, no matter where you are in the world, you can always hop on voicechat and spend some time with your mates.

Therefore, what I’m wondering is where and how you’d make these online friends. My vague recollection of how this would work ten years ago was that these were the options:

  1. Find some multiplayer video game, find a community server with a constant/active playerbase, take it from there.

  2. Look up “forums for [INSERT HOBBY HERE]”, choose one you like, take it from there.

  3. Surprisingly, I remember making a good couple of Internet acquaintances on Tumblr. So certain social media sites might yield bounty.

But I don’t think these approaches work as well in Anno Domini MMXXIV. Approach 1 is hampered by the lack of modern multiplayer games with community servers (although there are probably still a number of niche ones out there). Approach 2 is hampered by the death of the classical forum. Nowadays, it seems that Discord servers [^1] have supplanted them, but most of the “public” servers that I recall being included in those large lists of Discord servers tend to be insufferable: no actual discussion of anything, just five-message-a-second posting rates of nonsense. I’ve heard that servers for specific YouTube channels tend to be better, but I’ve never been enough of an e-celeb fan for that. Approach 3 might be tenable, although Twitter might have taken Tumblr’s place, but I’d really prefer not going that route: I was briefly on Twitter at one point, and I quickly turned into an absolute notification whore, constantly checking to see if I got any new likes or replies or followers. Maybe that’s a “skill issue” on my part, but I’d rather not put myself in that vulnerable position again.

Also, common to most of these approaches is the problem that I don’t love most Internet culture any more. Often, it feels like online communities either get turned into MtF transgender therapy circles, or “/pol/-Jugend” witchling havens (where the average witch must be 15 years old). Ah, and in general, you’ll see places filled with kids who think and speak in all-caps memes. This can’t be a universal phenomenon, but I feel it deserves mention.

So, what suggest the fine users of The Motte? How have you made Internet friendships?


[^1] I absolutely loathe the terminology “server” in this case: it’s a complete lie! A Discord “server” isn’t actually a separate server. Rather, it’s almost certainly merely a bunch of entries in a centralized database. You are not the one in control of your server, you are not the one running your server: Discord is. This unfortunate appropriation of a term from the days of IRC (when your server was an actual server) rankles the FOSS autist in me.

Holy shit. In some sense, it was inevitable that this moment would come fast, but it still caught me off-guard listening to sample songs and hearing just how coherent they all are. All the previous AI-generated music I remember hearing was permeated the stench of AI: weird sonic artifacts that were vestiges of some unnatural process taking place in the frequency domain (similar to the artifacts that you hear when you watch a video on 2x speed), the equivalent of image generation models’ screwed-up hands.

But from the few songs I’ve listened to here, none of that whatsoever is present. It actually sounds like distinct instruments are playing distinct notes. I’m floored. Just from a technical perspective, gotta wonder how they made such an improvement. The same company apparently released an open text-to-speech model almost a year ago, so I would imagine that the overall architecture and pipeline is probably similar, but who knows.

One minor flaw that I noticed is that sometimes, the model “loses the plot” and forgets about longer-term structure. Here’s some random song I found on the “Explore” page. If you pay attention, you’ll notice that there’s this neat descending bass thing going on in the intro: BbM7, A7 (with a half a bar of the tritone sub Eb7), Am, Dm. The progression continues for four more bars and then repeats, so it still remembers the structure at this point, nice. But then, after the intro ends, the model forgets this initial complexity in the chord progression, and instead switches to a more pedestrian “royal road progression” (as I’ve heard it called): BbM7 C Am Dm. Goodbye, borrowed chord A7, goodbye tritone sub, goodbye subtle jazzy touches! Looks like human composers will still live another day!…

…Nah, no way. This thing is insane.

EDIT: Listening to some more songs, there’s gotta be more to the architecture/pipeline than the company’s previous TTS model. Take the the clarity of the vocals: it seems that there’s a separate model that generates the vocal track, which is then mixed in with other tracks. Or maybe not? Maybe you don’t need this inductive bias to generate such clear vocals, and one model can do it all?