It took me approximately 6 months to get a billing dispute resolved with Google. Somehow my maps key got leaked, and someone ran up 1.8k in charges on it.
Super, super painful. That being said, I'm still using Google for geocoding (mostly batch) because their service works better for my data.
I mean, the easy thing to do is revoke US-EU data transfers. The ECJ is definitely going to do it anyway, and it provides a lot of leverage. If you keep it gone for long enough, then the Mag7 are basically forced to store data in the EU, and respect the laws.
As a bonus, it would nuke the markets, causing the US administration to backpedal on whatever. (Obviously I'd prefer not to nuke the markets, but something needs to happen to push back against the US).
This would only happen in a world where the US has entirely abandoned Ukraine though (i.e. no intelligence sharing).
> This. Companies like Nvidia, Google et-al and investors, don't care about and won't leave the US over morals, they'll go and stay where the money is good as long as it lasts. Trying to lecture them about morals from the EU won't change this. Otherwise they wouldn't be using slave labor in Congo and sweatshop labor in China.
Nobody will leave over morals (well except possibly the Norweigan sovereign wealth fund), but it's worth noting that for non-dollar investors, the US markets have basically been flat since the start of 2025, because the dollar has declined.
It's entirely possible that the US no longer takes in more global capital, if this continues. It's very unlikely that all the foreign investors will leave quicker, but it's much more likely that they'll leave as they sell their investments over time.
If investors leave, where will they go though? Most of EU economy isn't doing amazing right now either, with the economies of France and Germany being propped up on life support by government spending, and there's more political turmoil at the horizon. Asia?
Large European pension funds are rapidly decreasing (as rapid as a pension fund can without causing too much panic, devaluing remaining assets). E.g. some large pension funds have removed 1/4th of their investments in the US in less than a year. That is pretty unheard of.
>some large pension funds have removed 1/4th of their investments in the US in less than a year.
I saw the news about the danish fund dropping some of their US investment and on closer inspection, in absolute terms it was a drop in the bucket. Mostly an optics maneuvre.
Again, non dollar investors are flat since start of 2025. This isn't just politicisation (although that's part of it), it's that other markets are doing better than the US for now.
This will be a slow process, but the direction seems pretty clear (I fully expect to see a major economy introduce capital controls within the next twenty years).
> it's that other markets are doing better than the US for now.
Which? US currently has a rocky status due to Trump's interference, but Trump will pass while the likes of EU and Japan won't be able to fix their structural issues of low birthrates, crazy high debt welfare speeding, etc.
> Which? US currently has a rocky status due to Trump's interference
In non-dollar terms, the US markets have been flat since 2025 (so basically since "liberation day").
> fix their structural issues of low birthrates,
This is a problem basically everywhere. It's definitely worse in Europe than in the US, but the US is on the same trajectory (modulo immigration).
> crazy high debt welfare speeding
Where exactly are you talking about? The US government has been spending more than it takes in for the past decade at least, mostly on entitlements (i.e. welfare spending).
A single Dutch pension fund that was much larger (ABP, IIRC one or two orders of magnitude) retracted 1/4th (10 billion). But they only found out after journalists checked out a year report. Most pension funds just don't talk about it, because (1) they do not want the value of their assets drop too much as long as they haven't moved them; and (2) they do not want to draw the ire of the Trump administration in the meanwhile.
> the economies of France and Germany being propped up on life support by government spending
The US government is running (and has been for at least a decade now) a substantial deficit, which is basically propping up the economy with government support.
> there's more political turmoil at the horizon
Again, look to your own house. Even if you ignore all the Trump noise, the attempted politicisation of the Fed is very dangerous for the US economy.
> Asia?
Asia & Europe. It's beyond absurd that the US stock markets have 65% of total value, and was never going to last forever. All this craziness from the government is just speeding up something that was always going to happen.
>Even if you ignore all the Trump noise, the attempted politicisation of the Fed is very dangerous for the US economy.
Yes, but Trump is a passing issue that will eventually go away, and won't be able to fuck with tarries and the economy anymore just so his friends can do insider trading.
>Asia & Europe.
why do you think so? Japan's economy has no great future prospects, and neither EU's with many German bankruptcies and companies relocating abroad. Chinese companies and workers outside of the largest metro areas have bad time too.
