wahern

If you hosted your own email server, why would you need to be familiar with SES, especially if you've been hosting longer than SES providers have been around?

Like email, I run my own HTTP server, serving directly to clients. I know about Cloudflare just from HN threads, but have never used it or know much about that whole reverse proxy universe beyond the very low-level details from having written several reverse proxies myself.

It's not that simple because it's asymmetric. The companies track you (via firms that can see your purchases across different retailers) but how can you track the companies to find a bargain? Historically you might follow Consumer Reports, your neighbor's advice, or just look at advertised prices, but if everybody is offered a personalized price, now it becomes more difficult for you to compare or even identify prices. And what you're offered today might not be--and if Amazon is any indicator, won't be--the price offered tomorrow, making it more difficult to plan and budget across purchases (individuals don't have the scale and analytic prowess retailers have).

Fair price discrimination would involve some kind of bidding system, but that has high transaction costs for both sides. And naturally sellers aren't interested in maximizing global efficiency or competitiveness. Like any company they prefer "solutions" that give them an advantage over the consumer.

It still might be a net win overall, and I don't have strong opinions one way or another beyond being apprehensive about knee jerk public policy proposals, but surely the sellers will see a larger share of the benefits, possibly accelerating wealth inequality and people's growing sense of economic insecurity. That sense of security is critical for maintaining free (or freer) markets, otherwise the population is prone to using their political powers to try to take back bargaining power, and very often that winds up being a loss for everybody.

They can't know, they infer. AFAIU, normally they just detain you at the airport and harass you to try to break you. To jail you they're technically supposed to be confident enough about you knowing the password to be able to charge you with a crime (presumably something like obstruction, possibly specific to immigration law, otherwise right against self-incrimination might prevent a conviction on failure to disclose alone), or have other evidence of some other crime. Then you end up in the legal system, where courts handle due process and a judge, preliminarily, and then a judge or jury decides if you knew the password.

Note that the recent high-profile case of a man being jailed involved him refusing to decrypt, rather than claiming he didn't know. He was deliberately trying to test the law regarding the permissible scope of inspection of digital data, to force the matter into the courts so the issues could be litigated in a controlled context untainted by other potential crimes; being arrested and charged was part of his plan.

AFAIU, the San Francisco peninsula (San Francisco + San Mateo counties) has one of the densest networks of ALPRs in the country. You can't drive more than a mile or two in the peninsula without passing an ALPR. And my educated guess is this is a response to the rise of property crimes, including car break-ins and theft, over the past 15 years, which became a huge political issue, locally and of course nationally. Those crimes began to drop precipitously after it became evident that criminals were being tracked and apprehended by the cameras, and I doubt it's a coincidence. IOW, at least in this region the cameras exit because of crime, and they're working.

That's not a defense of these wide ALPR networks, and certainly not a defense of how they're managed. But my point is that it's a thorny political issue because in many areas they actually work, at least to some significant degree. Whether or not they're worth the privacy cost, or whether or not alternatives would work sufficiently well, are fair questions, but precisely because they're fair and contentious questions means preventing their use won't be easy politically. Many of politicians and LEO officials pay lip service to privacy concerns, but behind the scenes they're wary of doing anything that would rollback the capabilities these systems provide.

Has Apple backported anything recent to their forked gmake? Have they ever backported any features for that matter?

A scholarly and far more rigorous argument against copyright and patents can be found in Against Intellectual Monopoly by economists Michele Boldrin and David K. Levine. "We show through theory and example that intellectual monopoly is not necessary for innovation and as a practical matter is damaging to growth, prosperity and liberty."

It's a published book you can buy in hard copy (I bought the ebook on Kindle) but practicing what they preach it's freely available online, http://dklevine.com/general/intellectual/againstfinal.htm

I didn't want to guild the lily. Switzerland also has a significantly higher per capita GDP than the US. 1/4-1/3 higher.

