Fnord

“These safeguards prevent foreign adversaries from using our most capable models in ways that pose severe safety risks. The US and its allies hold an edge in frontier chips and the highly optimized software that runs them at full potential,” the company said in a statement to WIRED. “These safeguards ensure Claude isn't used to erode that advantage—by optimizing chips developed by those adversaries, for example.”

and now Anthropic's noncitizen employees can't use Mythos! Which isn't to say that there is a direct causal link there or that Anthropic is responsible for the Trump administration's actions, but one might hope that it would make Anthropic warier of that kind of framing in the future.

Anthropic was just trying to get ahead of the curve on the issue - it is going to happen, best be at the table when it does so we can have a voice. It isn't like OpenAI opposes all regulation themselves either; they are opposing state-level efforts at regulation by backed federal-level regulations, and have gone in almost entirely on the "national security" angle. Because they have to! This is a done deal.

Well, there is one thing...

In "social media vs education" discourse, there is a tension between the macro data not really showing any large declines in student learning & minimal benefits from things like cell phone bans, and the more anecdotal/smaller survey data of people telling you the kids in the classroom have really changed. I don't actually think you should toss out the personal experiences, that is being naive, while I think "oh the test scores are all corrupted by declining standards" narrative is equally naive - so what is the story?

The squaring of this circle is that when people talk about their student learning changing, it is almost always about books. They aren't unable to understand instructions, they can do math fine (again mild declines sure), people are taking calculus and chemistry and playing piano just like they always did. What has shifted is that people's attention span for reading books is vanishing, to the point where school districts are giving up on even assigning summer reading and shifting assessments to short form reading assignments. This cycle is just extremely hard to break because books are the one thing you can't "internalize" into the classroom - you can't ask them to sit in the room and read a book during an exam - and cheating the assignment is universal and free. Kids always wanted to skim their summer reading, but skimming, Cliff Notes, "asking a friend", those were all work and sometimes just an approximation of doing the reading anyway.

Now you just google/claude the answer to any of your homework questions about a book, you don't need even crack the (pdf's) spine. And of course the entire backdrop culture has shifted to either short form text or fully audiovisual content, so you don't care much about the object-level skill anyway. The bottom has dropped out on the incentives & value for many students.

But the rub is that this hasn't impacted learning all that much. The Kids Can't Read Bleak House, yes, but they were never great at that and it wasn't that related to other skills. "The mind needs books like a sword needs a whetstone" concept was just too limited; there are a lot of different whetstones out there. The test scores are doing fine, people's skills in the workforce are holding up, none of the "breakpoints" of someone's personal trajectory are flashing red. And of course many people do still read books! We have no shortage of historians because they were always a particular type of person anyway. Books aren't going away; their market share of global attention is just shrinking.

Are there consequences for this beyond the personal, though? Yes, individuals can still learn to be nurse practitioners and marketing analysts just fine, but is the ~body politic of the nation~ or whatever suffering from an endemic shallowness in its comprehension of itself? Maybe! I don't think that idea is crazy, actually, though it is a much more complicated story with many moving parts. But it is one that sits primarily outside the education system; school was never a particularly effective vaccine against "disenlightenment", and if the internet has done anything it has neutered the power of formal institutions to be that force.

It is sad that we aren't making progress, that learning isn't really improving and is actually getting worse. We both A: thought we were gonna keep better at this, and B: there definitely was some bogus pedagogy movements in this era causing damage. But these are changes that are within the normal range. If you compare what these test say, to the tone of pieces like "The Elite College Students Who Can't Read Books", I think the claim "there is a gap here in understanding between the two sides" holds up well. And that is the gap I am talking about - if you want to just talk about the test scores in isolation from the narrative, then it can be a different conversation.

In "social media vs education" discourse, there is a tension between the macro data not really showing any large declines in student learning & minimal benefits from things like cell phone bans, and the more anecdotal/smaller survey data of people telling you the kids in the classroom have really changed. I don't actually think you should toss out the personal experiences, that is being naive, while I think "oh the test scores are all corrupted by declining standards" narrative is equally naive - so what is the story?

