bzbarsky

In your analogy, where the grid corresponds to roads and railways, the equivalent of the supermarket or farm is a wholesale electricity provider or power plant, right?

In Massachusetts you can absolutely comparison shop for specific power plants or sets thereof as an individual, at least to the extent of explicitly picking which generation company your money goes to by default. Of course in the end the way the grid works these companies end up effectively buying electricity from each other as needed... So this analogy only goes so far; supermarkets do not usually buy food from each other, I'd think.

Not all retail customers have fixed rates.

Some project managers use Jira. Developers use Bugzilla.

> "I think ..." already started subjectively

But fundamentally me saying "I think the sky is blue" and "the sky is blue" are the same thing: the latter just has the "I think" implicit but clear to anyone who stops to think for a moment about what it means that _I_ am saying it. Maybe most people don't think, of course.

> But courts decide who is a criminal or not.

Sort of. The courts decided OJ Simpson is not a criminal by the legal definition. The descriptive statement "OJ Simpson committed a crime" has fairly contested truth value.

> hopefully in an unbiased way.

I guess the key is that the US approach starts out by admitting that the "unbiased way" is an unattainable fantasy and then works on routing around that problem. The real discussion here is whether the cure is worse than the disease in this case.

Not everyone agrees that disinformation should be curbed. And the reason many people, including the authors of the US Constitution, don't agree, is that they feel the risk of false positives is far too high.

That is _because_ I have no confidence that anyone doing the curbing, including myself, will correctly identify "misinformation", I don't want such curbing to be happening, period.

First of all, "insults" are not excluded from free speech protections in the US, except in very narrow circumstances.

Second, the distinction between normative and descriptive statements, as you define them, is not always clear-cut. As a simple example, is which bucket is "I think these doors are the same color" when said about a red and green door? What about (to use your example, but with a slight tweak), "I think all <insert group> are criminals"? Note that the "I think" was implicit in your phrasing, though maybe the implicit vs explicit distinction matters.

Third, even for statements everyone agrees are descriptive there may be widespread disagreement as to whether they are true. Example: "Donald Trump is a criminal." The US approach to this is to generally try to avoid having the government be the arbiter of truth, with some narrow exceptions for libel and the like. The German approach, as you note is different, not least due to different historical experiences. Which approach is better depends a _lot_ on circumstances and culture and norms and so forth. The German one places a lot more trust in the government not abusing it's truth-determination power than the US one does.

If they were going away for real across all browsers, yes.

Historically getting some browsers on board with that program has been very difficult.

Concretely: a large fraction of website developers would much rather put up "only works in Chrome" notices than fix their crap.

[Disclaimer: I used to work at Mozilla, and have done my share of trying to push for turning off third-party cookies.]

I have third-party cookies disabled, and have for years. A non-exhaustive list of sites where I have login or other problems as a result:

1) One of my local banks (who use weird third-party hosted modules for some of their functionality).

2) Verizon.

3) T-Mobile

If I were a normal user, any one of these ("I have to do _what_ to see my FIOS bill?") could be a show-stopper.

Which is what makes it hard to turn this on by default without driving away users.

> But the fact still remains that most Americans are living paycheck to paycheck.

I am willing to believe that a large fraction are, but it is not 70% by any means and given all the bad numbers I have seen floating around I can't even tell what that number is. I would be shocked if it were below 10% and equally shocked if it were over 40%. If you have a good source of data, I would love to see it

> See the graph:

Just to be clear, that chart is the "other ways" chart. And people can select more than one item. The numbers being shown add up to way over 100%, if you look at the chart you linked, even though it excludes some options (options (a) and (c) from the actual question, which correspond to "I just pay it, what's the issue") that between them were selected by some fraction of people between 29.5% and 59%, depending on how many people selected both of those optons. Hard to tell what's really going on given this survey setup.

So as a simple example someone can check both "With the money currently in my checking/savings account or with cash" and "By borrowing from a friend or family member". Would you categorize them as being able to pay the $400 or not?

> only 19% said they could pay for it by "selling something"

Only 19% said they would sell something as an option to pay it.

I don't know about you, but if I had an unexpected cash flow issue I would probably be far more likely to borrow from friends/family as a tide-over than to "sell something". Depends on amounts, obviously....

> IRAs or 401ks are not liquid

I agree. My main point with those was that "savings" can mean many different things, and people are really bad about differentiating what they mean by it.