Switzerland is just a really bad example to use because they're significantly wealthier than the US. The only other two non-city-states with a higher per capita GDP than the US are Ireland and Norway, but Ireland's is inflated because it's a tax haven (less real money available for public spending), and Norway's is inflated because of their oil and gas fields. Though, comparing the US to most any other country, especially individual European countries, is already specious. Switzerland is ~9 million, compared to ~340 million for the US. Scale matters, because of variance, selection bias, and the policy conclusions you can reasonably draw. A better apples-to-apples comparison would be between individual US states to individual European countries, or to compare the US to the EU as a whole[1]. People don't like to do that because such comparisons tend not come out the way they want. It would be interesting to read a follow-up comparing Colorado to Switzerland, not least because they're both mountainous regions. Not nearly as easy an analysis, though, because you'd have to figure out Colorado's share of Federal road infrastructure expenditures, plus Colorado-specific road data.

[1] To be fair, comparing the US to the EU can be a chore because many publicly available datasets, even from EU agencies, encompass not just the EU proper but larger geographic regions, or subregions of the EU. And of course Switzerland itself is just such an example as they're not a full member state. Like, try finding out the number of heat-related deaths. I couldn't find any numbers just for the EU proper. Some numbers included Russia, the Balkans, and/or Turkey, some were for subregions of the EU, and often it was a chore just to figure out which countries or regions were used in the report. It's a mess.

Switzerland has ~440 people per paved road kilometer. The US has about ~80 people per paved road kilometer. That means Switzerland has more than 5x the number of tax payers to support each kilometer of paved road. Maybe Switzerland was smarter than the US in terms of not building infrastructure it can't adequately support, but in any event....

Hmmmmm. It's late, but FWIW I think this comment from Martin from last year might be relevant: https://news.ycombinator.com/item?id=40838721 And I'm guessing the note that was added is fn#150 in C23. Volatile accesses aren't general memory barriers, but I think Martin's point is the potential for UB behavior means the compiler has to preserve the sequence order of the potentially UB operation relative to the volatile access.

What if x is a register to toggle divide-by-zero exceptions? Normally this is none of the compilers business, but the whole point of volatile is a mechanism to express stuff like that, so it needs minimal semantics, e.g. no time traveling (compiler barrier), to be fit for purpose.

Almost all states offer pensions, or hybrid pensions and direct contribution plans. You have to add 1/3-1/2 or more (especially for states with retirement health plans) to yearly salary to appreciate real compensation.

It's not just a burden for the immediate parties. Having the ability to dispute rights and obligation going back indeterminate amounts of time adds risk to the rights and obligations of untold numbers of third-parties; everybody's interests become more interdependent and intertwined the more time has elapsed. One of the important functions of a legal system is to settle rights and obligations. Settled, transparent rights and obligations are also integral to notions of fairness and justice, so it's not a zero sum thing that statutes of limitations sacrifice fairness for cold transactional efficiency.

Good one. Not quite the same thing, but Saint Catherine's Monastery in the Sinai Peninsula claims to hold a covenant specifically guaranteeing their protection, signed by and possibly given to them in person by Muhammad: https://en.wikipedia.org/wiki/Ashtiname_of_Muhammad That would make it over 1400 years old. The authenticity is disputed, both whether it's original and whether it named and was given to them specifically, though AFAIU it's well known that Muhammad gave out that type of thing to various groups, for example Yemeni Christians[1]. Nonetheless, it seems to have been honored, albeit sometimes half heartedly, for most of the time since. The ironic thing is the greatest threat to its continuance is the nominally secular Egyptian state, which in its secularity feels less obliged, notwithstanding the motivation to infringe upon the monastery's autonomy principally arises from the Islamic culture and pressure groups.

[1] In the Yemeni case seemingly to forestall the Aksumite/Ethiopian Empire coming to their defense, which they had done pre-Islam when, IIRC, the Christians were slaughtered by the Yemeni Jewish kingdom.

