galimaufry

Agreed that Foxconn doesn't sound malicious here. But it does seem like more than typical bumbling.

In particular, I can't see how this management structure is supposed to work - competing subdivisions is fine, and having divisions just do what the CEO says seems like it could work, but combining them is wild.

I'm kind of surprised that being an island would help, actually. Is international travel into prosperous islands really less than into equally prosperous contintental countries? Most long-distance travel is by air anyway.

> That can easily be contrasted to Manhattan that was cleaned up by "broken windows policy".

I believe this is more a just-so story than a fact. While New York did use the broken windows policy, they also grew the police force by 35%[1]. Rather than a special effect of the broken windows policy, it could just be with more police they were able to solve/prevent more crimes.

One interesting thing about policing in the US, is that the US has fewer police than Europe and also more crime. People want to say interesting things about how politics or culture affect crime, but the simple story (where crime is high in the US because it is relatively easy to get away with it because there aren't many police) may be the best one.

[1] https://www.nber.org/digest/jan03/w9061.html

Have any games tried frequent savepoints + only allowing the player to reload 24 hours after dying? Seems like that would make reloads feel a bit more fresh/consequential, and also prevent addiction.

US higher ed already is a public service. Only 25% of students are in private colleges/universities, and ~every professor who does research is funded mostly by government.

I believe that what's actually happening with US higher ed supply is that we stopped founding large research universities. Many of the great public schools were endowed using federal land[1] and eventually the government ran out of valuable agricultural land.

[1] https://en.wikipedia.org/wiki/Land-grant_university

I run tensorflow on my laptop for small jobs, and AWS for big ones or in production. Not sure what the "push" is supposed to be.

I really don't think that Chester Pierce was an idiot who valued being emotional over being logical.

While there is disagreement over how prevalent they are or what if anything should be done about them, micro-aggressions obviously exist. I find this sort of hostility to the very concept perplexing.

> In China, English language students are taught to break that habit for fear of it being misinterpreted by a native English speaker.

Based on my Chinese colleagues this is not true, or at least is not universal, or it changed within the last 3 years.

Politics isn't going to solve this. You could elect an anti-PC theorist like Donald Trump, or an old school liberal like Nancy Pelosi, and this sort of thing would keep happening.

The real solution is just awareness. As people see more examples of this sort of thing going wrong they will be more careful about it.

Just spitballing, but you could have composition eat the name of an input to the outer function: Given f(x,y) and g(z), write

let q = (f .y g)(z=a) then q(y=b) = f(x=g(z=a), y=b)

let q = (f .x g)(z=a) then q(z=a) = f(y=g(z=a), x=b)

Aesthetically it would be kinda like the notation for fiber product.

I am wondering this as well. Disclaimer I did not read the paper.

On the one hand, there are infinitely many geodesics on a dodecahedron, of unbounded length, so you can't really brute force all of them.

On the other hand, the actual solution only goes through each face at most once. There are infinitely many unfoldings to try, but for each unfolding there are only finitely many paths to try, and you can get the solution on one of the minimal unfoldings.

Could just be very risk-averse plea-bargaining on Ramesh's part.

People do foot races even though gazelle is invented.

I'm using 'finite' in the same colloquial sense as GP:

> there is a fundamental difference between finite, rent-generating resources like land, and value-added goods

Sorry, I should have consistently specified "urban" land. Urban land is not finite (cities grow) and is very high value-added.

American policy makers have chosen policies (eg zoning law) that make it easy to produce valuable urban residences. Now they have a massive competitive advantage in high-end residences, so of course they will become exporters.

Land is just an export like any other.

China produces lots of manufactured goods, and sells them to the US. The US produces tons of valuable urban real estate and exports it to China. (Similarly the US produces lots of higher education, and exports it in the form of international students)

The point is that, with hindsight, we can say that this espionage was a good thing. It allowed the industrial revolution to spread to America, creating a major driver of global prosperity. Perhaps mankind should repeat this highly successful experiment.

I don't think that's true. There have been frauds in materials science, but people still trust it. The fields that are untrusted are the ones that are so hard that it's plausible that the conventional wisdom might be wrong even if everyone is honest.

This illustrates the problem with calling out virtue-signalling. I see two references related to ethics. It seems the only way to avoid accusations of virtue-signalling would be to have 0 references to AI fairness.

In practice, comments like this encourage self-censorship, even if that is not their intention.

> It is both, in the same way as a function is both its implementation and its "extension" (its effect on all inputs).

Note that, according to definition of 'function' most common in math, this is false. A function need not even be implementable.

Aside: As someone trained in math I've always found explanations of type theory "fractally confusing" because of stuff like this: type theorists' words mean something different from my words, and in the explanation of what they do mean, they use words that also mean something different.

A partial answer is that big fully connected neural networks _are_ pretty much untrainable. Neural networks only became successful once programmers started constraining the space they were optimizing over pretty radically (like requiring convolutional layers if you know you are trying to detect something local).

> It seems like the vast majority of photons would just fly through the universe without interacting with anything, unless there are other ways for photons to interact with matter besides being absorbed.

Maybe someone who actually knows will chime in, but afaik:

- Most light doesn't have a fixed frequency. If it did, it would have a fixed momentum, but then you would have no idea where it is! Instead it is some superposition of many frequencies. That could be part of the story.

- Light could be stopped by something other than absorption into an individual atom. Metals don't have discrete spectra.

I think the best motivation for the theory of Lie groups and Lie algebras is representation theory. Like you don't need anything besides linear algebra to know what SO_n is, but if you want to know about how it can act on a vector space then you need to think about the Lie algebra.

The other great thing about Lie groups is you can discover new and valuable groups just from pretty basic topology. Like the Spin group, which you know has to be out there as soon as you know the fundamental group of SO_n, but otherwise would be very hard to think of.

The fancypants but I think most intuitive way to think about Galois theory is also with topology. It's an algebraic version of a much more geometric, visible story, the correspondence between {subgroups of the fundamental group} and {normal covering spaces}.

> angular momentum is only conserved if you take spin into account as well.

Do you know an example of a process that moves angular momentum from one kind of spin to the other?

Are parents' incentives really that strong? We need all sorts of laws to prevent parents from exploiting or maltreating their children. It seems that parental instincts are not very reliable.

This. Preventing cheating in an online-only course requires meatspace infrastructure: testing facilities where students can be monitored. That infrastructure is still patchy.

> Why would someone like that work for an online school when they could work for one that would give them tenure, funding, lab space, grad students, etc?

Well, someone like that would love to not have to waste time lecturing. By not being on campus they could go to more conferences and spend more time in the field. They could recruit ideal grad students from anywhere on earth without worrying about immigration restrictions or family ties.

I strongly suspect that the cause of isolation is that we don't treat our workplaces as our community anymore.

My parents' community all met their spouses and friends at work, whereas people of my age seem to find something unprofessional about this. And I believe that old-fashioned civic organizations like the Elks were built from a tangle of professional relationships.

And realistically, where are you going to find a stronger community than among the people with whom you spend 40 hours a week?

The Bachelors by de Montherlant. About two socially isolated impoverished turn-of-the-century aristocrats. The plot is very bitter and the main characters are awful people so you'd expect it to be fairly cynical, but it isn't at all, really sympathetic in fact. The narration is insightful and hilarious.

The problem I have with this is that love is a complicated thing. I love playing soccer, but without the social validation of my friends who also play it I would absolutely never practice. Similarly for most careers, even highly lovable ones.