Other sources have it as 'increase'.

Actually, it is 'increase' already in the linked article. The quote is

And it actually makes sense.

I'm pretty jealous of American billionaires who *do* things.

Elon Musk is south african.

This is simply a stats arms race.

one that seems overheating, too. You can buy quite well speced smart phones (way better than an iPhone, as you have correctly noted) for a very decent prize. Manufacturers seem to be running out of ideas on how to get traction in this market, so this is what they come up with: over-the-top-specs.

A market full of smartphones that can't find a way to differentiate themselves from each other seems to me like a market ready for collapse.

Why do scientists falsify? Or how can they? They must know they will be found out - especially the more sensational the finding.

The answer to that is that they fool themselves. If you ever have been at a top institution of this kind you might have witnessed a certain mix of hubris, megalomania and groupthink. These people tend to be really good, but their selfconfidence, their lack of understanding of statistics, their mutual reinforcement, and the huge pressure to keep producing blockbuster research can warp their thinking. It would not surprise me that they believed the results to be true, but thought it was just the damned data that kept being wrong.

This sectlike atmosphere at some of these institutions is compounded by the fact that people there work so insanely hard that they don't have time to take a step back and think things through.

Sure, it's not going to set any records, but with 500-1000 cores and 5-10TB of RAM, it's a lot more than most users will ever see.

Can you provide any links? I'm interested. Thanks.

Many of those systems have no (or minimal) idle time.

In my experience, this is true only for the top machines in terms of reputation of the institution where they are run. Many more supercomputing facilities actually idle around a lot if not most of the time. They were bought to confer bragging rights and are embedded into a context that makes them unable to operate effectively.

A lot of these machines are hard to program for, and the institutions that own them hard to deal with (often universities with bad bureaucracy and ridiculous internal rules) which means few bother and even fewer get to run code on them. Of course that is something that is rarely admitted in public.

I think the guy did wrong and should be punished. OTOH, I think he also deserves an award for showing (again) how ridiculous the whole HPC thing actually is. Here we have these supposedly super-high-end machines (in terms of running benchmark software) which just aren't competitive by a hilariously large margin with what is out there mining bitcoins. How embarrasing.

It is coincidence. Our team idiot is the company owner.

Even after factoring in that he's most likely making more money than you?

I would say it was only amoral if exploited for one's own gain or to others' detriment.

So if a hack gives reputation to a security researcher while embarrassing the website owners - how is this not exploitation for the researchers gain to the website owners detriment? You go there and pull off an I-am-smart-and-you-are-a-moron on these folks that are trying to make a living. How is that different from being an asshole?

The argument that security researchers are actually doing good is just an unsubstantiated assumption that needs closer scrutiny, and it is quite likely not true in many situations. For example, the SCADA vulnerabilities have not led to any major or even minor problem, yet they have generated a lot of FUD and maybe even given ideas to criminals and terrorists. Researchers have gotten their nice reputation out of this, but what has the world gained? And look at how the credit card industry works. A lot of their shit is fundamentally flawed from a security point of view, yet it works and is quite convenient. How can that be?

Security researchers make a nuisance of themselves in many situations, and don't even realize it. Their "told you so" can be extremely costly to a company when there is trouble, because of how it affects liability issues. Most companies would not be viable if they had to fix every bug unearthed by researchers or face full liability claims when their unfixed code fails. The kind of talent needed to get security stuff right is just not available in the needed quantities at a reasonable price (i.e. hourly rates comparable to that of a janitor) so it is unreasonable to expect things to be secure. The alternative to insecure stuff is no stuff. Everybody who's not a propellerhead knows this.

If you're doing C++ everything by Meyers.

If you are doing C++, you absolutely must read the Frequently Questioned Answers:

It's a little (though not much) out of date, as it does not cover C++11. But the author has some comments on it, too.

Obviously I am going to be modded down, but hey. Truth is truth.

If your Fortran program ran correctly on a PC it would run correctly on a mainframe, mini, or supercomputer. More importantly, it would produce the same result. It didn't matter which compiler you used.

This isn't true. As in C, optimizations might have changed the order of fp operations resulting in subtle differences that often matter. Memory allocation (yes, the static arrays) has some really funny weirdnesses across compilers that make buggy programs produce very different results on different versions of the same compiler. The F77 language has very little support for avoiding bugs, and quite a few booby traps. Most F77 codes are just riddled with bugs and depend on undefined behavior that varies a lot from one compiler to the next.

Just look at all the code generation flags of gfortran to get an idea.

This is a very real problemâ"it's not just some rich people being assholes, but rather some rich people who stand to become substantially less rich if things go the way they seem to be going.

I thought the actual story was that if you or me dislike some policy we can go fuck ourselves, whereas if the Kochs dislike it, they get a real chance to change it.

An oligarchy indeed.

What I also find a little unsettling is that most commenters, including you, don't seem to think much of that power imbalance (or even be aware of it) directly jumping to the solar vs. no solar issue.

3) Start writing alternatives in safer languages

Like Ada.

I expect Ada to make a comeback. It is a safe language, but without the bizarrness of a functional programming language, so it can be used to do actual work.

Your law may not be properly upheld in practice but that does not change the situation of Germany being in the very small club of countries where the art of street photography is effectively illegal or at least very cumbersome.

Yes, and they keep constantly weeping about all the street photography they miss out. Well, actually, they don't. It seems they like their law like that. For some reason, people assume that street photography, or being able to shoot photos of whoever you want, is a right nobody would object to. Well, I do, especially in times of facebook, and it seems I am not alone.

Things like google glass belong, as far as I am concerned, banned, and its use in public places punishable by jail. In the same vein, cell phones should be forbidden from having a camera.

There is this thing with privacy. For some reason, everybody is for it as long as it is not them who have to respect it.

The best tools in the worst hands are far worse than the worst tools in the best hands. Yelling for tools is a specious argument. Someone has to do the work, and that someone may well bone the job.

A similar argument was put forward against the use of seat belts in cars. It just does not hold water.

The point of safer tools is to keep the reasonably good programmers from shooting themselves in the foot. Because as good as they may be, they are human and make mistakes. C needlessly invites a lot of mistakes, and even good programmers fuck up in C all the time.

Again one needs MOTIVATED AND QUALIFIED eyes AAAAAND good QA and test cases.

Perhaps using a safety aware language like Ada would be helpful too. C is known to be brittle, yet people insist in writing all sorts of mission critical code in it. I really wonder why.