@VoxelVexillologist's banner p

VoxelVexillologist

πŸ‡ΊπŸ‡Έ Multidimensional Radical Centrist

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

				

User ID: 64

VoxelVexillologist

πŸ‡ΊπŸ‡Έ Multidimensional Radical Centrist

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

					

No bio...


					

User ID: 64

That might be a concern at some point, but current hardware that can (effectively) run bleeding edge models is really rare. A machine that can effectively run GLM 5.3 Flash (not a frontier model even, but close-ish) at nontrivial rates is, last I checked, six figures. Can it even run on a single 12A (120v US) breaker? A rogue frontier AI isn't going to spin up instances on smart TVs or even most desktops.

That might change in the next decade assuming demand and technology make these things commonplace, and that is worth preparing for, but seem oversold as risks today.

You would not expect a "pandemic safety organization" to create novel viruses.

There have been plenty of comments here about how much the CDC and NIH were funding gain of function research, despite a Congressional ban, some of which was done in Wuhan, and some of that proposed included specific mutations later seen in a pandemic coronavirus that first appeared in Wuhan. Whether or not this was the source of said pandemic remains unproven.

Isn't all of it hobbled without login now? There are some mirror services, but I don't expect those to last forever.

I'm sure Aaron would have been proud. /s

Surely Providence should have thought of, uh, safety before allowing this cellular life stuff to engage in evolution by natural selection. You can't just let that happen, or FOOM.

But that same Creator also didn't consider the gray green goo hypothesis, and now that stuff is everywhere, even deep underground and ocean vents.

The argument in favor of the "top ten six five percent" rule was that high schools were heavily segregated in practice --- barely now within living memory, by fiat, but now mostly by school zoning area demographics. Ideally the winners are supposed to be poor schools in the Rio Grande Valley and historically-Black communities getting to send graduates from poorly-ranked high schools to top-tier UT Austin, but I've definitely read anecdotes of parents of means, uh, venue shopping come senior year of high school. I can't speak to how common that is, though.

The most rabidly xenophobic nations are generally the ones people are least interested in migrating to

This partly depends on your definition of xenophobia. I've known no shortage of weebs dismayed that Japanese immigration laws are basically "no". I'm sure there are KPop stans that have similar situations.

Japan is an odd example here because general claims of "Japanese success" happened mostly after a dramatic period in their history in which Western culture was aggressively adopted by (local!) top-down fiat.

I'd be remiss if I suggested that Western ideals were the only reason for this success, but it certainly complicates the example. China has also had its own run-ins and adaptations from Western ideas too (literally Communism! Although that's not the only one).

That lets their population grow yet further past their decreasing food supply.

It's not a requirement that any region, state, or even nation be self-supporting in terms of food production, even if it is probably a good idea where possible for a few regions. But Singapore is never going to grow enough food for itself, for example.

The writing style was far clunkier than I expected

It's been a few years, but I recall thinking that the prose in Hyperion was pretty good. It's not something most (many?) science fiction authors are very strong in, though, and it may have been a relative observation.

It's interesting the Houthis have been remarkably restrained.

One possibility, although I have no specific evidence for it, is that the open conflict has allowed interdiction (or just dissuaded) Iranian shipping to their proxies. Sinking Iranian-flagged (or obvious false flagged) ships smuggling missiles in peacetime would cause trouble; less so today.

I don't think Yemen is making it's own long-range ballistic or anti-ship missiles. Small arms seem less of a logistical concern. But you could be right that restraint is being exercised locally.

Assuming we don't all die in the process (IMO unlikely, but not ruled out), I look forward to explaining our assumption to our new overlords that they'd go all murder robot on us. Ripley wasn't wrong to say "nuke it from orbit", but I'd expect bringing thermonuclear warheads to a First Contact event would be poor form.

The bar for "tried" here isn't related to success rates, but to whether she thought it would be effective. It is, admittedly, a more difficult thing to prove either way.

