This article is full of horrible bits, and that quote is just the bit that stands out the most.
The link from that quote is to Quora, with only one random guy holding forth with an opinion about Friendster. Not exactly authoritative. But even if it were, the Quora answer is not even close to implying that AJAX had a consequence on server load (it suggests that AJAX makes for better UX).
The link about "friendster is too slow" links to... a story about a change of CEO at Friendster.... with no reference to website performance at all? Wtf?
The "$15 a month" link and the "so easy" link, while they do mention AWS/EC2 IASS, are only peripherally connected to it. They're about caching Wordpress, i.e. automagically converting your dynamic-content webapp into a static-content website, and have nothing to do with elastic scaling or any of the other really interesting aspects of AWS/EC2.
It's weird, because it's like there was a vaguely reasonable (but fluff-weight) article here, that someone pissed all over and ruined.
I generally enjoy the pieces I read at the Atlantic... now I'm having that experience of wondering if any of them had any truth at all to them.
> IaaS cloud computing is not always cheaper than owning your own hardware
Oh really, not always cheaper ? For the same computing power, I observe almost x10 pricing on amazon compared to dedicated server. And still, I compare to rented dedicated servers, not owned (never worked with those, I suppose it's even cheaper).
Selling (because that's what this article is trying to do) aws as a highly scalable platform where one could add resources in a matter of seconds is ok. Implying it can be cheaper than other alternatives is either lying or having no clue.
As someone who boths owns and rents physical servers, I can tell you that owning is 1/10th the cost of renting, but that the rental price includes remote hands and other overhead, so call it 1/5th the price.
And remember, everything else you read in The Atlantic (and other publications) includes similar whoppers... just on topics where you won't recognize them.
It's very hard for media - even specialist media - to be as well versed in the technical knowledge of the practitioner. You see it all the time in financial journalism as well where people through around large numbers that are grossly distorted from reality. If the journalists were as good at the technical aspects of what they covered as the practitioners, they would be making a lot more money as practitioners (see for example Michael Arrington, or Michael Moritz before him). Celebrity journalists such as Michael Lewis or Tom Wolfe might be excepted from this rule, although even Andrew Ross Sorkin makes a lot of financial errors in his writing. Dan Primack is one of the few financial journalists who puts so much energy into understanding what he covers to get it right, yet loves the beat so much that he doesn't seem to be interested in becoming a money manager any time soon. I've digressed a bit towards financial journalism, but tech journalism and other journalistic areas in which technical knowledge is required suffer from similar problems.
Either they can do it or they shouldn't do it at all. It makes little sense to read an article that you know from the beginning is wrong because the journalist couldn't understand what he's writing about. Bad information is worse than no information.
That's an unrealistic standard: journalists are very rarely experts in what they report, but their flawed accounts can still be useful. The reader just has to be aware and avoid overconfidence in the particular details, when those are filtered through flawed, erroneous, and oversimplified sources.
You can still extract some signal from a noisy or biased channel: bad information is better than no information, if you have some inkling of how it's bad.
Is this seriously an article titled "An Amazon Engineer Had a Little Idea That Turned Into a Billion-Dollar Business" that doesn't actually name said engineer?
I will never understand the instinct for writers to attribute everything to a company's CEO. (The article fronts a big picture of Bezos). Nor will I understand certain CEO's not taking the effort to make sure their people get credit. It's possible The Atlantic made no effort to reach out to Amazon on this. But that seems unlikely.
This article is terrible. What is this doing on the frontpage? It's full of technical errors, and the title is pure linkbait. I was expecting the inside story of the birth of EC2, instead of this, the "Amazon engineer" is only mentioned in the title.
I was hoping they will reveal the name of the employee who came up with the idea and give him the credit he deserves. That's what made me click on the link at least.
I spotted the atlantic in my web analytics this morning, turns out their linking to a blog post of mine inside the article.
It seems a shame that an otherwise decent article is spoilt by technical mistakes, when it's incredibly easy to contact almost anyone who they've linked to and ask "We're going to publish this, does it sound reasonably accurate to you?".
Still, at least they don't perpetuate the myth that EC2 was initially selling spare capacity from the main amazon site.
I still fail to understand why AWS is growing. It is slow and expensive. It may be useful, and a lot cheaper if you could plan the AWS with spot price, long term rental, and generally play the AWS game.
[1]http://en.wikipedia.org/wiki/Chris_Pinkham
[2]http://en.wikipedia.org/wiki/Willem_Van_Biljon