The only thing we can conclude from the bank survey is that somewhere around 30% of those surveyed _did_ have > $1k in a savings account. Which says nothing about how they'd pay an unexpected $400 expense, by the way...

> So I would say that half of Americans not having the savings to pay a $500 expense is accurate

I don't see how you can conclude that from the presented survey data. It was about a $400 expense (minor nit), and it's very hard to determine what people _can_ do vs what they _will_ do, and the latter is what the survey asked. And, again, allowed selecting all the things they might end up doing.

In the extreme, I know people who _could_ buy a house for cash but what they _do_ is take out a mortgage, for various reasons, including liquidity considerations, opportunity cost, etc.

> For example age 35-44, median net worth is $91,300

True.

> and the bottom quintile is net worth negative

Also true.

> I think it's easy to imagine how someone growing up in such a household would have a different experience than someone in the top quintile

Here we have an implication that may or may not hold. I know a number of people in their mid-to-late 30s and with negative net worth. All of them are doctors. Their kids are not obviously worse off than the kids of software engineers the same age (who generally have quite positive net worth).

That is to say, net worth numbers do not correlate straightforwardly with standards of living. Nor do income numbers necessarily.

There are things that seem like they should, like disposable income, but even there it's hard to tell apart someone who just has less income to work with and someone who prioritizes nicer house vs disposable income differently.

None of which takes away from the large number of people who really _are_ living paycheck to paycheck, or the children whose parents really _do_ struggle to make ends meet. And there really _are_ a lot of people who both have large student loans and have no idea how they will pay them off. And there are people with large incomes who still manage to live paycheck to paycheck, for various reasons, and obviously plenty of people with low incomes who live paycheck to paycheck. But as I keep repeating, all the data I've managed to find on this has been nearly useless because it equates things that are not equivalent (e.g. negative net worth with a low standard of living, or low income in your early 20s while lumping together students and non-students, etc, etc). This allows people to just read their preexisting biases into the data and come to widely divergent conclusions based on the same exact numbers.

Again, I would welcome any pointers to better data here.

That third link points to a survey by a bank that had 69% of respondents saying they had less than $1000 "in a savings account" in 2019. That's not the same thing as "in savings". "In savings" would include things like bonds, CDs, high-yield checking accounts, 401(k)s, IRAS, etc, etc, etc. Even if we exclude random equities in a brokerage account.

The proliferation of high-yield checking accounts in particular has significantly reduced the use of savings accounts for a lot of people who have quite a lot of "savings".

I keep seeing people cite scary-looking numbers like this (the other good example being the "half of Americans don't have $400 in cash on hand", which is ... not quite true; see https://www.politifact.com/factchecks/2019/apr/19/kamala-har... ), but each time I dig into the actual data it turns out the actual question asked was not what people seem to think it was and the answers don't mean what they seem to mean, in the context of the actual question.

I am very aware that Rhode Island has significant numbers of people in poverty, especially around Providence. I don't think it's anywhere close to DC's situation, though I could be wrong, of course.

And just to be clear, right behind DC and Rhode Island we have Alabama (OK, maybe poverty). Then just a few % above them New Hampshire, Kansas, Alaska, Pennsylvania.

These are all quite different from each other on various socio-economic and geographical axes.

And of course the states with the most successful vaccination campaigns in the US (with ~98% of vaccine supply used) are West Virginia, North Dakota, and New Mexico. Which are also quite different from each other in various socio-economic ways.

So it's hard to blame the socio-economic angle for the state-by-state disparities here, though I would be quite willing to look at data showing otherwise.

Unfortunately, the article doesn't make clear which exact vaccines it's discussing.

So it's entirely possible that the specific vaccines approved in Europe are supply-constrained, while other vaccines are in the situation described in the article. All that means is that saying that "the problem problem is production" is very specific-vaccine and specific-regulatory-unit dependent and needs to be evaluated carefully.

> The bottleneck is vaccine production

Given claims like "I have 70m doses that I can’t ship because they have been purchased but not approved. They have a shelf life of six months; these expire in April." from https://www.theguardian.com/global-development/2021/feb/14/w... it seems like claiming that the bottleneck is production needs some serious caveats....

> Isn't the vaccination distribution a zero sum game between nation states?

