I think so, too, and I don't think it's a problem. Rather than just telling people they're talking out of their ass, why don't you explain what value is gained by watching it? Obviously there's no factual information in the video that can't be expressed in a few sentences of text, so the only think I can suppose is that you're of the opinion that the greater emotional impact of seeing it has value.

What, precisely, is that value? For me, personally, I can't imagine what it would be. I don't think anything could make me more strongly opposed to the act of beheading an innocent journalist. Seeing it would make that opposition more visceral -- perhaps in an almost literal sense -- but it wouldn't increase my opposition. It wouldn't lower my opinion of the terrorists, either, since it's not possible to hold a lower opinion of them than I do.

So what is the value of seeing it?

While it would be entertaining, I don't think that's a very useful method for evaluating the performance of self-driving cars, unless you're trying to design a car for demolition derby competitions. I understand that your'e trying to design an extreme environment on the theory that if the car can perform well there, it'll definitely do fine on real roads, but I don't think that theory is valid. In real life, the vehicles on the road try not to hit one another, and the method they use (in most countries, at least) isn't hyper-alertness and evasion skills, but rather cooperative rule-following.

We avoid accidents by collaborating on a set of rules, some written and enforced by police officers, most not, that tell us all how the other drivers are going to behave in a given situation. That is the context in which self-driving vehicles need to operate, at least until we eliminate the human drivers from the road -- at that point self-driving vehicles can use their high-speed wireless communication channels to collaborate more directly. Of course, while human drivers are on the road, we (human and machines alike) have to be wary of drivers who don't behave in the expected way, so there is some value in being able to avoid bad or aggressive behavior. But I don't think optimizing for that is likely to be the most effective solution.

The Google team recognizes this and is optimizing for proper cooperative behavior, and even behavior that optimizes for the comfort of passengers, as in the example in the summary.

Which is why you should do both. A simulation can test millions of permutations -- including arbitrary combinations of events, and in far more variety than could be tested in a reasonable amount of time on real roads -- and can verify that software changes don't introduce regressions. Real-world testing introduces an element of randomness which provides additional insights for the simulation test cases.

Ultimately, governments should probably develop their own simulators which run the autonomous car through a large battery of scenarios, including scenarios which include disabling some of the car's sensors. Then autonomous vehicles from different manufacturers could be validated on a standard test suite before being allowed on the roads, and when real-world incidents occur in which an automated car makes a bad decision, those incidents can and should be replicated in the simulator and all certified vehicles tested. They should also do real-world testing, but I suspect that in the long run simulations will provide much greater confidence.

This is among the worst advice for programmers I've ever read. And it's pointless advice because it's where the majority of programmers already are.

Oh, I certainly agree that clever code is a bad idea, but you should never stop thinking creatively about how to make your code better. Focus it on finding ways to structure your code that are elegantly simple and obvious, on finding the perfect name for that variable, function or class, one that precisely captures the meaning and intent -- and if there is no such perfect name, focus it on finding ways to refactor your code so that there is a perfect name. Programming -- done right -- is an inherently creative task, and the scope for beneficial creativity is vast.

This even applies at the micro level. It's almost always the case that any handful of lines of code that contains branching logic can be structured in several different ways. Take the time and try each of them! See which is most concise, which is most readable, which highlights one aspect of the logic flow or another... and then spend some time deciding which aspect will be most important for the next programmer to read it. Think about how you can write code a little bit differently to eliminate -- and visibly eliminate -- important classes of functional or security bugs.

One of the more important insights I received, after nearly 20 years as a professional programmer, was that comments are evil. Comments are a hack to work around the failure to write code which is sufficiently clear and expressive (note that I'm talking about inline comments, not comments used to generate documentation). When I find myself typing a comment, I step back and look for ways to improve naming, or refactor, until the comment is no longer necessary.

Those are just a few examples, there are many more. Programming, like any art, is a never-ending opportunity for learning and improvement, because perfection is unachievable. Doesn't mean you shouldn't try, though. I can already hear the complaints "But I don't have time for that crap, I have deadlines, and..." that's just another set of constraints to be optimized. When time is tight, I focus on simplifying and making absolutely sure that my code is bug-free and has thorough automated tests, because there isn't any time for extended debugging.

