@VoxelVexillologist's banner p

VoxelVexillologist

๐Ÿ‡บ๐Ÿ‡ธ Multidimensional Radical Centrist

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

				

User ID: 64

VoxelVexillologist

๐Ÿ‡บ๐Ÿ‡ธ Multidimensional Radical Centrist

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

					

No bio...


					

User ID: 64

Silicon Valley VCs seem hugely reluctant to back hardware startups, and even then, those take longer to come to market. It's not my wheelhouse, but I feel like we've seen lots of robotics demos recently that were previously the domain of Boston Dynamics. I've assumed those were driven by machine learning and neural models, as opposed to cheap MEMS sensors that have been around at least a decade now, but I'd be curious to learn more if you have links.

Aren't the head "admins" typically drawn from the tenured faculty (sometimes hired across institutions)? I doubt there is a college president (or provost, or dean, or department head) out there without a doctorate. Maybe not all the mid-level admins ("Deputy Title IX Coordinator"), but those aren't intended to be steering the ship.

Snarkly: As I've heard it described, it doesn't include freedom from consequences.

Do you think biologists are about to walk all their support of that back, in favor of what used to be (and, in my mind, still is) an unquestionably obvious conclusion?

Observably, the medical industry seems to have been able to successfully walk back from several scandals of seemingly similar magnitude (on a logarithmic scale). The lobotomy as a procedure won a Nobel Prize in medicine. The sister of a future president had one! It had similar arguments over its ethics and efficacy, but in the end we don't do them anymore (I hear there are some rare similar procedures, with much more oversight and gatekeeping). It's largely been swept under the rug, though there wasn't zero introspection on the topic. At smaller scales I could point to "repressed memories," Freudian analysis, and such.

But it might take a decent fraction of a generation.

While it is important to consider the strongest ideas of a movement, so as not to be knocking down straw or weak men, the insistence on that coupled with the lack of disavowal on their end makes for an insidious motte and bailey.

Doubly so because Scienceโ„ข claims to be the process by which we find "strongest ideas" generally. It's both a direct and a meta-level failure.

Heh, I'm old enough to have owned a pocket electronic spell checker at one point. The hash table seems the right way to do it these days, but it will take up more memory (640K shakes fist at cloud). And sometimes you do want to scan faster than the user types, like opening a new large file.

Software spell check (on early computers at least) required a cute algorithm --- the Bloom filter --- to work reasonably efficiently. Actually checking each typed word against the whole dictionary wasn't (and likely isn't) practical, but a statistical guess of correctness is good enough.

Interestingly enough, using a computer is a mitigating circumstance, doing the same without a computer, network etc carries a penalty of two to 15 years.

At first glance this sounds wrong, but I'm not sure I'm actually upset that trying to "lure a child" IRL (effectively) is considered worse than typing into the Ethernet on a keyboard.

As an example, deployed US soldiers and others stationed in allied nations often get processed by the American (UCMJ, as appropriate) justice system. It's obviously unfortunate when it happens, and sometimes leads to local protests and upset local officials, but at the high level neither side sees it as worth ending the arrangement (historically, "better US troops than Soviet ones", I'm sure). See the death of Harry Dunn in the UK and various incidents in Okinawa and Germany.

I'm not here to defend the process, merely to note that it happens.

LLMs truely can boost productivity in most tasks except for pure manual labour we would see GDP roaring.

Ironically, I personally suspect that the current AI revolution may find its best use cases in manual labor fields. Replacing software developers is sexy, but hard. Training a model to operate robotics for conventionally "hard" problems (picking fruit, sewing garments, sorting trash) that require a bit of intelligence due to inconsistent inputs (fruits differ, fabric bunches in non-deterministic ways, trash isn't worth sorting) seems much more viable than groking a full software stack.

This makes sense to me. The impact of the spreadsheet has been huge, but it took a long time to settle in everywhere, and the accounting department still exists, even though the guys "running the numbers" don't anymore. There are still plenty of operational systems running on DOS or OS/2: if it isn't broken, don't fix it, and things take time to replace.

The annual US revenue of Novo Nordisk (Semaglutide) seems to be 45G$,

Isn't it a Danish company? I've wondered about that because I've seen lots of concern about cuts to research funding citing GLP medications as an example, but it seems odd that my tax dollars payed for the research, and now I would have to pay the Danes, as it were, to use it.

Obligatory: "The Earth isn't a sphere, it's an oblate spheroid."

"Actually, I prefer an equipotential geoid model. EGM84 or better."

Observably, humans have these same problems, or ones that look similar from a distance: organizations run with sycophantic "yes men" seem to produce worse output broadly across fields from engineering to governance to film production. Is it really surprising that universally warm and empathetic text responses don't always produce "good" outcomes and sometimes reassure our worst instincts? It takes a certain level of, well, something to want friends and colleagues that will challenge your bad ideas.