It's complicated. Right now there are vaccines that have been purchased and produced but not delivered due to regulatory approval issues (70 million from just one manufacturer, according to https://www.theguardian.com/global-development/2021/feb/14/w... ), vaccines being delivered but not used (see https://www.bloomberg.com/graphics/covid-vaccine-tracker-glo... "Supply used" statistics), and so forth.

Blaming supply constraints is a great cover for these organizational and regulatory failings, though. "We don't have enough vaccine" sure sounds better than "we've only bothered to use 60% of the vaccine we got our hands on" (hello, Rhode Island and D.C.).

The US, which is what I have seen data for, is nowhere close to 100%. It's at ~75% per https://www.bloomberg.com/graphics/covid-vaccine-tracker-glo...

Now some states _are_ close to 100%. Some are close to 60%. Investigating the differences in approach that lead to those differences in numbers mostly comes down to whether states prioritize vaccinating people or creating complicated rules and sticking to them.

When combined with things like this from https://www.theguardian.com/global-development/2021/feb/14/w... : "Instead we have a patchwork of approvals and I have 70m doses that I can’t ship because they have been purchased but not approved. They have a shelf life of six months; these expire in April.", it sure seems like a large part of the apparent supply bottleneck is self-inflicted. Some basic "if it's approved in another G-8 country (or whatever criterion you want to pick) we should auto-approve it here" logic would probably significantly help with this sort of thing....

While there was some pre-production, I disagree that vaccine supply is the only binding constraint we have going on. And the lack of sane regulatory approvals and sane distribution setups has been heartrending.

If you look at https://www.bloomberg.com/graphics/covid-vaccine-tracker-glo... and scroll down to "U.S. Vaccination Campaign" you will see that in the US 74.5% of the vaccine doses that have been delivered have actually been used.

This could happen for a combination of two reasons, as far as I can see: either delivery is ramping up so fast that most of the delivered doses were delivered very recently, or because people are doing a terrible job of using doses that get delivered.

My impression is that the latter is a larger factor than the former, so far, based on the wide variation by state in the "Supply used" column. West Virginia has used ~96% of their supply. D.C. and Rhode Island have used only 60% (and it's not like their populations are larger or more spread out than West Virginia's!). And I should note that I can imagine various poverty-and-social-whatever issues D.C. might be facing, but they don't obviously apply to Rhode Island in the same way.

Things like http://www.arnoldkling.com/blog/virus-update-3/ item 3 are pretty common: crazy vaccine-distribution setups meaning people who want vaccines can't get signed up to get them, while vaccines are getting thrown out due to spoilage.

Things like https://www.nytimes.com/2021/02/10/us/houston-doctor-fired-c... are, I hope, not common, but you can bet that case will have a chilling effect on efforts by doctors to avoid wasting vaccines.

https://www.theguardian.com/global-development/2021/feb/14/w... includes this incredible from a major vaccine manufacturer: "Instead we have a patchwork of approvals and I have 70m doses that I can’t ship because they have been purchased but not approved. They have a shelf life of six months; these expire in April."

Does that sound like a manufacturing supply constraint to you?

Similar is https://marginalrevolution.com/marginalrevolution/2021/01/ap... where we have a vaccine, and a factory that can produce it, and the vaccine is approved in the UK, but it's not approved in the US. This is also not a supply constraint.

Yes, if your app is a high performance server app, measure in that environment.

But user-facing apps (the sort people run on laptops, say) have async I/O as table stakes, really. It's not even about throughput or CPU cycles: it's about the fact that if you have I/O latency on any thread the user interacts with the user experience will be terrible.

Now in practice maybe that means "just make the I/O async, but the performance details of that don't really matter too much".

Anyway, the overall comment was about performance profiling in general, not just async I/O.

Yes, you have to be very careful with measurements, I agree.

> If you speed things up by 10% on your server, they'll get 10% faster on your laptop as well.

Depends on the speedup and techniques to achieve it. For example, speeding things up via more parallelism can lead to wall-clock improvements on servers but not laptops, precisely because the latter just end up doing more thermal throttling....

Ideally, you want to measure both ideal hardware and actual-user-hardware; often speedups on one will not be visible on the other and vice versa.

Are your users going to be running your application on laptops? Will they have the same "conserve power by limiting performance" going on? If so, that is _exactly_ the environment you want to do performance work in, generally speaking.

