67IsTheNew69
No bio...
User ID: 4109
67IsTheNew69
2mo ago
I am very much not a lawyer, but reading the bill, it seems like it specifically makes it illegal to "sell or deliver any three-dimensional printer in the state of New York unless such printer is equipped with blocking technology". So the company is definitely fine, and I think the individual might also be fine (doesn't "deliver" imply some other person you're delivering to?).
Unless I'm misreading, the definition of "qualified product" to which much more strict restrictions are applied only includes "digital firearms manufacturing code", not unrestricted 3D printers. It still seems like a huge problem for Internet companies, and a First Amendment issue, to apply gun-control restrictions to drawings of guns on a computer.
67IsTheNew69
2mo ago
Maybe I'm pessimistic, but I suspect we've reached the top of the sigmoid for the original "train a GPT model on all the human-written text in the world" idea, and it turned out not to be possible to generate arbitrarily long/coherent/good fiction that way. The recent advancement in AI models has come from other ideas, like "multimodal" training (training on image data / audio / etc), or doing reinforcement learning on synthetic tasks.
Those other ideas can incidentally benefit fanfic -- maybe the reinforcement learning also helps the AI keep track of plots for longer, for example -- but even then I sort of suspect we're at a point where you shouldn't expect models to get much better at fanfic unless AI labs take specific action to target that use case (or they try some significantly different architecture, like text diffusion models). You definitely shouldn't expect them to be great at fanfic just because fanfic is in the training data.
67IsTheNew69
2mo ago
I'm sure that Opus/Mythos is trained on more objectively-verified examples (RLVR) than subjectively-critiqued ones (RLHF), and for the former examples, it's trained to produce shorter solutions with less reasoning (GRPO, rewarding shorter output).
Yeah, I think this is the the main culprit. Making models write their reasoning traces in a more terse style has been a big focus of more recent model releases -- it's usually expressed in the marketing as "uses fewer thinking tokens".
In the extreme case, you get Claude Mythos's "Neuralese" where sometimes the traces stop looking like comprehensible English. In more normal cases, though, reasoning traces look enough like normal prose that I suspect the "write more tersely" training leaks out and makes the model's creative writing worse. I've noticed newer models sometimes slip into a weird "reasoning-ese" writing style when writing things like software documentation and step-by-step explanations, for example.
67IsTheNew69
2mo ago
So, I'm kind of clueless about this, but are reasoning models are actually different models, as in different neural net weights?
Yes. Generally, reasoning models are trained to use special "start thinking" and "stop thinking" tokens, and to generate a specific kind of monologue in between those tokens. Similar to how RLHF biases models towards producing text that's appealing to human readers, reasoning models use techniques like RLVR to bias towards generating monologues that end in a correct solution to a problem.
Many reasoning models are trained in a way that lets you disable the reasoning by forcing them to never generate the "start thinking" token -- Claude Opus 4.8 probably uses the same weights regardless of whether you enable or disable thinking, for example -- but their weights are different from models that were never trained for reasoning in the first place.
With that being said, people used to use "chain of thought prompting" to get a similar kind of result out of regular LLMs. (I think reasoning models basically got started when AI companies saw the early success of chain-of-thought prompting and started baking it in at the training stage.)
67IsTheNew69
2mo ago
The sample size of women was N = 13, so even if 9% of them also fall in the "has never been in a relationship, and fills out a survey about infidelity just to report that fact" category, it seems decently likely that none of them would have happened to do that.
67IsTheNew69
2mo ago
Yeah, the "thinking" process itself also counts as output tokens. When you use a reasoning model, it's basically writing a long monologue about how it's going to solve your problem and then immediately throwing it away at the end. (Different providers have different policies about whether you're allowed to see this monologue, but it often significantly exceeds the length of the actual code or whatever that the AI is writing.)
67IsTheNew69
3mo ago
2032: [...] Places like Japan are receiving modern websites for the first time.
I know everyone's already piling in to tell you "this is already happening", but... Unless you expect "modern" websites to look dramatically different in 6 years, I'd say websites in Japan are already pretty modern? There are definitely a few Craigslist-style companies out there with outdated design, and there are some differences due to local design taste and language (web fonts are much less practical to use for Japanese text, for example). But overall web design standards haven't changed much ever since the flat design trend took hold a decade ago, and most Japanese organizations have caught up.
Compare www.city.osaka.lg.jp to www.chicago.gov, for example.
67IsTheNew69
3mo ago
I enjoy that the DNC can't even agree on what the outcome of the 2024 election was. From page 17:
Democrats netted two seats in the House, flipping ten seats from the Republicans while losing eight.
[Red box: Data appears to be inaccurate and contradicts public reporting.]
The Wikipedia article "2024 United States House of Representatives elections" is also confused about these numbers in a few places (e.g. the values in the "By state" results table don't actually add up to the stated totals), I think maybe because it's not sure how to handle the fact that George Santos had been replaced by a Democrat in a special election earlier that year, but even it doesn't actually say that Democrats flipped 10 seats. (The generally accepted narrative seems to be that they only flipped 9 seats, for a net gain of 1.)
