Panasonic also made computers in 1980 and earlier, and still have a good line of PC notebooks today.

US does exactly the same towards foreign car manufacturers.

So, truly unsupervised algorithms cannot do useful recognition - that is, classify objects the same way people do.

That's an overly narrow defintion of "useful recognition".

Same thing with deer, moose, pheasant, duck, most fish, octopus, squid ... And they all go by the hundreds of thousands or millions, not a couple of hundred. Might want to begin by outlawing hunting and fishing in general then.

BTW, bottlenose dolphins are not endangered. They apparently aren't all that good eating either. Without all the noise, the attention and the furor the tradition would probably be well on its way to disappear altogether. Now it looks set to continue for decades to come or longer.

It's not effectively free,

Which is why I did not say "it is free". It is "free" as in "we pay nothing extra for attending, and get no tax money back if we do not". Which is "effectively free" from the point of view of the people attending university.

Only in the US (temple of individuality) does the state subsidize your studies (if you're fortunate enough) but then you are not compelled to give back. Individuality taken to the extreme, and then we ask ourselves why all the worlds big problems stem from that country.

Sweden doesn't require it either. University is effectively free, and you get a part-stipend, part-loan for your living expenses. If you go abroad to study the stipend and loan will follow. You do need to pay back the loan (in proportion to your income), but other than that there's no strings attached, even if you decide to move abroad right after graduation.

Which really makes sense. Most people that move abroad end up returning at some point - bringing a valuable load of work experience, skills and contacts back to Sweden. And the people that stay abroad become contacts for people and businesses back in Sweden. Having people leave is overall a large net win for the country even if some end up never returning.

Artists presumably know record companies screw them over already. It's not as if it's been a big secret for the paf|st fifty years after all. But nowadays they do have much more choice on who they do business with - they can elect to sign with independents, join music collectives or go it alone for instance. If they're still working with the big labels it must be because they, after all, still provide a service that is worth it for the artists.

But there's more than one 20% feature out there, and any one user depends on multiple features. It's easy to imagine where just about every single user depends on at least one 20% feature. Remove them all and you have no users left.

You can even imagine a system where almost every single feature is used by only 20% of the user base. Remove them and you've removed your system.

Apple is to Betamax as Android is to VHS.

There's no porn on Apple devices? You have to flip iPhones over to access the second half of the memory? Come on, let's stretch the analogy a little :)

Yes, we all know the definition. But knowing the definition doesn't mean you really understand the implications. My experience (I am a researcher) fits with his description: most researchers effectively think of SD as a measure of average deviation, and treat it that way when they (informally) reason about their data. As another post mentioned, even statistics teachers sometimes describe it that way for non-statistician students.

And yes, one approach would be to make people learn statistics at a much deeper level. In fact, that would be a really good idea in general. But another, complementary approach would be to simply use the mean (or median) deviation when that's what you really try to use anyhow.

In the situations this comes up, people aren't actually using the SD to derive anything else; it's simply used as a concise description of the underlying data. And the benefit of using mean or median deviation is that it's valid and sensible even when the data is not Gaussian; another common form of misuse of the SD.

What he is saying is not that statisticians should stop using SD in statistical theory or anything. What he's saying is that non-statisticians should stop using SD as a measure of variability when describing their data to each other. And since everybody (except statisticians) think SD is the average deviation from the mean, then people should perhaps use that instead, and reduce confusion for everyone.

I think that's the point, really: to make it taste good you need to cure and smoke it to the point where you could have used cardboard dipped in lard without noticing much of a difference.

Humans are the best long distance runners on the planet, and we evolved that way so that we could chase our prey until they died of exhaustion.

You wouldn't believe the stamina of an onion on the chase. No wonder our forefathers could run so well.

As other people point out, wasd and esdf are the exact same layout, but wasd are right next to the tab, caps and shift keys which are often used as well. And without any extra keys in between it's easier to hit them accurately. Also, as pointed out, in two-player games it'd give you more room for two players to use the same keyboard.

Me, I always preferred mapping my Wico joystick to whatever buttons I needed in any game. Never had any input device work as well as one of those, and I used one of their trackballs for years instead of a mouse.

"New treatment kills some but not all metastatic cancer cells in mice, but only while they're traversing the bloodstream and so far only when the cells are injected into the mice in the first place".