tybris

The difficulty of scaling is not in getting the hardware, but in the operational burden and frequency of failure. With EBS, machine and disk failures are effectively irrelevant and you can instantly migrate to another machine. You don't need to worry much about data loss. If a disk breaks in one of your machines, EBS keeps working. If the machine breaks, the data is still in EBS. If EBS breaks, the back-up is still in S3.

Dedicated hosting and rackspace cloud may be cheaper, but if you expect to be in a situation where failure becomes frequent, you'd better be on EC2.

If you expect to grow your start-up quickly, it makes sense to use a provider that allows you to scale. It is very costly to maintain two separate infrastructures for different providers.

I don't recognize your description of the political landscape. Republicans don't do any more to shrink government than Democrats do to expand it. This has never been true. They do have different priorities. Republicans think the government should be like your father who protects good citizens from harm by fighting bad citizens and foreign threats. Democrats think the government should be like your mother who nurtures and cares for all of its citizens.

As science and economics are showing us a mixture of father and mother turns out to work best, the political differences between Republicans and Democrats are shrinking every year. As a consequence, the parties feel it is necessary to distantiate themselves more and especially the Republicans have spent several decades creating a vast narrative around where their priorities lie. Unfortunately, this has led to big changes in voter's habbits and ideological perspectives.

Wealthy opportunists have helped promote and taken advantage of the narrative to push an agenda of lower taxes for the rich in return for lavish campaign funding. A group of people bought into the narrative too deeply and now think that the myths about taxes they heared over the years are the hard facts, and that Democrats have misguided beliefs. Supported by the wealthy opportunists, these people have taken control of the Republican party and its policies. This is a corruption that has taken many years to develop, but now runs deeply into the belief systems of voters and the wallets of politicians.

Amazon.de went down when the outage started, but they quickly recovered.

Both S&P and Moody's make good points, but have a different assessment of the uncertainty. Compared to other nations, The U.S. has a surprisingly healthy budget. Cut out the bush tax cuts, end the wars, stabilize health care spending, and you could create a surplus in no time. Few other countries have that luxury. However, it has not only become questionable whether Washington will be able to make the necessary changes in the foreseeable future, the corruption that underlies this inability is very hard to get rid of and may cripple the government even further.

I think these two different ratings appropriately reflect the uncertainty of the situation.

Is that true? My computer savvy friends are slow to switch from Firefox, because they've come to rely on certain plug-ins. My non-computer savvy relatives, who kept switching back to IE if I gave them Firefox, are all using Chrome.

Not sure, an open source 3D printer is already cheaper than most laser printers.

Sarcasm can work, but it needs to be based on widely accepted truths that people can identify with.

Problems with great grandparent:

-Campaign finance is heavily regulated in the UK and about half of it comes from the state.

-UK media is less of a political frenzy. They have other things to talk about, like the premier league, or human affairs.

-The closest thing to US liberals language-wise are the Liberal Democrats, who are part of government. The closest thing to US liberals politically are the Conservatives (e.g., combating climate change is high on the conservative agenda)

-Liberals aren't gaining any seats in polls

That's a few too many errors for sarcasm to work.

Their standard policy concerns the unilateral decisions. This was a mutual agreement outside of the standard policy. The benefit to the app maker is a guaranteed free-app-of-the-day promotional spot, the benefit to Amazon is not having to pay the 20%. Seems like a (potentially) fair deal.

> 'Amazon is building their market on the backs of apps by giving them away for free.

I don't get that. It's just a deal. They could have said no. The rules of the game are clear. Amazon can make your app free unilaterally, but you'll still get 20%. Or, you go into any arbitrary mutual business agreement with them to guarantee that your app gets promoted, at the cost of lower/no revenue. I found the fact that the latter exists informative, but in no way threatening.

My take from this whole thing is that in the technology industry, people are not used to making business deals and get scared when they do.

I see no reason to think that's not the case. This was a mutual agreement outside of the general terms. That's hardly unusual.

Why is it the top 10? Their profiles seem really standard for developers aged 19. Who didn't start playing around with PHP and JavaScript at age 14?

Ancient literature can predict floods, lumberjack says.

My point is, I'm skeptical of a physicists' insight into the psychology of insurgents that would give this model any legitimacy in terms of making accurate future predictions, rather than just modeling past data. The former requires a lot more faith.

That's pretty smart. It basically anonimizes individual responses, without harming the statistical outcome.

Almost makes me want to go to school again.

Deprecated in 5, 4, 3, ...

Debates among scientists sometimes look almost indistinguishable from public debates on science. The difference is in the premise: in the former case the merit of the methods is questioned, and in the latter case the merit of science itself is questioned.

