Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)C
Posts
9
Comments
326
Joined
3 yr. ago

  • Hash functions are not known to be quantum vulnerable (i.e., there's no known quantum algorithm that provides an exponential speedup, best you can do is to use Grover's algorithm to slightly speed up the brute force search). So maybe never.

  • Well, that's not the notable bit. Almost every year has seen increasing global CO2 emissions since the start of the industrial revolution. But what's interesting is that this year or next year are when global emissions are projected to peak, and then come down. So sometime next year, fingers crossed, we'll be seeing a different headline.

  • BYD also put lots of resources into electric buses. Anyway the point is that there's multiple game plans EV makers can follow, not only Tesla's.

  • If you think LLMs hallucinate too much, wait till you check out code literally written during hallucinations.

  • There's more than one way to skin a cat. The Chinese EV companies that have come up in the last few years use a diversity of business strategies, not all involving high margin SUVs. BYD's cars, for example, are spinoffs of its battery manufacturing business.

  • At least having those GPUs training neural networks is vastly preferable to having them mining Bitcoin.

  • Pointing to Japanese money supply versus inflation is irrelevant because Japan doesn't fund its fiscal deficit via monetization. It issues debt, just like every other non-basket case economy on Earth.

    The distinction is important. Debt is tied to a promise to repay later. Monetization has no such promise, so it's functionally equivalent to issuing debt and then immediately defaulting. So long as lenders believe debt will be repaid, the effects are different from monetization.

  • You might be confusing debt issuance with money issuance.

    Governments often issue debt to fund various kinds of spending. And despite concerns about debt levels, they can have a pretty fuzzy relationship with inflation; Japan has public debt of over 200% of GDP, and an underinflation problem.

    But issuing money for the purposes of government spending -- the monetization of fiscal policy -- is almost always a bad idea, outside of wartime. The practice is behind every single episode of hyperinflation in economic history. And governments know this. Fiscal monetization is only resorted to by countries that have exhausted their ability to borrow; if cutting spending isn't politically feasible, the remaining resort is monetization. That's basically how you get to Argentina's situation.

    As for giving up monetary autonomy, it is indeed a serious drawback to dollarization. But this is a second order problem compared to the kinds of problems facing Argentina, like findng a guy bleeding out after a road accident, and worrying about his obesity.

  • Argentina's runaway inflation is caused by the central bank printing money (to finance the government's out of control spending). The rationale for dollarization is to remove the ability for the government to do this. It's not an inherently crazy idea, since (i) there are smaller Latin American countries that use the dollar, and (ii) the dollar is already used de facto for many purposes in Argentina because of how debased the peso has been. But there are lots of practical problems; notably, Argentina simply does not own enough dollars in the entire country to keep the economy running normally if they switch (whatever "normally" means for an economy like theirs).

  • Argentina's been so wrecked by the Peronists that this is the one place in the world where bringing in a "hard right" president may not be a bad thing. Milei probably doesn't have the right personality, or enough backing in the legislature, to have a successful presidency. But the alternative was worse.

  • Ideally, they'd just blow the entire $330M training an LLM, and release the weights. In reality, much of that money will probably go into paying salaries, various smaller research projects, etc.

  • He's not wrong. But it's worth remembering that when China faced a far smaller provocation from their own restive Muslims, in Xinjiang, they responded by locking up a large fraction of the population in vast reeducation camps...

  • As a rule, multinationals agree to do what the government wants. You'll notice that TikTok also agreed to follow the EU's various new rules on social media. It's only a few American firms that have the power and chutzpah to ignore local governments and do whatever they want.

  • Reading between the lines, it sounds like one of the other Five Eyes (if it was Canada, it would have been easier to say so) bugged Indian diplomatic comms. Probably US or UK.

  • It's not 15 or 20 years ahead of time, though. The "ghost cities" came alive within only a few years; for example, this page points to Zhengdong New District, which was singled out as a ghost city by 60 Minutes in 2013. It had a population of 5 million seven years later. For district development (as opposed to constructing a single building), seven years is nothing.

    Coming back to their current property crisis: let's assume the article is correct that there's an excess of 7 million homes. We can plug this into China's current urbanization rate, and suppose China will get to South Korea's urbanization rate in 20 years (that's roughly how far they're behind SK, by GDP per capita). At one home for every 3.5 people, they need 3.4 million homes per year on average. So they have overshot by about 2 years, which is hardly going to make buildings crumble.

  • Enough of them filled up that even the press outlets that pushed the ghost cities narrative most aggressively, like Bloomberg, have run follow-up stories acknowledging it.

    Yes, some developments worked out and others didn't, but building out housing in advance of increasing urbanization is a good thing, not a bad thing. It's how you avoid housing unaffordability in urban centers, or worse, the rise of slums.

  • Nah. China's urbanization rate is currently 65%. South Korea for comparison has 82% urbanization rate. So the Chinese have plenty more (say, a hundred million or so more) homes to build. The current difficulties are more to do with (i) loss of consumer confidence caused by the leadership's bad economic management, and (ii) the deliberate restriction of credit to developers because of the government's concerns about debt.

    This analysis reminds me of the hoo-hah about China's "ghost cities" circa 2010. Those ghost cities ended up being filled up.

  • Vietnam ending up getting F-16s would be the very height of historical irony. It's good stuff, and example #39572 of how Xi Jinping's CCP has botched their foreign policy.

  • I almost wish for a nuclear winter. On account of patrolling the Mojave so much...