I mean, their counter-parties know all of this, but the fact that PE assets don't need to be marked to market on a regular basis can be good for a lot of these investors, as it introduces a delay in the spiral that can otherwise occur with public assets.

Like, if AI collapses, everyone's gonna sell Treasuries to cover losses as they are super liquid (mostly), but the PE assets can pretend that they're still worth whatever, thus reducing margin calls.

PE is generally bad, but their LP's are not entirely stupid and the ability to mark to imagination is worth a bunch of money sometimes.

To be fair, this is because the US figured this stuff out way earlier through credit cards, and now there's a bunch of stakeholders and legacy changes which get in the way of making the services better.

Indeed, and there are some good reasons, too: US regulators want to prop up smaller regional banks and avoid large national monopolies (for what is essentially a natural monopoly).

The externalities of the crappy US banking system are so vast though. Musk, crypto, ...

> Maybe you get put on a list so US banks can't send you money anymore too.

This is a good example, because the US government routinely passes laws that prevent people from transacting using the dollar system (which is basically the world financial system) and this is OK, but the EU requiring companies that operate in their market to obey different laws is not OK?

I don't really get the logic here, but perhaps I'm missing something.

> And this is why the EU is stagnant and unable to innovate.

Can you help me understand how the EU is stagnant? Granted, they have lower economic growth than the US, but they're (mostly) not running large fiscal deficits.

And unable to innovate is quite simply, untrue. Deepmind (you know the people that invented LLMs) were a UK based company and were purchased by Google. Spotify & Skype were also both relatively innovative.

If by innovative, you mean are highly valued in the stock market above what a rational person would pay, then yeah Europe doesn't have as much "innovation". Now, if there was a single EU capital market (which honestly should be in London, despite the political complexities) then that might not be true.

Also worth noting that a lot of the US market is propped up by EU/EEA investors. Like, the Norwegian oil fund owns an appreciable amount of the US stock market. What would happen if all the European money was withdrawn from the US market? Nothing good for US "innovation".

And on the core point here, social media is now the public sphere, and as such is definitely worthy of investigation by academics. Like, if FB can do this (with much more personal data) then Twitter/X can do it. In fact, it would be super easy as they used to do it before Elon decided to attempt to monetise it badly.

Like, most studies of social media were performed on Twitter data, precisely because of this.

Do you think Russians have a better quality of life than Germans?

If they do the same thing to Republican presidents then I'll believe it's principled. I haven't seen much sign of that, though.

Like, whatever happened to the major questions doctrine? Feels like that should apply to a bunch of Trump's actions.

It's a term of art. The audience for this paper would understand, rather like many denizens of HN know what TDD is.

The paper doesn't even use it consistently. At first it uses "all-cause mortality" to mean "all causes except COVID", and then in the results section it uses the same phrase to mean "all causes including COVID". The whole purpose of terms of art is to increase the specificity of language, but they're not doing that here. Their usage of the term is confusing.

Edit: I'm wrong. I could have sworn it said that the groups had similar all-cause mortality, but it doesn't.

Where do you see them using "all-cause mortality" to mean "all causes except COVID" in the beginning? I skimmed over all uses of the term before the "Results" chapter, none of them seem to exclude COVID deaths?

I mean, it's plausibly something around trust in medical authorities which correlates with both vaccine and other treatment refusal.

I think that's mostly fair, but given that we can't randomly assign vaccine administration this sort of study is the best we're gonna get.

Like, the major takeaway is that the vaccine is safe, I think that we've already established that it works to reduce Covid hospitalisations.

More importantly, you are not your code.

This is actually harder for more senior/managerial folks, as often they'll build/buy/create something that's big for their level and now they're committed to this particular approach, which can end up being a real problem, particularly in smaller orgs.

Once upon a time, I worked for a lead who got really frustrated with our codebase and decided to re-write it (over the weekends). This person shipped a POC pretty quickly, and got management buy-in but then it turned out that it would take a lot more work to make it work with everything else.

We persevered, and moved over the code (while still hitting the product requirements) over a two year period. As we were finishing the last part, it became apparent that the problem that we now needed to solve was a different one, and all that work turned out to be pointless.