My feeling was this critique was of the former kind, particularly with regards to scientists' communication. His view is fairly similar to that of most academics I know, though less nuanced.

In that light, I think you may have misinterpreted this particular comment.

The word "truth" is a diabolical invention that has a fairly subjective interpretation that changes when you start looking into what it means. However, once you dig deep enough, you find the inevitable: There is no spoon.

Empirical science is not about truths. It's about models. It's about modelling our world 'as well as we can'. The theory of relativity does not explain how the universe is architected, it's just a model that makes predictions. Any conclusions about the nature of the universe drawn from the structure of the model or any valuations of the model are entirely your own.

What does "as well as we can" mean? Here it comes down to merit. Accuracy is a virtue, but accuracy often implies complexity and limited applicability (e.g., simulations are relatively accurate, but only concern a single set of situations). Practicality is a virtue, because few people use relativity to calculate gravity. Correctness is a virtue, but probability can make things much simpler. Simplicity is a virtue, but not for any fundamental reasons. Generality is a virtue, but often comes at the cost of accuracy. Completeness is a virtue, but we do not know how to achieve it. Communicability is a virtue, because many good models are forgotten for lack of sex appeal. Usefulness is perhaps the greatest virtue, but it's always too early to judge that. As scientists, we are constrained by the same reality that all systems face: There is no right solution.

Why make models if they don't reveal any truths? Two reasons. One is that they (sometimes) allow us to calculate things, which is the most powerful concept in human civilization. It allows us to build bridges, launch satellites into orbit, produce the materials that we need, etc. without relying solely on trial-and-error. The second reason is that while the model does not reveal the truth, it does reveal a piece of the mathematics or processes that play a role in shaping the system being studied. Usually the model only reveals one aspect of the system, while in reality all things are related and influence each other and no calculable formula could describe it. However, it gives us a sense of direction, like a compass, rather than a map.

Once you grasp these subtleties of science, you become much more careful about your claims. Not so much because you think they might be flawed, that option should already be off the table, but because you know your claims do not (intend to) fairly represent the complexity of the truth. Part of the author's frustration, which many in the sciences department would share, seems to be that some scientists don't see or willfully ignore the subtlety when making claims in public. This leads people to misrepresent science as a magical truth producing machine, and make preposterous claims that distract attention from what science actually does: create models, for whatever purpose they have.

You can actually make your public folder available through CloudFront by creating a distribution and pointing it at the Dropbox web servers. Unfortunately, it's a bit of a risky business, since all the other Dropbox users will be able to use your distribution as well.

They backed down alright, by creating a competing app store and presumably a competing tablet.

Google invests in Chrome. People use Chrome to buy things from web shops. Should web shops pay Google for this privilege? Are web shops immoral? I thought I paid for an app platform. Not a secret Apple sales platform.

I can envision the board meeting now: "It's so much better for our users if we take 30% of the earnings on every sale".

No.

Is that a bad thing?

I think that's completely besides the point. Sony actually bothered to make a couple of plain-text letters look like a captcha image just for the sake of it. That's just a clear misunderstanding of what these things are for and why they are the way they are.

As to your point, Captchas have a completely legitimate use-case that you can sometimes see on Google and Amazon. If they start to have some doubts about whether you are a human being they will present you with a Captcha image to solve. Otherwise, you should be regarded innocent until proven guilty.

The software industry does not lend itself well to direct competition. One technology will usually be the winner for a particular set of use-cases and others incompatible or lagging too far behind. If you don't distinguish yourself on use-cases it's a losing battle against the establishment.

A better strategy is to find use-cases that make your competitor obsolete. Microsoft lost some of its significance when Google started producing office software for the web. Google lost some of its significance as the starting point of the web when Facebook became successful. Microsoft could make Google less relevant in search by commoditizing and commercializing Bing's search technology as part of Azure. Google is dependent on it's screen real-estate and has been reluctant to open up APIs to developers. That's where Google is weak and vulnerable, while it's one of Microsoft's core strengths.

This is called debt restructuring, not defaulting.

PDF has static contents and a static interface. CDF has static contents and a dynamic interface. HTML has dynamic contents and a dynamic interface.

This is not an HTML replacement, but a PDF replacement. HTML has never been, and probably never will be, a popular format for static documents.

I don't disapprove of free and open publication, but without publishers, who is going to organize conferences with gala luncheons?

Why not http://en.wikipedia.org/wiki/List_of_sovereign_states_by_ext... ?

Most computer scientists I know don't like computers either. What we all like is that within them lies a whole new universe with the infinite possibilities of math, yet a strong connection to the physical world. At least, until it breaks.