@VoxelVexillologist's banner p

VoxelVexillologist

Multidimensional Radical Centrist

1 follower   follows 0 users  
joined 2022 September 04 18:24:54 UTC

				

User ID: 64

VoxelVexillologist

Multidimensional Radical Centrist

1 follower   follows 0 users   joined 2022 September 04 18:24:54 UTC

					

No bio...


					

User ID: 64

those corporate HR policies he’s setting aren’t getting implemented in places like Texas and that middle management is simply lying because they don’t want to look bad.

I think this is really a function of workplace culture and class rather than geography. Most of your college-educated, white-collar workforce is cognizant enough to recognize that "no homo" in the office probably won't fly. But I doubt that the truck drivers, technicians, assembly line workers, and even janitorial staff are really watched by the liberal panopticon so closely. For many working-class gigs like restaurants, the working language (Spanish, most frequently) isn't necessarily understood uniformly by the educated left anyway.

But I could just as easily write this about dropping "fuck" in every other sentence, which is also a distinctive class marker.

Keep in mind: there is absolutely no sex/gender distinction in our local language.

There is barely one in English either, to be honest. It seems to have been shoehorned retroactively because the sex descriptors are adjectives -- "female" as a noun is, er, quite objectifying as used, and I can see why it upsets some feminists -- and the gender descriptors are nouns: "woman [career]" is awkward too.

It is no exaggeration to say that most of STEM innovation in US academia is now being carried out by foreign-born people.

This may be true overall, but I think it's somewhat exaggerated because what we discuss as "STEM innovation" is colored by high-profile tech companies that generally broadcast their research far and wide. If you're talking OpenAI and Google, sure. But there are significant classes of employers that are restricted to either citizens or at least permanent residents (green cards). SpaceX isn't hiring Chinese nationals. The largest employer of mathematicians in the US only hires citizens and generally holds its research very close to the chest, as do the national labs.

I think an unbiased sample of "STEM innovations" is harder to measure in real-time than it sounds. But also that concern about US research output is not necessarily misplaced.

I imagine these kinds of things still happen in the US but aren't federally funded. And US university people know anything about this?

The typical workaround is that you can host a "women in [field]" event, but you can't restrict who actually attends. To some extent everyone knows what's expected, but I do recall my local Society of Women Engineers chapter was pretty explicit about recruiting all comers, so it's not all a wink and a nudge.

At least in rural areas it's pretty frequently legal to engage in target shooting on private property: as far as I know in Texas it's legal to discharge firearms on your own property outside of city limits provided you're at least 300 feet from neighboring occupied buildings. Within city limits it's generally a local law issue. Rifle and shotgun shots (presumably mostly for sport or hunting) are not an uncommon sound if you start wandering backroads.

Which would make this another example of "just enforce the law you losers" cases.

As far as I can tell, the suspect in question wasn't in the US legally, and thus couldn't have legally acquired the firearm in question.

Because Title IX has been interpreted to require universities to referee adolescent relationships! Title IX created the problem (via campus administration), and Title IX "fixed" the problem (via the judiciary)

Worse: the requirement was codified by a now-rescinded Dear Colleague letter from the Obama administration requiring schools adopt these policies.

It was rescinded by Betsy DeVos in 2017.

IIRC the controversial "refereeing of adolescent relationships" portion was driven by a requirement to review sexual harassment and sexual assault allegations with a preponderance of evidence (civil) standard, which put school administrators in a position of having to establish parallel judicial systems because an act that didn't meet "beyond a reasonable doubt" in criminal court can absolutely meet a preponderance of evidence.

IMO punishing students with things like expulsion needs higher than a 51% standard of evidence.

I am pretty firmly of the opinion that any post-scarcity society we can build looks less like UBI socialism and more like massive deflation in the pricing of essentials driven by automation and capital investments.

There are plenty of commodities that used to be expensive and are now basically free. Salt and pepper used to be valuable commodities, but now are tossed carelessly in paper packets into food packaging and largely discarded. Within the last century we've driven food prices down to where basically nobody in the West dies of starvation because they can't afford food: there are plenty of charities that together manage to make sure everyone is fed, although I'll concede the nutrition is often lacking.

But I also think the hedonic treadmill is a powerful thing and we can relatively easily convince ourselves that things haven't objectively changed: to me post-scarcity seems doomed to always look like the distant future, but is actually an asymptote we can steadily approach.