67IsTheNew69
3mo ago
I think a more straightforward way to notice that this scenario detaches P(heads|you just woke up) from the optimal betting strategy is to compare it to the following scenario:
Some researchers flip a coin without showing you the result. On Monday, they interview you about the coin and ask you to make bets about its status. Then, on Tuesday, if the result was tails, the researchers play the videotape of your interview from Monday and perform all your bets a second time on your behalf.
Here, your belief that the coin landed on tails should clearly be 0.5 even given the condition that you're currently being interviewed. But if you make any bets, you need to keep in mind that they'll be executed twice in the tails condition. The optimal strategy is the same as in the original Sleeping Beauty problem, since that problem supposes that you were going to do the same thing on both days anyway. (That strategy is not as straightforward as "assign probability 0.6667 to tails" if you can bet things like "all the money currently in my checking account" rather than just fixed dollar amounts.)
So within the problem, the concept of "credence" is not as broadly applicable as it normally is; the conditional probability is different from the optimal betting odds (and those odds themselves differ based on details of the bet). You can either stick to the conditional-probability definition, say that the odds are 0.5 (0.6667 in the original problem), and not use that value for any practical purpose. Or you can say "I think there is a 50% chance that the coin is tails, and if that is the case any actions I take will happen twice", which is a more useful fact to know when strategizing.
67IsTheNew69
4mo ago
I'd say my health is decent. Like many people, I should maybe exercise a little more and eat a few more fruits and vegetables. As it is I take a multivitamin and walk thousands of steps a day, and have no ongoing medical issues or concerns.
I wouldn't say I get tired easily or feel exhausted all the time. I spend a lot of time and energy on leisure activities, but I do find long social engagements (multi-hour dinner parties etc) pretty uniquely exhausting. My guess is that I'm not used to those kinds of events so need to spend a lot of mental energy analyzing what's going on and figuring out what I'm supposed to do; that and I tend to act a lot more energetic and outgoing than usual when I'm around other people.
67IsTheNew69
4mo ago
The blue button is the optimal choice here. Either it gets 50% of the vote and nothing happens, or it fails and I get to go to Heaven, because risking my own life to save the lives of all the people who don't understand Monty Hall problems is the ultimate act of self-sacrifice.
67IsTheNew69
4mo ago
67IsTheNew69
4mo ago
I really appreciate the insight here. Personally I think that I am genuinely good with how things are, but people sometimes express mild concern about my lifestyle, in a "you really should get that mole checked out" kind of way.
To your point about culture, I do think part of it might be that society sees Internet activity as not really being "real" social interaction, even though hanging out on a forum is probably filling the same psychological need that hanging out at a pub would've done.
67IsTheNew69
4mo ago
All of my grandparents had pretty severe dementia, so I think I'm cooked no matter what I do. I'm just hoping my country gets MAID / right-to-die sorted out in time.
67IsTheNew69
4mo ago
Is "schizoid personality disorder" a real thing that the medical industry can help me with? I did some psychometric testing a while ago and was told I might have that -- I suspect primarily based on some questionnaire responses where I indicated that I was content with not really having any friends or engaging in any social activities. (I still mostly feel that way, but it's a bit of a panic whenever I need to scrounge up an emergency contact or a personal reference for something, and I sometimes wonder if it's like one of those Mr. Beast videos where a guy sees color for the first time and can finally appreciate the full richness and beauty of human existence.)
67IsTheNew69
4mo ago
as far as I know, Google was never a major provider of the game in the first place
The game wasn't available on Android (or iOS) at all until four months ago. I remember it doing pretty well when it came out -- I think it hit a million downloads -- but not nearly as well as the original Steam/Itch release back in 2017.
67IsTheNew69
4mo ago
Unfortunately the mobile and console versions of DDLC compromise pretty heavily on the
I almost think it's a net good to take down the mobile version of DDLC, if it directs more people towards the superior original PC version. (Of course, that's definitely not why Google took it down, and Google's actual reasons are probably bad enough to merit opposing this move regardless of its immediate effects.)
67IsTheNew69
5mo ago
I've heard that videos under ~5 minutes become shorts which don't get monetized
There was some panic over this in the early days of Shorts when YouTube was retroactively, non-consensually converting existing videos into Shorts, but these days, YouTube is pretty clear that only vertical/square videos less than 3 minutes in length become Shorts. So as long as your video is shaped like a normal video, it can be as short as you want without getting the Shorts treatment.
You can technically monetize Shorts now too, but I've heard it's not as lucrative as traditional videos.