I generally only play with free models, but I've had ChatGPT tell me an idea was wonderful and I should look to get it published despite me knowing that it had clear flaws in the math.

Note that Columbia ended up settling a lawsuit with the guy in question on account of (presumed) Title IX violations in the course of its disciplinary actions and allowing the performance art piece, including issuing a formal apology.

super homogeneous

Is there any other country in Europe that has as many official languages (4, not including English as a common lingua franca)? Granted, Singapore has that many too, but I'd hesitate on calling either "homogeneous" across the board.

Just finished Spring Snow. I'd seen it recommended a few different places (maybe here, maybe HN). I've never been a weeb, but I've visited Japan -- it's a beautiful and very interesting country -- and I can appreciate why Mishima is seen as such a prominent Japanese writer of the 20th century. Some of the vibe was to be expected, like the very Japanese aesthetics, and the tension between Japanese traditions and incoming Western norms during the Meiji era.

That said, I was intrigued at the following author blurb at the back of the book:

In 1970, at the age of forty-five and the day after completing the last novel in the Fertility series, Mishima committed seppuku (ritual suicide) -- a spectacular death that attracted worldwide attention.

I felt I had to read a bit more to understand this. Wikipedia tells the story slightly differently:

[Mishima] was a Japanese author, poet, playwright, actor, model, Shintoist, ultranationalist, and the leader of an attempted coup d'รฉtat that culminated in his seppuku (ritual suicide).

Well, that escalated quickly. Having read that article, it's interesting to see how story beats from the author's life: "After briefly considering marriage with Michiko Shลda, who later married Crown Prince Akihito and became Empress Michiko" sounds a lot like the story of the novel.

Looking for something shorter and more sci-fi, I've picked up Asimov's The End of Eternity, which I think someone recommended a while back. I might consider continuing the Fertility series at some point, though.

To be fair, the fact that Costco (et al) sell giant containers of acetaminophen is kinda scary to me. It's substantially more dangerous than naproxen (also available in that size). The 24-count restriction sounds pretty reasonable to me.

During WW2, China lost nearly all its modern equipment and trained forces in the early days of the war, leaving them to fight the remainder of the war with only obsolete or crudely made small arms, against an enemy with machine guns, trucks, tanks, artillery, and air support.

The Allies (primarily the US) provided a total of something like 650,000 tons of materiel to China via The Hump, at the cost of nearly 600 aircraft and around 1700 crew lost. There were also notable Allied air units (and some ground units) in China during the war.

Which isn't to diminish their accomplishment, but it wasn't completely a solo effort.

Red tribe defection doesn't look like random acts of terror.

It demonstrably can: look at Oklahoma City. But I think you're right in the general case.

One of the biggest sources of illegal guns is parked cars.

I'd be curious what fraction of guns left in parked cars are left their because their owners can't legally keep them on their person where they've parked --- bars (depending on state), private property that disallows carrying, etc. I'm not going to blindly use that to push for legalizing carry everywhere (honestly, probably not my preference), but the numbers would at least be interesting.

There's probably a defensible, if aggressive, claim that the universal suffrage has proven more dangerous (in terms of deaths per capita) than (universal) firearms ownership. Definitely some error bars on the hypothetical impact of more gun ownership, and exactly which government actions are attributable to "voters" specifically --- How many bodies sit at the feet of voters in the 1932 Weimar Germany elections? Every victim of state violence in democratic states? Do we count communist "elections"? --- but I think the Libertarian premise that "nothing is more dangerous to a people than their government" has at least some academic merit.

Yeah, the distinction there is pretty subtle, and I can't think of a case in which it's ever actually mattered. Maybe we'll see that in the lawsuits over White House press pool access: is there a relevant act of Congress establishing that at all? But it generally seems to be considered to apply to the government as a whole.

the First Amendment protects citizens and probably permanent residents, but possibly not those on temporary visas.

The text of the First Amendment specifically binds Congress: "Congress shall make no law...", while other amendments say "the right of the people...". Courts have generally read the First pretty broadly with respect to free speech protections. Excluding the no-longer-binding Schenck, I can't think of too many cases the speech advocates have lost. On the other hand, I can't see anyone reading it as preventing a ban on literal propaganda from foreign enemies. Limitations on issued (or renewed) visas for foreign nationals on the basis of their speech seems rather borderline.

I don't think the Founders intended to require letting in visitors who call for, say, the violent overthrow of the US government --- citizens doing so are already barred from security clearances, for example, without too much fanfare.

In the past I've known people who've made a simple HTML page with some links on it set as the home page for easy access from new tabs. Not sure if file:// would send a referrer, but a lightweight local-only server isn't crazy either. I doubt they know they're sending referrer headers.