Tokamak style approaches for "sustainable fusion" require scale. There's no "sustainable fusion in the lab" unless your lab is at the scale of a full blown power plant.

This should be understood to be exactly that. The project itself would most likely not be a standalone commercial win -- if all goes right it would put out net electricity on the grid, but you have to consider the capital costs of the project. The real hope here would be that the engineering effort involved would produce a design that would then be commercially viable if built at larger scale, i.e. more and maybe larger plants.

A power grid does not need to be the worst in the world to complain about it.

Obviously Texas is its own special level of shitshow, and don't get me started about the fact power has cut out briefly twice in the past half hour at my hotel in Bangalore, making me miss the stable California grid back home.

This policy calls out redirecting ads "for monetization purposes." That is, an app that uses the VPNService API so that users from Moldova appear to ad providers to be in the USA so the app creator gets more money.

If a bona fide VPN makes you show up in a different country and so changes what ads you get served, that's an entirely different story.

The basic idea here is that, if it's possible for single cellular life to migrate between star systems through "fluke" events on timescales of hundreds of millions of years, you have orders of magnitude more planets out there where an unlikely "start of life" event could happen.

If the number of extra planets available in the galaxy times the probability/rate of transmission is enough higher than the rate of native life generation, most planets with life would get it through seeding rather than developing it natively.

On the other hand, we don't really have data for any of these probabilities. Unless we can refine any of these probabilities somehow or we start seeing signs that extrasolar life is very similar to life on Earth, there's little reason to have an opinion either way here.

I've lived in the SF Bay area for 16 years now. I'm not "from Texas and the South".

Prop 13 systematically lowers tax on people with multimillion dollar homes who bought their houses before the inflation of the 1970s, or inherited such houses, with relatively milder effects for people who bought more recently. Meanwhile, the rest of us pay thousands of dollars a month in rent, live with a bunch of roommates, or commute for hours.

The best part about this: Prop 13 homeowners get all the benefits of higher property values (higher cost of housing) but don't need to pay the higher property taxes that would ensue everywhere else. Then, for some reason the homeowners happily vote for city councils that rubber stamp more and denser office development, but the moment anyone starts talking about denser housing that would bring down housing costs -- or at least match the dense office development to keep prices stable -- they turn into entitled NIMBY pricks who whine about the "character of the community" and fight tooth and nail, bringing up issues like "burrowing owls" that mysteriously aren't a problem when bulldozing vast amounts of dirt to make a baseball diamond park in the same neighborhood.

Yeah, you're talking about the property tax law that creates a landed gentry of multimillion dollar home owners who pay negligible property taxes, while the rest of us pay through the nose on rent. If we do go buy a house we're saddled with a multimillion dollar mortgage and tens of thousands of dollars a year of property taxes the landed gentry don't need to pay.

No.

It is (usually) easy for C++ code to call into C code. The reverse, however, is not the case. While it is easy for C++ code to provide a C callable API via "extern C", it is not possible for C code to call directly into typical C++ APIs without a shim layer.

There is plenty of truth to attack the healthcare industry with without resorting to baseless allegations of widespread fraud in covid reporting.

If you think Japan didn't effectively close its borders, you're delusional.

To this day Japan is the only one of the G7 countries that doesn't allow tourists, and for months you couldn't even get a visa if you were moving there. "Special exceptional circumstances" only for non-residents/citizens, and for the first few months you couldn't even get back in the country as a resident. Heck, my friends living there were afraid to leave to visit family over the holidays for fear that they might close the borders again even to residents, as they did regionally in response to Delta and Omicron.

Also Japan doesn't use Ivermectin, at least in anything resembling a standard treatment.

And I will remark that, with the possible exception of South Africa, Japan is the only developed country on your Ivermectin list, and if you're holding up the numbers of borderline failed state Venezuela in comparison with Canada maybe you should rethink how in touch with reality your news sources are.

The reasons why those Ivermectin countries have such low counts is largely explainable by the effects -- such as robustness of reporting systems and demography -- of them being the least developed countries on that list.

This isn't Google Wifi -- OnHub was the thing before it, and the devices in question date from late 2015 into 2016.

Factoring and discrete log (to go from public key to private) are QBP -- they're fast on a quantum computer.

Finding interesting inversions of hash functions is not. "find a valid public key of length N with this hash" is susceptible to Grover search but that's still 2^(N/2) -- it's just the generic "you need to double the length of your keys" speedup that quantum computers provide.

