VoxelVexillologist
Multidimensional Radical Centrist
No bio...
User ID: 64
VoxelVexillologist
Multidimensional Radical Centrist
27d ago
I'd personally endorse (B), although I could see an argument for (C) in some cases. I think the novel claim of "POSIWID" is that under the analysis of (B), some systems are revealed to have a "purpose" that contraindicates it's mission statement under (A). The idea that systems are complex and efforts to push a given indicator in direction Y might actually move the needle the other direction should be taken seriously: eliminating phonics instruction "to improve literacy" has quite possibly worsened outcomes. Just because a system exists "to fix Z" doesn't mean it's actually helping.
The steelman for (C) is that looking at anything other than outcomes risks endorsing systems that are actively counterproductive but happen to "sound nice" and have "good vibes" on paper: "The purpose of NEPA is to heavily curtail new construction," or "The purpose of the NRC is to prevent new nuclear reactors from being built" (literally the NRC had never approved the construction of a new nuclear plant from its 1975 inception until Vogtle Unit 3 began construction in 2009).
VoxelVexillologist
Multidimensional Radical Centrist
27d ago
I would have loved to see that viewpoint diversity report on an Abstract Algebra class. It should at least require the elimination of radical ideals.
I've long thought it would be amusing to portray (as farce) an Inquisition within the Math Department to root out heretics that accept the Axiom of Choice. Probably as a musical.
VoxelVexillologist
Multidimensional Radical Centrist
27d ago
It feels very germane to this issue that "immigration judge" is not a judicial-branch role here: it's someone the AG has delegated their Congress-given power to decide the merits of these immigration cases. Honestly, it's IMO a terrible name for the role because it's so misleading. It seems like the relevant questions in this case are less "the administration (and perhaps the AG specifically) disobeyed a judicial order" and more "the administration (and perhaps the AG specifically) didn't follow its own documented procedures": the former is a very legitimate separation of powers concern, while the latter is, still legitimately, more a question of proper procedures -- capriciousness, notice-and-comment requirements, and documentation.
Naively, "if the Attorney General decides that the alien’s life or freedom would be threatened" in (b)(3)(A) seems completely satisfied by the AG testifying "I have decided" regardless of what her delegates have documented in the past. That said, I am absolutely willing to accept procedural questions about what "decided" needs to mean for good governance, weighted against the idea that past executive branches shouldn't be able to forever bind future executive actions.
VoxelVexillologist
Multidimensional Radical Centrist
28d ago
Is that process defined by Congress, or by executive policymaking? Admittedly it could be some combination of both (policy specifically to implement details called out in law).
VoxelVexillologist
Multidimensional Radical Centrist
28d ago
IIRC the US opened a completely new shell production line about a year ago that didn't exist at all before 2022. I think the EU has also ramped up production.
how many contact methods does DHS use?
I suspect DHS could have decided not to grant parole without three valid, tested methods of contact, but that was either deemed too much effort or not in line with the previous administration's political objectives.
IMO we should design these systems to avoid even accidental incentives for misuse: establishing "if you provide incorrect information, we can't reach you and therefore you can't be expected to comply with otherwise lawful orders to your perceived detriment" is not a sustainable precedent. There is a reason you can't just send your W-2s to the address of the municipal landfill and tell the IRS you didn't know you owed income taxes.
VoxelVexillologist
Multidimensional Radical Centrist
28d ago
Does a pretrained, static LLM really measure up to your "actually von Neumann" model? Real humans are capable of on-line learning, and I haven't seen that done practically for LLM-type systems. Without that, you're stuck with whatever novel information you keep in your context window, which is finite. It seems like something a real human could take advantage of against today's models.
VoxelVexillologist
Multidimensional Radical Centrist
28d ago
Kolmogorov complexity is, IMO, a "cute" definition, but it's not constructive like the Shannon limit, and is a bit fuzzy on the subject of existing domain knowledge. For lossy compression, there is a function of how much loss is reasonable, and it's possible to expect numerically great performance compressing, say, a Hallmark movie because all Hallmark movies are pretty similar, and with enough domain knowledge you can cobble together a "passable" reconstruction with a two sentence plot summary. You can highly compress a given Shakespeare play if your decompression algorithm has the entire text of the Bard to pull from: "Hamlet," is enough!
VoxelVexillologist
Multidimensional Radical Centrist
29d ago
I think the problem is that we still lack a fundamental theory about what intelligence is, and quantifiable ways to measure it and apply theoretical bounds. Personally, I have a few suspicions:
- "Human intelligence" will end up being poorly quantified by a single "IQ" value, even if such a model probably works as a simplest-possible linear fit. Modern "AI" does well on a couple new axes, but still is missing some parts of the puzzle. And I'm not quite sure what those are, either.
- Existing training techniques are tremendously inefficient: while they're fundamentally different, humans can be trained with less than 20 person-years of effort and less than "the entire available corpus of English literature." I mean, read the classics, man, but I doubt reading all of Gibbon is truly necessary for the average doctor or physicist, or that most of them have today.
- There are theoretical bounds to "intelligence": if the right model is, loosely, "next token predictor" (and of that I'm not very certain), I expect that naively increasing window size helps substantially up to a point, and at some point your inputs become "the state of butterfly wings in China" and are substantially less useful. How well can (generally) "the next token" be predicted from a given quantity (quality?) of data? Clearly five words won't beget the entirety of human knowledge, but neither am I convinced that even the best models are very bright as a function of how well read they are, even if they have read all of Gibbon.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
You could organize a conference for US residents, but that might exclude two thirds of the researchers.