> why do you think so? Japan's economy has no great future prospects, and neither EU's with many German bankruptcies and companies relocating abroad. Chinese companies and workers outside of the largest metro areas have bad time too.
I think Europe has a bunch of issues, mostly downstream of energy prices. It's basically all the low value added manufacturing (fertiliser etc) that is leaving. This is a real problem which needs to be fixed (the energy costs, that is). At least we're not trying to kill green energy.
Look man, I think the US will be ~fine, but a situation where the US markets account for 65% of world equity value is not going to last forever, and I think it will become apparent as this changes how much else needs to change to benefit the American people.
> You do see some large players like in Canada and Europe, but even they are not similar in size to American pension funds and endowments, let alone other American institutional investors.
Look, I haven't dug into this, but if one wants a fair comparison, then you need to account for the size of an economy. If 330mn people need pensions, then you'll obviously see much larger pension funds. If 400mn people across 27 countries want pensions, these will often end up being national level and will look individually much smaller than the ones from the US, purely because the US has more people.
> Unfortunately most European countries don't pension funds. It's a pity...
Many Europeans prefer bank deposits to investment in markets, that's true. I assure you though, there are lots and lots of pension funds in Europe, as well as many, many insurance companies who represent similar capital profiles.
> With the amount of talent working on this problem, you would be unwise to bet against it being solved, for any reasonable definition of solved.
I'm honestly not sure how this issue could be solved. Like, fundamentally LLMs are next (or N-forward) token predictors. They don't have any way (in and of themselves) to ground their token generations, and given that token N is dependent on all of tokens (1...n-1) then small discrepancies can easily spiral out of control.
To solve it doesn't mean we have to eliminate it completely. I think GPT has solved it to enough extent that it is reliable. You can't get it to easily hallucinate.
It depends on how much context is in the training data. I find that they make stuff up more in places where there isn't enough context (so more often in internal $work stuff).
Yeah, closed to new entrants in 2015, entirely gone by 2020.
Hilariously enough, all of the anti tax avoidance actually ended up getting Ireland an absolutely absurd amount of corporation tax (which we'll almost certainly waste, but them's the breaks I suppose).
Yup, the term ethnic cleansing became popular during the Yugoslavian civil war, so that UN states didn't have a legal obligation to intervene, as they would in the case of genocide.
Ethnic cleansing and genocide are obviously not the same. If Israel's intent were to kill virtually all Gazans, that would be genocide, but it seems very plausible that they would be entirely satisfied if all Gazans just left Gaza, which would be ethnic cleansing.
Look, I remember this being discussed at the time as a euphemism to avoid the necessity for intervention.
The wikipedia article suggests that I'm not alone in this belief:
"Both the definition and charge of ethnic cleansing is often disputed, with some researchers including and others excluding coercive assimilation or mass killings as a means of depopulating an area of a particular group,[6][7] or calling it a euphemism for genocide or cultural genocide.[8][9]"
This study seems both wildly underpowered, chooses relatively bad methods for analysis (splitting between below average and above average, really?) and has far too many comparisions (they claim to have used bonferroni corrections where necessary, and then don't mention it again).
All in all, it's a pretty good example of modern psychological research. Bad statistics, hyped up findings and (probably) wild over-generalisation about what this tiny study means for society/the world/my research funding.
I think the biggest flaw here is around the measurement of VO2 max. So they first ask people how often they exercise (cos no-one ever lies) and then use a linear equation to map that to VO2 max. Granted, the equation has lines for age and sex, so it's not entirely useless, but if you're only going to sample 40 students then why not try to measure things a little better?
The most unintentionally hilarious thing about their methods section is that age was not normally distributed (according to a shapiro test). They sampled students, so of course it's not going to be normally distributed. Students have a well-known bias in age.
Ultimately, regardless of whether or not this finding turns out to be true (I like the idea of it, myself) this study provides absolutely no evidence for the effect.
Note: I have a PhD in psychology, but left the field a decade plus ago. I'm both horrified by this study, and also having a lot of fun poking holes in it, maybe I should try to get back into reviewing? ;)
I get the KYC concerns for API access, but I'm sortof baffled at why they'd need all of the AML stuff, given that they're not payment processors/financial institutions.
Or does Persona provide that by default? Don't know much about their service...
Super, super painful. That being said, I'm still using Google for geocoding (mostly batch) because their service works better for my data.
reply