Anything going by sea to or from Bahrain has to pass the Strait: it could just be ineffective (high risk) for logistics at the moment compared to more distant but clear options. But you could also be right.

Time-wise, I would expect some Iranian (or proxy) action or attempted action to embarrass the US before the mid-term elections. I'd also expect the Americans to expect such a plan. "Baiting" like you suggest doesn't sound crazy.

I have seen maps that split across the Atlantic rather than the Pacific. IIRC China prefers that choice. Seems like you'd be less likely to lose New Zealand (a common map error), too.

I suppose on that front it's not surprising that the Hugging Face incident didn't occur at Meta or Google. But the hubris of self-declared experts here is ironic, at best.

the general perspective projection

Anyone who adopts anything else thinks they can map the surface of the earth to a Euclidean plane: they're literally flat-Earthers! Although if you have concerns about perspective distortion I can be sympathetic: no maps -- everyone can only use globes!

/s

Aerospace welding has enough precision going into it that "we spent tons of money automating it to reduce process variation" (structural flaws caused by human tolerances) has long been on the table, regardless of the financial efficacy. Even literal amateurs have been building automation for this for quite a while.

The real jobs to watch are the ones that are cost-sensitive and sufficiently-varied to make automation difficult. I'd watch industrial textile sewing, fruit-picking, or things like that.

You're not wrong, but I usually find that incompetent people aren't particularly self-aware. Knowing that what you're doing is dangerous (they say so themselves!) and choosing to do it anyway (they did!) together IMHO go beyond incompetence and into questioning-motives territory. But I can't rule out that they are just spectacularly incompetent and should (be forced to?) step back from their goals for the sake of the rest of us.

I, for one, have trouble believing that the same folks preaching the dangers of ASI unwittingly used just a proxy to poorly-sandbox for "take off all the guardrails" pen testing. At least a few corporate networks I've known have, pre-AI, built more layers of security than this. Accounts of "my work computer can't access the Internet" are something I've heard plenty of times.

If this is their normal model (maybe believable: move fast and break things), I'd also be concerned about the other direction: a nation-state actor would only need to pass through Artifactory (maybe a supply chain attack on hosted packages) to start egressing model weights, which is their entire trade secrets.

The methods of running a true air-gapped system are well-documented, and I'm pretty sure someone at OpenAI is already doing it for government contracting.

The Manhattan Project scientists had only recently convinced themselves that an atomic bomb wouldn't set off an atmospheric nitrogen chain reaction a couple years before Trinity. Despite their confidence at the time, I'd point to the Castle Bravo test a few years later (which was drastically larger than expected) as a suggestion their knowledge was imperfect.

But also it was wartime, which at least emotionally changes the calculus.

Art is meh, writing style is terrible. Part of this is the lack of RLVR

I think these are categorically different than math: the theorem prover languages are at least a clear Boolean success criterion. "Is this painting/book good?" is canonically an opinion question without a singular answer.

From what I can tell, human-driven post-training has historically been done with English speakers hireable at low hourly rates, but probably not ones that appreciate Ulysses (I certainly haven't read it). Even if they could afford English department grad students, would they be willing to train, effectively, their replacements? How do you rank the writing styles of Faulkner versus Hemingway? It seems there is at least some room for legitimate creativity there.

Gaming journalism also had a problem with fairly consistently giving high scores to games audiences didn't like. You see it occasionally in film with divergence between critic and audience scores (The Last Jedi was well received by critics), but games are longer than movies and sometimes it was clear the reviewers hadn't played much of the game at all.

As in literal slot machines? I'd always heard that the Vegas gaming authorities made that a surprisingly-regulated software industry, up there with aerospace or automotive. Not that those are perfect, but they at least have standards for their code to adhere to.

But maybe "write this in MISRA C" (automotive) isn't as much of a moat as it sounded, now that I think about it.

Previous results have been basically irrelevant to the sciences.

I don't know if you'd consider it, but Anthropic (Mythos) found an issue with HAWK, one of the finalists (but not the finalized standard) for post-quantum cryptographic signatures.