greatzebu

Wouldn't that make them entirely beholden to search engines that rely on ad revenue to pay Chrome? It doesn't seem to change the underlying incentives very much.

In 2018, 53% of votes for Wisconsin State Assembly went to Democrats, and Republicans won 64% of the seats (see https://www.jsonline.com/story/news/blogs/wisconsin-voter/20...). So in this case it's not true that "if that party stays in power, it's because a majority of the people in that state have voted for that party".

Grades are also fraught with bias, and in a lot of cases measure compliance more than mastery of the material.

I don't think this is borne out by the data on manufacturing output: https://fred.stlouisfed.org/series/OUTMS

The tip is for the person coming up with the rule--even rules that seem very simple will be harder to figure out than you expect.

See "Considered Harmful Essays Considered Harmful" (http://meyerweb.com/eric/comment/chech.html)

"We should stop doing X" will always be a popular essay topic, but the "X Considered Harmful" framing is a bit stale at this point.

It never attracted a giant user base. I miss Reader (especially the pre-G+ integration version) and I wish it was still around, but it's not mysterious why it got killed. If it had hundreds of millions of users it wouldn't have died.

This is a very reasonable question. One way to get at the question of whether the reduction is real is to look at crimes where it's hard to game the statistics and see if they follow the same trends as other crimes.

One of the best crimes to look at is murder, since it's much harder to reclassify or not report a crime when there's a dead body involved. Looking at the statistics for murder and non-negligent manslaughter from http://en.wikipedia.org/wiki/Crime_in_the_United_States seems to support the idea that the reduction is real--the rate has been cut in half since 1990, and the reduction is broadly similar to the reductions in other types of crime over that period as described by that article.

I think this should be read as "kids under 6 failed [to perform better than chance] every time"

The principled way to avoid bad worst-case performance for quicksort is to select the pivot using a median of medians approach. That gives you good asymptotic performance even for adversarial inputs, but in practice it tends to be slower. Which fits well with the theme of this article.

A "Pareto optimal" solution is one which, given all the possibilities of action, produces the best outcome for both parties (with some negotiable surplus).

This isn't quite right. In a Pareto optimal solution, there's no way to improve things for one party without making them worse for another party, but there's no guarantee that the outcome is good for everyone. "I get everything, you get nothing" is a Pareto-optimal way of diving things up.

The network looks relatively anemic at first glance. I would guess that this will be a LINPACK monster, but it'll be difficult to scale applications that aren't pure linear algebra up to the point that justifies a single machine this size.

"Lede" is supposedly a holdover from the time when news was printed using actual metal lead type, and "lead" was accordingly shorthand for type itself. The alternate spelling avoids a potentially confusing homograph.

I'd say that the two communities here are just people working at opposite ends of a continuum of applications that run on large clusters. The HPC community is all the way at the end with tightly coupled applications, low data to computation ratios, and diverse communication patterns. The big data community is characterized by giant data to computation ratios, highly constrained and regular communication patterns, and loose coupling.

The fundamental problems are similar (fault tolerance, load balancing, scheduling) but the best approaches depend on where you are on that continuum.

One possible explanation: the cost of initial IO to write your data is amortized over that data's lifetime. The early-delete penalty ensures that Amazon always makes enough on your data to justify the cost of writing it into their system in the first place.

If you have a lot of heavily-read data distributed across machines, you're probably constrained by available spindles rather than available storage space. So co-locating data that will almost never be read with heavily-read data is effectively free. I would guess that Amazon stores Glacier data alongside S3 data, and the lower price reflects the fact that the limiting factor in their storage system is IO rather than capacity.

The difficulty ratings on these questions are cruel mockery.

I was trying to build WRF from source a few years ago for a project in grad school, and the bulk of the program was one giant file that crashed gfortran when you tried to compile it. So compiling with a non-Intel compiler can present some problems.

Does it? I've always thought of a mid-life crisis as an attempt to recapture the thrills of youth, whether through a fast new car or a hot young fling or whatever. If you take him at his word, he's trying to do the opposite and move beyond the problems of his youth.

One of the biggest ways I became a better programmer in 2013 was essentially the opposite of this advice from item 3: "Don’t be afraid of your code. Who cares if something gets temporarily broken while you move things around? A paralyzing fear of change is what got your project into this state to begin with."

I was formerly working in an environment where no one was really responsible for making sure that features worked consistently or that new features didn't incidentally break old features, and the result was a huge amount of frustration and wasted time. Learning to spend some time and effort up front to make sure that new features and small-scale redesigns won't break old code has really clarified my thinking about system design. Taking it for granted that no one really cares if things are broken for a bit while you refactor really presumes a lot about the problem domain you're working in.

If you want to get into proper parallel numerical algorithms, most useful programs are C, C++, or Fortran, sometimes with CUDA or OpenCL kernels. I don't think I've seen any large-scale HPC applications in Haskell, as much as like that language.

Google does everything in a very orchestrated campaign. Bash Adwords or Search here and you will see "them" show up to defend it.

I would assume this is because a lot of Google engineers read Hacker News and they want to defend their work, rather than because of an orchestrated PR campaign.

I clicked through in the hope that it would say "it barely matters, just get started and fix mistakes as you go". It started off with promise when he asserted that the people asking this question usually didn't actually care about his answer and that other factors are much more important, but then for some reason we got the long list of languages and bullet points anyway.

If you're relying on giant corporations to engage in coordinated civil disobedience on your behalf, you're in a pretty bad place.

I usually enjoy math problems that are motivated by some interesting real scenario like a man evading a lion, but in this case I don't think it works very well. You end up concluding that the man can evade the lion, but the solution only works if the man and the lion have zero volume and you don't mind getting arbitrarily close to the lion. So the math ends up being very much at odds with the physical scenario that's ostensibly motivating it. And this solution doesn't say anything about the possible existence of a solution that maintains a finite distance between the man and the lion, so you're left wondering whether or not a real solution might still exist. Still a neat technique, but the motivating scenario doesn't work for me.

The moral compass that led us to slavery, the genocide of the Native Americans, and Japanese internment camps? The idea that the USA was formerly moral but has given it up lately isn't really supported by history.

There's no bright line between food and drugs. Alcohol is a caloric substance, and also an addictive one. That's not to say that we shouldn't distinguish between potato chips and heroin, but it's not clear to me that foods are unproblematic by virtue of the fact that they're foods.

You're taking a pretty narrow view of parallel computing. For example, there are plenty of systems that use essentially actor models for large-scale parallel computation, e.g. Charm++ and HPX.

There's a sense in which this is true, but it's a really facile argument that doesn't get to the heart of the relationship between Google and the people who use it. "You're the product" implies that they're just using you, man, and that if people were more sophisticated they wouldn't put up with this treatment.

In practice I find that the most sophisticated internet users are among the heaviest users of Google products, which implies that the value proposition that Google provides is good even if you understand how their business works. The product doesn't generally extract value from the producer when it's sold to the consumer, but I get an enormous amount of value from using Google.

It seems more likely that the reporter made a mistake, and there are actually a quintillion possible internal states generated from a relatively small number of variables, not a quintillion variables.