The equivalents in 1975 were saying that the Cold War would inevitably end in nuclear annihilation. This was a terminally unhelpful position

IMO this is a fair comparison, although the Cold War MAD scenarios were explicitly designed to cause annihilation. The Bulletin of the Atomic Scientists, probably the premier Cold War doomerism group, is practically a laughing stock these days because they kept shouting impending doom even during the relatively peaceful era of 1998-2014, finding reasons (often climate change, which is IMO not likely apocalyptic and is outside their nominal purview) to move the clock towards doom. Do you think they honestly believe that we're closer to doomsday than at any point since 1947? We supposedly met that mark again in 2018 and then moved closer in 2020 and again in 2023.

There are all sorts of self-serving incentives for groups concerned with the apocalypse to exaggerate their concerns: it certainly keeps them in the news and relevant and drives fundraising to pay their salaries. But it also leads to dishonest metrics and eventually becomes hard to take seriously. Honestly, the continued failure of AI doomerists to describe reasonable concerns and acknowledge the actual probabilities at play has made me stop taking them seriously as of late: the fundamental argument is basically Pascal's wager, which is already heavily tinged with the idea of unverifiable religious belief, so I think actually selling it requires a specific analysis of the potential concerns rather than broad strokes analysis. Otherwise we might as well allow religious radicals to demand swordpoint conversions under the guise of preventing God The One Who Operates The Simulator from turning the universe off.

As a counterexample, I think the scientists arguing for funding for near-Earth asteroid surveys and funding asteroid impactor experiments are quite reasonable in their proclamations of concern for existential risk to the species: there's a foreseeable risk, but we can look for specific possible collisions and perform small-scale experiments on actually doing something. The folks working on preventing pandemics aren't quite as well positioned but have at least described a reasonable set of concerns to look into: why can't the AI-risk folks do this?

I do think there are many Republicans who would not like to see mass deportations, but I think 80%+ of Republicans would like to see a secure border.

The (admittedly moderate) Republicans I know generally would agree with this: the sentiment is generally in favor of enforcing the laws on the books, with the caveat that changing the laws to reflect reality is acceptable -- for example, by massively expanding work visas to provide a legal basis for those who currently cross as illegal immigrants to find work, which would require facing the thorny issues of what conditions would be applied for those visas. But there's a concern that the generally uncontrolled state of the border allows all sorts to cross: not just cartels, and not even just Central/South Americans. Take a look at the 2020 statistics (last page): there are literally hundreds of citizens of places as far away as China, Ghana, and Bangladesh, and Romania apprehended by Border Patrol crossing the Southern border (many of these may present valid asylum claims under existing law, but that's not clear from the data and is honestly a pretty poor route to encourage even if true). Most developed nations aren't opposed to or incapable of tracking names and dates of those that enter the country.

What are the specific statements she made you think constitute defamation?

In a '98 NBC interview she called the sexual assault allegations against her husband a "vast right-wing conspiracy".

During her NBC interview, when Lauer said, "So when people say there's a lot of smoke here, your message is, where there's smoke..." Clinton interrupted.

"There isn't any fire," she said. After the allegations and motivations are dissected and the truth emerges, she predicted, "some folks are going to have a lot to answer for."

This was in an NBC interview, not a courthouse (like Trump's comments), and appears to directly suggest that all of the women are liars, which appears at least modestly similar to Trump's supposedly-defamatory remarks.

Ultimately, I think I have to conclude that denying allegations should probably enjoy specific privilege from defamation concerns.

If we can reasonably expect every police officer to carry a body camera on duty in 2023, we could absolutely put a camera on every subway car. A little searching suggests that every car in the Tokyo subway has had cameras since at least 2020. Like the body cameras, at some point it becomes an obvious choice to not provide surveillance, although it's not completely obvious to me that 2023 is that point.

The NYPD claims that its rollout of body cameras to "all Police Officers, Detectives, Sergeants and Lieutenants regularly assigned to perform patrol duties throughout the city" was completed in August 2019, for a total of around 24,000 cameras. There are approximately 6400 subway cars in NYC.

And yet when asked to apply the same logic to that same sort of men in America, an impenetrable mental block descends and makes it impossible for even the same commentators to reach the same conclusion.

Part of the issue here is that America defines nationality not along ethnic lines, but by citizenship (so does France, although this appears less in Anglophone discussions). Swede refers both to citizenship and generally to a specific subset of Scandinavian heritage. I'd say these map generally to language groups, but that's not fully the case (there are Swedish-speaking Finns): identity is a complex and often locally-defined concept, and Americans have generally embraced the "Melting Pot" outlook that "American" is not an ethnic group.