FWIW, the longest peace treaty lasted approximately 700 years, from the 7th to 14th centuries: https://en.wikipedia.org/wiki/Baqt

An analogy just occurred to me. Imagine an income tax not on what you currently earn, but on what you could earn given your schooling, credentials, experience, etc, and the available jobs within your region. Imagine how utterly unfair people would feel about that, and how quickly such a scheme would fall to widespread political pressure.

I get why an LVT makes sense from an economic perspective, but in practice it feels nearly as unfair as the above. Residents especially, but small and medium sized businesses would very quickly lobby to eviscerate an LVT. Imagine long-time residents getting kicked out not because a developer wants to make buck, but because the tax system requires that to happen. There's zero chance such a scheme could survive intact politically. In Pittsburg it's an LVT in name only, with a long list of exceptions and curious appraisals. And once it's applied ad hoc, which it must, there's all manner of distortions and incentives for corruption.

There are many ways for land use and tax policy to improve, but you have to come at from the perspective of improving the existing system, not revolutionizing it with an LVT. And you could probably get closer to an LVT by smart incremental reform than you could by trying to nominally pursue an LVT.

There was also a community in the deep south. And I once read a case study about a town that tried it in, IIRC, New Mexico.

Full LVT never works. It always ends up shot through with exceptions and loopholes and, in the case study, inducing corruption. And it's all predictable, too, if you know anything about taxation policy and American political culture. For full LVT you'd need to start from a blank slate, not just from a legal perspective but from people's expectations about property ownership.

I wouldn't call it legacy as that implies something people inherited but didn't want. The common sentiment among kernel developers has always seemed to be that device trees are nicer to work with than the complicated PC abstractions, and that sentiment remains today. And IIRC Linux didn't always have device trees; it was an approach that came later as it was ported beyond x86, Alpha, and SPARC, and was welcomed.

I'm not a kernel developer, but it seems to be one of those things that's great when you're adding support for a particular device on a particular platform, but becomes a giant PITA in the aggregate for the downstream ecosystem. It's a mismatch of incentives and preferences.

You left out the important part, that Greek influence made it all the way to Japan, in some cases identifiably intact: "[The Greek wind god] Boreas became the Japanese wind god Fujin through the Greco-Buddhist Wardo/Oado and Chinese Feng Bo/Feng Po ("Uncle Wind"; among various other names). The mother deity Hariti [Kishimojin in Japan] was inspired by Tyche."

I'm not anti EV. I would love to own an EV. I just can't justify the expense. In my case a big part of it is the cost of upgrading the electrical, otherwise I might be inclined to pay the premium.

I don't think housing costs would drop that much, if at all. What matters is the monthly mortgage payment. Property prices drop when property taxes go up, but mortgage payments tends to stay similar.

In one sense that sounds like free money for the state. However, home equity constitutes the largest share of savings for people, and that's more true the lower down the income ladder you go. So you're taking money out of the working- and middle-class. In some circles home equity is considered a poor savings vehicle, but they don't seem to consider that it's the only secured loan most people can access for leveraged investment. It's the only investment available to most people that let's them leverage capital markets the way the very wealthy can.

And FWIW, residential rents also track monthly mortgage payments, so renters aren't likely to see any difference, either. On average renters pay roughly the same amount per month they'd pay as owners, just without building any equity. I suppose they might be slightly better off after accounting for transfer payments (i.e. public services, entitlements, etc), but they'd be even better off if they could become owners rather than renters.

I once ate a dog meat stew in South Korea. It was tasty, not very different from any other red meat like beef, lamb, or horse, though the spiciness of the stew may have muted any unfamiliar tastes. However, I had to psyche myself up to eat it, and once I was full and my cultural taboos came flooding back I did start to get queasy. There's a significant mental component to what we find appetizing. Most people aren't that far removed from the stereotypical picky eater, myself included, despite having eaten balut, live shrimp, and many other "exotic" foods.

San Francisco receives over $3,000 per capita in property tax revenue alone. Dallas receives about $1,200 per capita. NYC is less than $1,000, though I gather NYC has their own property tax issues.

And California has an income tax, whereas Texas doesn't.