67IsTheNew69
5mo ago
I wanna say it was around 2020-ish when YouTube changed their algorithm to put a lot less weight on channels' upload record / consistency and a lot more weight on individual videos' watch-time. (Hard to know for sure because all information about the YouTube algorithm is buried under a mountain of "how do I get rich on YouTube" slop articles.) Previously the dominant strategy was to release a lot of short, punchy videos, ideally on a daily release schedule; Captain Disillusion griped about that in his 2018 parody video. But by 2021, creators like Quinton Reviews were seeing success with five-hour monstrosities that were actually just a bunch of shorter videos combined into one upload for algorithm purposes.
"Short daily video" YouTubers and "long infrequent effortpost" YouTubers have both existed on the platform for a long time (and still exist), but now that long videos are a popular/successful format, I see a lot more low-effort attempts at making them. It doesn't help that YouTube Shorts (and TikTok) provide a better path to success for people interested in making short videos.
67IsTheNew69
5mo ago
In court you must pay deference to the judge (who, being very high class indeed, is often very disrespectful- “I just can’t with that history.” I wanted to jump her too.)
This is tangential to your main point, but I found a video clip on Reddit with more context, and it seems like the judge was responding to the defendant's lawyer saying "I think the court can trust that he will perform well on probation" (at about 6:30 in the video). So I think she was trying to straightforwardly say "I just can't trust [that he will perform well on probation] given his history", and wasn't, like, invoking Tumblr slang to disparage his criminal record.
67IsTheNew69
6mo ago
I think if the government gave out a few standardized food items to every benefits recipient, that would actually heavily encourage trading. They'd basically be minting a currency, except instead of coins that all contain the same amount of silver, it's bags that all contain the same amount of Kraft cheese product or whatever. There'd be big opportunities to take those items and smuggle them back into the regular supply chain en masse for cash, or sell them to people who want a reliable source of cheap Kraft cheese product.
Ideally, if you want to prevent trading, you want to give people stuff that seems valuable to those people, but worthless to everyone else. So they're incentivized to consume the items themselves rather than try to sell them. The mushroom "superfoods" described in OP actually seem like a excellent example of this principle in action.
67IsTheNew69
6mo ago
Another factor is that benefits are only sent out on certain days of the month. (Looks like in California it's the 1st through the 10th depending on your account number.) So you're more likely to see people using EBT on those days, or the weekends after those days, especially if people are doing the responsible thing and planning out a single big shopping trip to get all the stuff they need for the month.
67IsTheNew69
6mo ago
I followed the citations, and the "sexual aggression scale" the researchers used in their questionnaire involves asking questions with five possible answers ranging from "not at all likely" to "very likely". However, they got that 30% statistic by re-coding the answers as either "yes" or "no".
So this seems like the classic social-science trick where you inflate the number of "yes" responses to a question by providing one answer choice that means "no" and four answer choices that all mean "yes". And because they asked about both "rape" and "forcing a female to do something sexual she didn't want to", you get the bias where people want to answer that one is less likely than the other.
(The "Materials and Methods" section of the paper makes clear that the "something sexual" wording was what they actually asked on the questionnaire. The researchers seem to have paraphrased that as "force a woman to sexual intercourse" in their results, which also seems kind of misleading.)
67IsTheNew69
6mo ago
The dates given in the Wikipedia article for the incident suggest that the victim was about 59 years old at the start of the incident, and it continued until she was about 68.
Where I'm from, I think that'd be called "old age". And to @ActuallyATleilaxuGhola's point above, it seems like a stage of life where it'd be pretty normal to have a lot of unexplained aches and pains.
Gender ideology is a big-tent movement, and many of its members can be pretty slippery about assigning concrete meanings and labels to things, so it's going to be difficult to make a specific set of claims that "those who adhere to gender ideology" believe.
With that being said, I think the main way your explanation differs from the usual story is that "sex" is not considered a category that exists. Each person has a "gender", which is an innate property that's not externally observable. Then they have several externally observable traits, things like what kinds of clothes they wear or what kind of genitals they have. Society generally has associations with which traits correspond to which genders (e.g. sundresses are typically worn by women), but it's not necessary for a person's traits to align with their gender (e.g. putting on pants does not cause you to cease to be a woman).
Taken to the extreme, it's not necessary for any of a person's externally observable traits to align with their gender. While many people feel an innate, personal sense of despair when their externally observable traits do not match their gender identity (according to their/society's general associations between traits and genders) and they should be given the resources and medical care to bring their traits into alignment with their gender identity if they wish, that's a purely personal decision on their part. It's important for everyone's gender identity to be respected regardless of the traits they have.
"Externally observable traits" here is a term I made up. The term used most often is "gender expression", which refers to the traits that other people can actually observe in practice, which is what matters most of the time. But things like "the feelings caused by the hormones in your body" are also usually considered to be something external to your gender identity, such that you might want to bring them into alignment but are not required to. (Things like the Y chromosome are not considered relevant here because they aren't easily observable. It's nice if you happen to get a bunch of traits that all align with your gender identity, but the "how" of that doesn't matter much.)
More options
Context Copy link