The problem is that "Those men are not [Americans]. They will never be [Americans]" is not obviously a true statement: one can absolutely (with appropriate vetting, which is hopefully not open to gang members shooting each other in malls) choose to become an American. One can become a citizen of Sweden, but our linguistic blurring of Swede referring to both heritage and citizenship makes "becoming a Swede" less obviously correct.

This is the equivalent of a Japanese Banzai charge straight into dug-in machine gun emplacements and sighted artillery.

I've long thought that if the Catholics really wanted to win a battle in the Culture War, they should start repeating "anti-Catholic animus" (or perhaps some catchier -phobia or -ism term I'm not going to consider) in the same way that "racism" and "antisemitism" get thrown around. The historical citations aren't really unjustified: the KKK was founded as, among other things, anti-Catholic. All of the historical bias against Italians and Irish immigrants is at least somewhat rooted in anti-Catholic bias, as is some of the bias against Central and South American immigration. The Nazis persecuted Catholics. And they continue to be victims of hate crimes in the US.

On one hand, repetition legitimizes and a constant drone of "we're persecuted" is functionally how various groups on the left have achieved their existing hierarchy -- this seems to bear more relation to the quantity and quality of complaints than to any particular metrics of measurable oppression. On the other, I respect that Catholics absolutely could claim (some degree of) martyrdom in the Year of Our Lord 2023 but choose not to because silent stoicism better aligns with their principles.

I know this is a hobby horse, but once AI is trained on gait recognition and body language of labelled examples of millions of hours of countless criminals’ movements recorded by CCTV, tiny little telltale patterns might well allow for effective pre-crime in the case of almost all premeditated criminal activity. People show their nerves, everyone has a tell, etc.

I have trouble believing there's enough information content present in CCTV streams to uniquely identify individuals confidently. I see how it maybe could work, but it's not something I'd focus on directly. Are human gaits really that different as to be identifiable from distant security cameras? Are they even consistent for a single person day-to-day?

The longer I think about it, I've also started thinking that AI likely scales sub-linearly (logarithmic?) with the size of the training dataset. "But the AI can viably consider a larger dataset than human experts" may be true, but may not generate hugely better results.

Interesting! Any good papers or summary articles you'd recommend?

One of my heuristics for good persuasive writing involves the number of citations, or at least clear distinct factual references that could be verified, as the clerk is doing for the rest of us here. Broad, general arguments are easy to write, but in my opinion shouldn't be weighted as heavily.

The amusing part here is that I have been doing this for years to weed out political hacks, long predating GPT.

It's not bizarre at all if you remember that ChatGPT has no inner qualia. It does not have any sort of sentience or real thought. It writes what it writes in an attempt to predict what you would like to read.

I don't think I disagree here, but I don't have a good grasp of what would be necessary to demonstrate qualia. What is it? What is missing? It's something, but I can't quite define it.

If you asked me a decade ago I'd have called out the Turing Test. In hindsight, that isn't as binary as we might have hoped. In the words of a park ranger describing the development of bear-proof trash cans, "there is a substantial overlap between the smartest bears and the dumbest humans." It seems GPT has reached the point where, in some contexts, in limited durations, it can seem to pass the test.

If so, I want to ask the DoD why they're giving people clearance to say untrue things.

For some categories of folks, I wouldn't be surprised if pre-publication review is mandatory. The DOD probably has to at least proofread any book by certain folks, even if they decide to write historical fiction. A memoir or anything close to their specialty could actually inadvertently disclose something classified. It's quite possible the folks here charged with "giving clearance" only care about a very narrow set of facts (names, places, dates) appearing.

Although I suppose claiming any degree of official statement or backing might be its own concern. But it's unclear to me this is actually claimed here: "the censors didn't censor my ramblings" isn't alone an endorsement.

IMO the fact that the universe exists is a fairly compelling argument for theism generally. Any science undergraduate will understand that zero everywhere is a very satisfactory solution to all of the relevant field equations, but the fact that anything is here at all implies a far more complicated arrangement. I personally find it more compelling than life, even intelligent life, existing within the universe.

But that's just my opinion and it's hardly conclusive ontological proof.

The issue around classification is effectively whether Trump could have by his power as President deemed any of the documents he took to not be information relating to the national defense, and also whether or not his claims to have done so are in fact true, or just something he made up after the fact of him leaving office.