Never, ever settle for mediocrity. One of my proudest days was when another programmer whose skills and code I highly respect called my code the cleanest and clearest he's ever read. I strive to impress my colleagues (and I work with some of the best) with clarity, simplicity and elegance. Sometimes I succeed, mostly I fail... but I always learn in the process. After 25 years, I think I'm learning more every day now than I did when I started. The lessons are more subtle and far less obvious, but I think they're more valuable.

Detroit got fat and lazy, and as a result foreign automakers ate their lunch. Japan in particular had cheaper, harder-working workers, coupled with more focus on efficiency and -- eventually, after they built enough capital and experience building cheap crap cars -- design and build quality. Detroit didn't believe they could lose, either the management, or the unions. In order to stay competitive, both would have had to make serious changes... almost certainly including some reductions in labor costs and some labor re-training.

Yep. And, frankly, it was and is obvious that it would be. I've been saying for years that globalism was ultimately a good thing, though in the short term it was going to be painful for the wealthy countries, as standards of living equalize. If this article is correct, the pain may be much less, and much shorter, than I'd expected. Not that there isn't still pain ahead, but if we're already getting to the point where overseas labor costs have risen enough to be offset by domestic education and infrastructure, then the future looks pretty good.

At the end of the day, though, I'm no more entitled to my job than some programmer in China. If he can do the job as well and will do it for less money, then he should have it. Cost of living differences make this painful in the short term, but if we just keep competition open, the field will level -- some of that leveling may come from decreases in my standard of living but most of it will come from increases in his. That's too bad for me, but great for him, and it's fair because he's no less a human being than I am.

Interesting. What do you mean by "operate the engine in private"? Who would use it? And given that information derived from what you search for is the primary source of information for ad targeting, and given that the search engine is the primary place the ads are displayed, how would that work?

You missed the point. You can make such a license if you like, indeed many people have made them. But it is a popularity contest in that unless a significant number of people agree with your priorities and therefore choose to adopt your license, you won't have accomplished anything.

And, of course, the GP disagrees with your priorities and wouldn't use your license. I see both sides, but I think I'd probably shy away from a license with such vague and potentially far-reaching restrictions.

Sure. What we can't have is the power of the sharp kitchen knife, plus the compatibility with existing code and libraries without the four years of training.

I can teach a novice to use a nice, pleasant, safe and very powerful subset of modern C++ in a fairly short period of time... as long as the novice is only working on code written in that subset. If the novice starts looking at and modifying other code, though, all bets are off until he's done his years on the mountain top.

The way I see it, C++14 is a very nice language with a bunch of baggage you should just ignore... except when you have to use because you're working with code that already does. This means given a clean, modern codebase you should be able to hire a bunch of smart novices and get them productive fairly quickly. Just keep an old salt around who can answer their questions when they step outside of the nice subset.

Non-sequiteur. X has nothing to do with the GUI, at least not any part of the GUI users care about. X is merely the tool used to draw stuff on the screen; it says nothing about what gets drawn. Everything users care about, including what windows, buttons, fonts, etc., look like, how applications interact with one another, and whether or not all of the above is nicely integrated and looks like it belongs together has nothing to do with X.

I insist on living and working in locations where I don't need an elevator... a remote tropical island would work well for this.

US courts have ruled that linking to infringing content constitutes contributory infringement. I'm not sure how "direct" the link has to be. Google search links to torrent sites which contain pointers to infringing content, for example, so that's apparently okay.

Courts in the US have held that linking directly to infringing content constitutes contributory infringement. Linking to another site isn't infringement just because the other site doesn't want you to link and benefit from their material (Tickemaster v Tickets.com established that), but linking to infringing material on another site does.

(Disclaimer: I am not a lawyer nor am I a Google spokesperson.)

Which would be relevant only if Larry, Sergey and Eric decided to allow it to be. As long as the three of them stay united, they outvote the rest of the shareholders combined.

Aside from the rather questionable assertion that Google is a monopoly, the company's cash reserves are nowhere near that large, or, rather, the company's expenses are much larger than you believe. Last I heard, Google has cash reserves of ~$60B (which, note, aren't actually cash; you don't leave that much capital sitting idle), and annual operational costs of about $40B. How long Google could continue to operate with hugely decreased revenues depends on just how far the revenues declined, and how much economizing the company could do, but I strongly doubt that it would be "decades". If all advertising revenue derived from the search engine disappeared and Google didn't economize at all, it would be bankrupt in maybe three years.

(Disclaimer: I work for Google, but I don't speak for Google. Everything in this post is derived from public information.)