Prop 13 has it's problems, but I don't think they're as significant as the rhetoric claims. It makes for a great excuse, though. Politicians can blame Prop 13 for why they can't spend as much on services as people demand, and now it's become the common wisdom--but for Prop 13 California could afford to spend much more than it does, and the housing affordability crisis would end, too. It's just wishful thinking.

Yes, the difference. I'm not comparing keeping an ICE to purchasing a new EV. The premium for EVs is $5-10k or more relative to ICE models, and the $5-10k is being very charitable.

Non-peak electricity rate in some parts of California is $0.45/kWh or thereabouts, 2-3x what it is elsewhere in the country.

Depends on where you live. In states like Connecticut, or certain areas of California (PG&E Bay Area customers), the ROI is very long or simply doesn't pencil out at all. I think for Connecticut, which has high electricity prices but low gas prices, it's just net negative. For me in SF, despite high gas prices the ROI is just too long (5-10 years?), and that's before factoring in the ~$5000+ required for updating my electrical box, which is a conservative estimate because now the city requires the meter to be outside the house, whereas it's currently in the garage.

Wireguard with PQ won't be Wireguard, anymore. It'll just be a rehash of IKE+IPsec. What made Wireguard better was the very simple handshake and minimal state, but no PQ algorithms can support that simplicity because the keys are too large and/or not as simple to use as ECC.

Might as well switch to IPsec. Everything is already in place, including hardware acceleration. But most people won't, and we'll live in a world with duck-tape hacks built around a compromised Wireguard-ish layer.

IIRC, the Navy stopped teaching officers traditional ship navigation methods for awhile, until the series of incredibly stupid incidents drew attention to the lack of basic navigation and piloting skills, then curriculums were partly reverted. Not sure if things have improved much; probably too early to tell.

I doubt officers or crew would ever have to resort to a sextant, but that's besides the point. It's the type of thing that's integral to learning navigation theory and building tacit knowledge that's fundamental to the job.

EDIT: 2016 article discussing restoration of traditional navigation instruction: https://www.npr.org/2016/02/22/467210492/u-s-navy-brings-bac...

On modern systems, including Linux and OpenBSD, getentropy will return entropy seeded from not just rdrand/rdseed, but other entropy sources on the system, e.g. many Intel NICs, the AMD PSP, etc, and at least on Linux some well-vetted jitter hacks as a backstop for embedded devices without hardware RNGs. These sources are unavailable to user space. You have nothing to lose and everything to gain by using getentropy as one source, among whatever else you're using, to seed your PRNG.

At least $16 billion for 45 miles. See https://www.deltaconveyanceproject.com/about-the-delta-conve... and https://en.wikipedia.org/wiki/Delta_Conveyance_Project

> the first VPN that can’t log your activity and outsmarts internet censorship.

I guess they never heard of Zero Knowledge Systems: https://en.wikipedia.org/wiki/Zero_Knowledge_Systems

You don't know if it's bad. Microcode updates might fix it, or break it for that matter. Revision history can be difficult if not impossible to comprehensively catalog.

What it is is unreliable. And that's fine so long as you have other entropy sources. OpenBSD is really good about this. Quite a few drivers for various chipsets and cards exist just to read their RNGs, not actually use them for their primary function (which can be a bummer if you want to use the the device, get your hopes up when you see the driver exists in the tree, then discover the only capability it supports is reading the RNG). If you have a CPU with a known bad rdrand, odds are OpenBSD is still sourcing strong randomness from some other chip in your system (PSP, NIC, etc). And because feeding bad (as opposed to malicious[1]) entropy is harmless[2], they don't have to maintain a pile of conditions. Nobody is worse off, and overall everybody is better off, including having stronger getrandom/getentropy output, by not trying to be clever.

[1] https://blog.cr.yp.to/20140205-entropy.html

[2] Presuming nothing is relying on an entropy estimator. I can't remember if Linux finally moved past the entropy estimator nonsense. IIRC they did add a software jitter RNG that runs early to try to set a minimum entropy floor, regardless of hardware sources.