Part of the problem with this is legal case is that with a few (largely nuclear-related) exceptions, all classification guidance exists in the form of Executive Orders. The current guidance is EO 13526 from 2010, but that revoked and replaced a whole list of orders from previous administrations dating back to Harry Truman. So if the question is "could Trump have declassified this?", he could have declassified (almost) everything by mere a executive order revoking 13526 without replacement. In addition, the EO 13526 explicitly designates the President (and Vice President) as a "classification authority" able to determine classification.

But what constitutes an executive order? In general, the separate powers of the US federal government are given broad leeway to determine their own rules and procedures (see Noel Canning, which found that the Senate is in recess only when it declares itself as such). I can't see any reasonable court deciding that failing to write on official White House stationary invalidates an executive order. There might be an argument that the President wasn't faithfully executing the laws as passed by Congress, but the Legislature has its own means (impeachment) for enforcing that.

If the contention is that the illegal acts happened after he was president, that's a potential case, but I think it still faces a fairly high bar to show that keeping the documents wasn't justified by actions taken as president: that would require a court to take significant leeway in interpreting how the executive ran its operations. A precedent of "just because a President [claims to have] issued verbal instructions to do things that are lawful except for violating prior executive orders doesn't prevent your prosecution for violating those prior orders" would be terrible.

Does an elected President (in particular, one with no prior service) even have to sign SF 312? That NDA is the vehicle through which most criminal charges for mishandling classified information flow, and without it it's unclear that any charges could stick to a non-signatory. That's why the powers that be can't charge the journalists at The Washington Post who published the Snowden leaks.

Now, the fact that classification is almost entirely due to Executive fiat is, I would agree, a terrible arrangement, and it would make quite a bit of sense to codify (much of) the existing ruleset through an act of Congress. But, in its great wisdom, Congress hasn't decided that doing so is worth its effort. Ultimately, I'm not a fan of Trump, but this really seems like a politicized effort to bring historically unprecedented charges.

Anything and anyone around big secrets tend to become secret as well. And eventually things must either end in everything being secret or nothing being secret.

It sometimes works this way, but there are certainly cases where specific facts can remain classified in isolation on systems that are acknowledged. The DoD likes to claim top speeds "in excess of" a public value, while the actual numbers are much more closely guarded, and they presumably avoid exceeding the claimed value outside of controlled conditions. Your missile silos in North Dakota are part of Minot Air Force Base, which even has a public web site, but I doubt the silo coordinates are on it. I can only assume that completely secret programs are particularly expensive and avoided when possible.

Although I'll agree that your concern is reasonable and there are plenty of folks worried about overclassification, even within the government.

This is why secrecy and classification is infectious. Right now if you work at all with certain government agencies, like any branch of the military, the CIA, FBI, etc then you will almost certainly have to get a top secret clearance.

The need to hold the clearance doesn't imply constantly working with such material. I've talked to folks with TS clearances who have offices with windows: sometimes it's a small fraction of the job in a separate space. But having a standing clearance makes it easier to have certain discussions on short notice (and, as someone once told me, "prevents paperwork when people make mistakes").

Separate and apart from that issue, the Espionage Act does not reference "classified" information. It references information "relating to the national defense," as it predates the modern system of classification by several decades. It's not clear that information being classified or declassified changes whether it is objectively information "relating to the national defense."

This is true, but I'm unaware of any (successful) instance of bringing charges against someone without a prior position of trust (see the SF 312 question). Nobody charged the editors of The Washington Post for possessing the Snowden documents, which were presumably "relating to the national defense," nor can I think of anyone other than the leakers charged for possessing leaked documents in the cases of Manning, Winner, and Teixeira. The Assange case is still outstanding and raises a lot of relevant questions.

Of course, I suppose that would be moot if someone produced a SF 312 signed by Trump himself.

But you have a reasonable argument that could be made, I just think this treads on lots of areas of novel law that will (given reasonable legal strategy, which is hardly a given for the parties in question) presumably raise lots of opportunities for challenges and appeals.

One of the most relevant parts of that game theory, though, is how to re-establish trust and fairness given a history of defection. Given a historically bipartisan-ly corrupt system, how do you begin enforcing the rules without appearing to play favorites?

Honestly it's a hard problem of soft skills: if there were an easy answer, any number of longstanding grudges (Israel/Palestine, etc) could be settled. There are a few successful examples: Northern Ireland seems pretty peaceful these days.

For the record, I'd much prefer a non-corrupt system, but I think a partisanly corrupt system is probably even worse.