The squaring of this circle is that when people talk about their student learning changing, it is almost always about books. They aren't unable to understand instructions, they can do math fine (again mild declines sure), people are taking calculus and chemistry and playing piano just like they always did. What has shifted is that people's attention span for reading books is vanishing, to the point where school districts are giving up on even assigning summer reading and shifting assessments to short form reading assignments. This cycle is just extremely hard to break because books are the one thing you can't "internalize" into the classroom - you can't ask them to sit in the room and read a book during an exam - and cheating the assignment is universal and free. Kids always wanted to skim their summer reading, but skimming, Cliff Notes, "asking a friend", those were all work and sometimes just an approximation of doing the reading anyway.

Now you just google/claude the answer to any of your homework questions about a book, you don't need even crack the (pdf's) spine. And of course the entire backdrop culture has shifted to either short form text or fully audiovisual content, so you don't care much about the object-level skill anyway. The bottom has dropped out on the incentives & value for many students.

But the rub is that this hasn't impacted learning all that much. The Kids Can't Read Bleak House, yes, but they were never great at that and it wasn't that related to other skills. "The mind needs books like a sword needs a whetstone" concept was just too limited; there are a lot of different whetstones out there. The test scores are doing fine, people's skills in the workforce are holding up, none of the "breakpoints" of someone's personal trajectory are flashing red. And of course many people do still read books! We have no shortage of historians because they were always a particular type of person anyway. Books aren't going away; their market share of global attention is just shrinking.

Are there consequences for this beyond the personal, though? Yes, individuals can still learn to be nurse practitioners and marketing analysts just fine, but is the ~body politic of the nation~ or whatever suffering from an endemic shallowness in its comprehension of itself? Maybe! I don't think that idea is crazy, actually, though it is a much more complicated story with many moving parts. But it is one that sits primarily outside the education system; school was never a particularly effective vaccine against "disenlightenment", and if the internet has done anything it has neutered the power of formal institutions to be that force.

Trump-IRS settlement 'forever' bars audits into tax claims for Trump and his family

Can...can he do that?? Is that allowed?? "As part of a settlement, you can't investigate me for future crimes", that seems like something that wouldn't be on the table!

Federal law prohibits the president, vice president and other executive officers from instructing the I.R.S. to start or stop specific audits. But that broad prohibition appears to include a carve out for the attorney general.

Like, if this were a normal law, there would be no problem interpreting the president telling the AG to stop an audit as legally equivalent to the prohibited act of the President telling the IRS to stop an audit. "I didn't do it, I told my subordinate to do it or I'd fire them" is not a defense that will ever save a private company CEO.

to pull on a thread in that prev post, there's all sorts of funny nuances to names in translation/localisation, and the norms vary a lot by location as far as I can tell.

sticking to target-language English, the Asterix localisations are generally very well regarded for finding lots of English puns to replace the huge amount of French puns. in this case, the translators prioritised sticking to the spirit of the original names (dorky wordplay) over directly using the French names in English where the joke would be lost. so for example, Asterix's dog is called Idéfix in French after idée fixe, and this is rendered Dogmatix in English, which manages to slip an extra dog pun in there.

but, because of the thorough localisation, it was only when I got older that I found out the broader context of bandes desinées that I'd read as a kid like Asterix or Tintin. as a kid, they were just 'comics', just as Pokémon was 'just a cartoon' that happened to be made in Japan, not an instance of a special thing called 'anime'. but isn't that an interesting quirk? we like to refer to the distinctive creations of xyz region by loanwording their language's name for the whole medium. (following this pattern, Chinese animation is sometimes called donghua, and Korean animation aeni... unfortunately the French for 'animation' is just 'animation' so we can't play that game there.)

but yeah, a similar norm evidently was in effect for American localisations of anime, generally taken to an absurd extent (jelly donuts etc.), but even so there are many cases where it is still uncontroversial. most English-speaking fans of Pokémon will use the localised names they grew up with (so Ash and Charmander, not Satoshi and Hitokage), and the translators still continue to come up with new wordplay for each new batch of puns. if you ever download a fansub of Pokémon which uses the Japanese names, it can be a little jarring at first to relearn the names of the creatures!