The point is that quantum computers at this point are nowhere near big enough to do that. Integer factoring isn't interesting until it is hundreds of bits, and the machine in question has less than 100 bits so it can't even hold a number to be factored much less run Shor's algorithm on it.

The Google "quantum supremacy" thing is on a toy problem, because that's only how big/effective the machine was -- it was only at the point where it was seriously competitive with a datacenter/supercomputer when running against toy problems specifically designed to be easy for it to solve and hard for classical computers to solve. Nobody is claiming it isn't essentially a toy problem.

"Quantum supremacy" is a fuzzy milestone that marks the beginning of a time period for quantum computation starting from the first point where you can talk about quantum computers that outcompete classical computers on toy problems up until the point where they are useful for real problems.

The flip side is that lots of classical side optimization after the fact doesn't make a huge amount of difference. These problems get exponentially harder to solve for a classical computer as you add more qubits. Saying you can keep up with a 53 qubit computer isn't going to help that much when e.g. 75 qubit computers show up; it just shifts the exact point of a milestone on the way to having something that's actually useful.

The "20 to 30 years" away is because for decades they haven't been getting the investment needed to make them happen even on that timeline.

The real problem is that while large scale tokamak reactors along the lines of ITER are almost certainly viable once engineering problems are worked out, from an economic perspective they're hard to distinguish from large scale multi-gigawatt fission plants with no clear way to scale downwards -- you get extremely capital intensive plants (billions of dollars for gigawatts of production) with fairly cheap fuel costs. It will be a nice addition to the energy mix but it isn't the "too cheap to meter" panacea that some people expect.

Why do servers need to know what time zone they are in?

There are network protocols (e.g. Kerberos) that rely on synchronized clocks. Those protocols strongly want some concept of universal time -- using local time will break if two machines in different timezones speak to each other. That means the server certainly needs to know what the "universal" time is.

There are certainly usecases where the server wants to know what time (e.g.) the client is in, sure -- there are plenty of good reasons why a server would want to deal programmatically with timezones, at which point yes you need something like tzdata. However, unless you're dealing with some weird regulatory requirement it isn't clear why a server would need to deal with its own timezone. If anything you really don't want the behavior of a server to change if you're failing over to a different site.

If you're not hearing about this stuff you're living under a rock.

First off, -- you probably understand this but it is worth calling out explicitly -- the vaccines *do* have a very real preventative effect. The problem is that the "prevention" is far from 100%, for various reasons. We know obvious factors: how good is the early immune response triggered by the vaccine, to what extent to variants reduce that, how much does it drop over time, how large is the incoming viral dose at initial exposure. As far as I know we don't understand the relative importance of these factors, on a spectrum from "most fully vaccinated people are mostly immune but breakthrough cases happen, more frequently during a major surge" to "most people will eventually get exposed to a large enough viral dose to have a breakthrough infection".

It isn't that clear that a hypothetical attempt to vaccinate a largely immunologically naive population against polio or measles wouldn't have the same issues with breakthrough cases we're having with SARS-CoV-2. On top of that, measles *did* have outbreak levels of breakthrough cases until they added a second dose to the standard regimen after the 1989 Chicago measles outbreak. Even then breakthrough measles is still a thing -- it's just rare and almost always really mild when it happens. It may well be that SARS-CoV-2 vaccines similarly need an extra dose, or need the two doses to be further spaced than 3-4 weeks to provide robust long lasting immunity.

Still, regarding treatments: the biggest problem with respect to treatment is that by the time you know that a case is serious you're past the point where antivirals help that much -- at that point you're dealing as much with the immune system destroying the body. Dexamethazone was a big deal in improving the situation there, and came with something like a 1/3 reduction in chance of death.

On the other hand, there are a number of existing antiviral treatments that improve the situation if applied early on -- remdesivir and the various monoclonal antibodies -- but they don't help *that* much unless they come early and they all require hospitalization level treatment. If hospital systems are overwhelmed, treatments that require hospitalization of likely otherwise sub-hospitalization cases cases are not that helpful except for high risk patients.

There are two game changers here:
1. There's a fair bit of news about an oral covid treatment that Pfizer just launched phase 2/3 trials for. That has the same "needs to be early to be helpful" issues that existing antiviral treatments do, except that it would actually be viable to take them far more early -- at first positive test or even after known exposure.
2. In an environment where hospital systems are not overwhelmed with covid cases among the immunologically naive, even those more intensive early treatments would be far more helpful than they are now.