That's all fair, and your hypothesis is pretty much spot-on in terms of what happened. The huge productivity boost was basically key to the success of the project.

I'm not sure whether the argument now comes down to "it would have been possible to gain that sort of productivity boost in C++ too, with the right design" or "it would have been possible to implement a new CSS engine without this sort of productivity boost". If it's the former, then you're right that one can't prove a negative. If it's the latter, then I think that's where the "schedule and manpower constraints" come in.

Mobile devices in the field typically have _more_ cores that desktop devices. The Firefox hardware report is strongly influenced by there being a lot more desktop than mobile Firefox users.

As you note, the modal number of cores on mobile has been 4-6 for a while now.

In the limit, you could implement a DSL in C++ that would allow you to implement CSS.

Or you could implement Rust in C++.

As you say, you can do whatever you want in C++. The only question is how much time and effort it would take and whether it's worth it.

I will note that the people who worked on the Firefox style system had plenty of experience working on complex, high-profile C++ projects. Every single one of them that I've talked to agreed that for the specific thing they were doing here Rust allowed much faster development of code with fewer bugs (especially as measured by how many issues the fuzzers found) than their past experience with C++ had been.

If your point is that the claim should be "parallelizing the style system in C++ was impossible within the schedule and manpower constraints Firefox was operating under" rather than the simpler "was impossible" claim, then sure, as a purely logical-statement matter. But in practice there are always schedule and manpower constraints.

Can I prove mathematically that some other team would not have been able to achieve the same results in the same amount of time with C++? No, I can't; such a proof would be quite difficult to construct. I do have the empirical observation that there used to be at least four fairly different modern non-parallel CSS implementations out there written in C++ (Gecko, WebKit, Blink, Edge), and that one of them has disappeared (Edge), one was parallelized in Rust, and the remaining two remain in C++ and non-parallel, even though people generally agree that parallelizing them would be good and so forth.

Maybe it's just that none of the organizations involved are any good at writing C++ code. Maybe it's that doing this in C++ is hard enough that it's not worth the effort. Maybe it's something else. What is your hypothesis on the matter?

Well, right, that's why the Firefox CSS implementation actually aims to be parallelized and has been for years. I'm just saying that the wins from that on current hardware are not as large as one would hope. And that the increase in number of cores has not been increasing nearly as fast as one would like, in devices that most people actually own.

Fwiw, I've seen a number of cases of the other way around: the WebKit/Chrome CSS implementation had some shortcuts for performance reasons that caused them to not follow the spec, where Firefox did.

The link you posted just indicates that Firefox rendered the page faster than Chrome; did it render it _wrong_?

One other note: the linked issue here is layout, not the CSS system per se; this part is not parallelized in Gecko and is still in C++, not in Rust.

For that last one: it's much faster in many cases, because Chrome's engine is not parallelized.

Back when I was last measuring this, stylo's single-thread performance was comparable to Chrome's or a bit faster in some cases. Parallelized performance was much better.

That said, actual hardware in the wild has a surprisingly low level of hardware parallelism in practice. https://data.firefox.com/dashboard/hardware shows that for the Firefox user base as of end of Jan 2021 54% of users had 2 cores and 34% had 4 cores.

When the time cost of getting it right and _keeping_ it right exceeds the amount of time actually available in a day, the result is equivalent to "the language is not up to the task".

I mean, you can implement anything you want in assembly, in theory. In practice, the cost of iteration tends to be high enough that implementing things with somewhat rapidly changing requirements (which includes CSS, to be clear!) pretty much requires a language in which iteration and refactoring can be done more quickly than in assembly.

Maybe the confusion on your part is that you think "CSS" is a static set of requirements, so if you put in enough effort to implement it once you're then done? Unfortunately, that's not how it works in practice.

> wood stove for water heater

Even ignoring the CO2 aspect, the particulate emissions from this would be quite problematic in any sort of dense area. Not to mention what fraction of those particulates would end up inside the house...

Also, you don't mention heating (houses, not water). That is a huge part of energy consumption in many parts of the world.

I would much rather not normalize naming and shaming and bullying, personally. Can't tell how much of your comment is irony, of course. :)

To be clear, journalists approximately never get to pick the title. Editors do that. Who should also have integrity, of course, but the clickbait incentives are very strong...

Or 30% truth and 70% confirmation bias, as the case may be....