however, in most cases nowadays in English translations of CJK media, the current norm is to transliterate the names into the Latin alphabet (using romaji, pinyin or one of several systems for Korean), except maybe to use the native spelling if it's obviously a foreign name in the original, or names that have a very literal meaning (e.g. ヨルハ二号B型 becomes YoRHa No. 2 Type B).

so, in an English-language context I would most often write "Haruhi Suzumiya" and not 涼宮ハルヒ. but even here, we've swapped the name order! it would be less common to write Suzumiya Haruhi in English (the Japanese style where the surname is first), and doing so would sometimes be seen as weebish. (though to be fair, I am not avoiding that if I'm talking about Haruhi in 2025!)

except of course, even just as far as name order, sometimes there are people like Yoko Taro (横尾 太郎) who are usually written lastname firstname in English, which causes no end of confusion to people who try to respectfully refer to him by his surname but call him Taro, which is his personal name. but that just seems to be the way Yoko Taro wanted to be known in English, and the gaming press just kinda went with it. even more confusingly, Yoko Taro's wife (who is known as the artist on the Taiko no Tatsujin series of rhythm games) is more often known as Yukiko Yokoo in English, the opposite order to her husband and a different romanisation of her surname.

however, these norms vary a lot by context. for example it feels more natural for me to write surname-first 'Ushiromiya George' for 右代宮 譲治 than 'George Ushiromiya', because that's the choice that the Umineko-Project translation team made, but this still hides the nuance that George's Western name is written in a really weird way using kanji in the Japanese. I can't think of a way to express that in English that wouldn't be really cumbersome.

another complication is that, even with names transliterated into the Latin alphabet, readers may not be familiar with the source language's phonology and may still end up reading the name differently. for example, when English-language commenters talk about Look Back, they may refer to the character Kyōmoto as 'Ki o mo to' and not 'kʲo mo to' (like the English 'skew'), because that's a more common way to interpret those characters in English.

and of course we can add to that that some languages use the Latin alphabet in a significantly different way than English, so if you've got some Irish characters in your story you'd need to know that Meabh and Siobhan are pronounced /meɪv/ and /ʃɪˈvɔːn/ (that's with the standard English pronunciations which are actually a bit different than the Irish ones...)

as another wrinkle, when fantasy character names are written in a different script for a different phonology (e.g. a character name in katakana), and it's unclear to the translators whether a specific Western name was intended, they just have to guess. For example, a character in Witch Hat Atelier was named アガット Agatto, and I recall reading a scanlation that reasonably assumed this is a gemstone name, 'Agate'. however, the official translation went with 'Agott' for the character's name. did they ask Kamome Shirahama about this and find out this is her preferred romanisation, or did they just make a different choice and interpret it as a fantasy name? I'm not sure, both are plausible.

in general, in names as in other areas, a translator has to be remarkably literate in all the random allusions a source work might throw out, and make a judgement call to what the original work was intending. I kind of love reading translation notes for this reason.

on top of all the above, there is also the fact that many people will have two or more names for different linguistic contexts, because their (say) Chinese name is hard to say for people who speak another language. that is a real aspect of many peoples' experience, which people likely feel all sorts of ways about that we probably want to depict in fiction!

I love scam emails so much

Only to learn that the illuminati DO in fact sacrifice rams and eat them. But without blood sacrifices.

So, here's a bit of homework: Gimme what bug you think you would be, BUT it cannot be a bee of the family Apidae (honey bee, bumble bee, carpenter bee, etc.), a Lepidopteran (moths, including butterflies) of any kind, a jumping spider (Salticidae), or a ladybug (Coccinnellidae). That should leave y'all with a good amount of more vague options, but bonus points if it's a particular species and double bonus points if it's an especially uncommon one (but remember, this is what bug you think you would be, so don't feel pressured to pick an extremely uncommon pick if you know for sure you would be, for example, a hercules beetle).

For sources of inspiration, there are plenty of places to go looking, but if you want to go with something you might be able to spot IRL then you can go to iNaturalist and check species observations around your region!