I think it's worth noting that these conferences exist: for research close enough to ITAR and defense stuff, there really are "US persons only" academic conferences.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
To be a bit glib, we could establish a third branch of the federal government with a nebulously-defined capability to charge the other two branches (and their agents) and remove them from power as necessary when a sufficiently-large group of geographically-distributed representatives find their actions to be sufficiently out-of-line with the general consent of the governed.
As to how you'd get Congress to stop napping and actually become accountable for things again, that seems much harder. You'd think "do a bunch of questionable stuff to draw their ire" might work, but I've seen quite a bit of questionable stuff in my lifetime, and it hasn't yet.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
I think the norms of this community strongly opposed copy-pasting the same response into multiple replies in the thread. Even if it is an apropos, relevant quote.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
I'd question the precedent the other way too, for the courts to dictate specific international relations outcomes seems a pretty slippery slope as well. If the court orders "return them by any means necessary" then I suppose we're all in for judicially-mandated ground invasions. If only Cheney had been aware of this One Weird Trick.
If there was the executive will to actually get this man back it's plainly obvious a phone call could be made to El Salvador to put this man on a flight back tomorrow.
I think you're probably right at some level on the circumstances in this case, but the man is a Salvadoran citizen and international law generally provides a number of protections when one country requests something resembling extradition. If Australia asked the US to return an American who overstayed their visa there and was now imprisoned in the US (for whatever reason), I'm not sure we'd just hand them over either. "Why didn't the Judiciary just compel the Executive to make foreign nations release their own political prisoners?" reads very much as "one weird trick" that you wouldn't expect to work.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
Without endorsing the idea, I have to wonder if the US is calling up Rwanda to see if they could get an agreement like the (unused) one they had with the previous UK government. "As legally required, this is not El Salvador."
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
Roberts is rather pragmatic, and I'm somewhat inclined to believe that at least part of the motivation here is that the judiciary shouldn't be issuing orders that won't actually be followed. Better to issue a stay then find that his orders won't (or can't, given international relations) be followed, and fight a better battle some other day.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
If you believe the FBI, there was that plot to kidnap governor Whitmer.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
With a level of charity that I would myself question the reason behind: "reasonable requests" get indefinitely referred to committees and yield "we'll see" (which means probably no in practice) responses. Witness that American presidents have been encouraging increased EU defense spending for decades, but it took the (IMO very questionable) announcements of the last few months to seemingly cause a change in mindset and priorities. It takes a (perceived) crisis to make people reconsider seemingly "forever" policies: "but we need our dairy tariff to protect our delicate cows and farmers and their ancestral cheese making practices" sounds so nice and reasonable, making it hard to ever get rid of. Neoliberalism is really bad at enacting policies that sound mean.
Even with all that said, I still am not convinced it's worth it. Or even that close to being worth it.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
I have it on good authority that the professionals also value "logistics" (warehouses, trucks, trains). Which isn't to say that the factory isn't useful, but the ability to ship it's outputs to where they are needed is important too. I'd be watching for investments in navy transport abilities there too.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
I have assumed the TikTok ban was at least in part retaliation for China's bans of Western tech companies, delayed until they had one worth banning. Which still fits the definition but feels different as a policy choice because we can say we didn't start it
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
Acquisitions often take the form of shareholder propositions for [party] to buy the stock for a given price, which is usually higher than the current price. They don't often occur purely as market orders, which would be vulnerable to price changes, that I've seen, although the buyer often has a decent ownership position to start with. Maybe the SEC requires or strongly incentivises the shareholder vote structure, though.
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
I always manage to kill everything I've planted (including an indoor succulent a friend gifted that I keep forgetting to water even occasionally). Any recommendations on getting started gardening as someone with no discernable green thumb, and who lives somewhere with dry, sweltering summers, and the yearly bout with a hard freeze or two?
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
Could you elaborate on the distinction here? I don't immediately see a difference in the simple case: if I import a widget for $1 completely manufactured from scratch in [country], I'd pay a percentage of that value as a tariff. Or I'd pay a (similar) percentage in a VAT regime, because it seems to me that in both cases all of the "value added" comes from one place. I guess there is a distinction for supply chains that go back and forth across the border in question to produce final products, but is that the modal case?
Would a tariff carve out for reimported intermediate products (excluding the value of American-made semiconductors used in iPhones, for example) meet your goals?
VoxelVexillologist
Multidimensional Radical Centrist
1mo ago
I don't think our philosophy of science has a good way to handle non-repeatable results. If you look at something like the Oh-My-God particle detected exactly once in 1991, I'm not sure how I'd distinguish from a miracle. Sure, a scientific instrument saw it, but those aren't immune to weird things, like the faster-than-light observed neutrinos a decade ago. As a one-off observation, it's a bit more believable than, say, a coherent message, but if we instead observed the alien equivalent of the Arecibo message (sent exactly once in 1974), we'd be talking about something that would look, to me at least, rather miraculous.
I haven't run the numbers myself, but I would be thoroughly unsurprised to find out that the large spike in murders starting mid 2020, which is IMO at least partially attributable to "BLM," actually caused an increase in the total number of murdered black lives. Uncharitably, "The purpose of BLM is to secure sinecures for friendly academics" seems a POSIWID-framing of the situation, which I'm partially inclined to believe as someone who actually wants to care about (all) lives.
More options
Context Copy link