robryk @robryk@qoto.org

To be more precise: the interesting thing is that one can interpret D_JS(A||B) as a mutual information between something for any pair of distributions. (One can do the inverse with D_KL: I(A;B) for any two variables can be interpreted as D_KL between some two distributions.)

TIL (somewhat embarrassingly) that https://en.wikipedia.org/wiki/Jensen%E2%80%93Shannon_divergence provides a connection of sorts between mutual information and KL divergence

@krzyz Ah, a thing I was missing was that `call` wants mutable access to Service. Thus, if we wanted to make `call` a method on the token, we'd need to have the token borrow the service mutably. However, that we cannot do: the token appears as a result of the future sometime in an uncertain future. So, we need to have call take both a &mut Service and the token, which sort-of requires the ability to have the type system enforce that we use the token with the service it came from.

I see why this doesn't work straightforwardly, but I don't see why GATs solve this issue and what does the issue have to do with behaviour on dropping.

Why isn't hyper's Service ensuring the poll_ready invariant with typing?

Docs (https://docs.rs/hyper/0.14.18/hyper/service/trait.Service.html#required-methods) say that the user is supposed to wait until poll_ready returns Ready(Ok()) with calling call. Why not instead have poll_ready return Poll<SomeType, Error> and have `call` be a function on SomeType (that takes it by value and consumes it)?

#rust

@unascribed It makes me sad that we have to resort to such half-measures that destroy the users' mental model of where security boundaries are, because apps try to create boundaries that can't exist. :(

@unascribed Huh? If it's so common to find this implemented on purpose, then how come we don't have the same thing being exploited by e.g. the "monetization frameworks" that pay app developers to include them in their app? Or do we?

@unascribed Unless there's a "url shortener" involved, that requests the url at realtime and issues a redirect. But then I'm not sure if handling of the url would be in browser's hands.

Anyway, this is a bug in discord and not in qr code readers: calling externally visible activities is something that should not be understood to carry user intent, because any app can do that.

@unascribed So, what happens if I redirect someone to that URL from a website I own? How is scanning the QR code not equivalent to opening a URL? If it is equivalent, why don't we see scans where people put redirects to that on random spam websites, or send people shortened urls to that via email, or sth like that?

@ckrypto Also, I would greatly appreciate a transcript of the image when you post images in the future.

@ckrypto WDYM by capitalism? The way I understand the term, agrarian societies of early 20th century Europe were capitalistic and didn't have these properties.

@icedquinn

You can view the patient as analogous to the military, or to all the people in the country. You could say the same about the country as the whole not having "another doctor" to go to.

That said, I'm not trying to say that these two situations are near-exact parallels of each other. I was trying to point out that we often view skin in the game as antithetical to objectivity. My examples did not have all the properties of the decision-to-declare-war one, but the reason we think people might lack objectivity in them is that they have too much skin in the game. Why shouldn't those problems manifest here too, however they are exactly created.

@Hyolobrika@berserker.town @dhfir

@Hyolobrika@berserker.town

A devil's advocate take on your approach (I don't know what I agree with here, but am willing to discuss from this POV):

There are situations when it's generally believed that the person with the skin in the game makes worse choices than someone without. I think the medical field believes that strongly: treating one's family or oneself is considered unethical in most circumstances and always discouraged. Performing noninvasive "fishing expeditions" for problems that show no symptoms is generally discouraged (except some cases where we encourage it for patients from some demographic), lest false positives cause patients to strongly wish for treatment.

All these examples suggest that people with more skin in the game might be over-eager to do something instead of nothing when a choice with unclear calculus of benefits appears. However, we are in a more confused situation regarding war: conscripts have much more skin in the game of war than in the game of suffering from consequences of not entering a war. From a purely selfish point of view, if a war can be avoided by means that cause the same amount of suffering by means other than military, all the conscripts should vote against it, because the suffering will be distributed among the whole population.

I don't think these two problems cancel each other. Each of them is a problem of biased noise, and even if the magnitudes of both are similar, they are likely far from anticorrelated.

@icedquinn @dhfir

@icedquinn

Sorry for being imprecise, you are right in what you said just now. The part I tried to point at was:
> Smedley's version was i think they had to put someone from their immediate family up for a combat role.

@Hyolobrika@berserker.town @dhfir

@icedquinn

Smedley seems to be describing the same system that Heinlein was: as far as decisions to declare war are concerned, he's proposing only people who would be drafted to be eligible to vote (in https://archive.org/details/war-is-a-racket-smedley). Did he say something else sometime?

@Hyolobrika@berserker.town @dhfir

When one responds to a boosted post, wouldn't it make sense to by default cc the booster? Is there some obvious reason why that would be unexpected/harmful?

#mastodon