It’s all about platforms…

A very interesting read about life at Amazon and Google from Steve Yegge (ex Amazon, now Google)

Steve posted this on Google+ publicly when it was intended as an internal message. Steve talks about the difference between Amazon and Google and how Amazon turned their technology stack into a platform.

Here are a few links to the post, just in case any get removed – also they have different comments.

At Caplin, we are not really in the same game as Google or Amazon, but we have a platform that our customers take and build on, as opposed to an online platform that customers use.

4 thoughts on “It’s all about platforms…”

  1. It’s something that struck me around 2005 when I was first playing with Amazon Mechanical Turk (a way of giving programs API access to human workers). The level of ‘getting it’ that Amazon display when it comes to API access is astonishing. All major computing companies had beem preaching ‘utility computing’ or some variant of it since the early days of computers, but Amazon was the first company I was aware of that genuinely did it, not some smoke-and-mirrors, “pay us some money and we’ll remotely switch on another bunch of CPUs already sitting in the server we sold you” gimmick.
    It’s perhaps a good thing that Yegge has so much else negative to say about them, otherwise it’d be tempting just to anoint them the ‘Good-Guys’ in the imagined morality play of the technology business.
    Of course, the real hero of the story is not Amazon, but APIs. APIs to do anything from anywhere. It irks me how much money, for example the BBC spends on making ipad and iphone apps. None of that needs to be paid for – if they made the content suitably available by good APIs, beautiful, brilliant apps would be created for every platform under the sun just for the love of it. If the service or content your APIs give access to is compelling, you can stop right there – people will batter down your door to extend your reach to hardware you’ve never heard of with UX designs you had never imagined.

    1. Agreed, although I think the BBC example may be a little bit irrelevant. The BBC and many other content producers do not wish to make their APIs available to third parties (whether this is right or wrong is irrelevant). This post is much more about the internal structure of interoperability between the different islands of your software, content and services and the BBC may very well have this behind their walls. This post is more shows that whether or not you are going to be making the software available outside of your own organization it is important to be thinking in a SOA way as it just seems to be far more usable by everyone involved.

      1. do not wish to make their APIs available to third parties (whether this is right or wrong is irrelevant).

        It’s not irrelevant, it’s almost the entirety of the argument. Yege doesn’t think Amazon is good because they use SOA internally, he’s praising them because they have created a platform that others can build on (and SOA enabled that). You can’t do that if you don’t make your APIs available to third parties.

        Larry Tesler may have convinced Bezos that he was no Steve Jobs, but Bezos realized that he didn’t need to be a Steve Jobs in order to provide everyone with the right products: interfaces and workflows that they liked and felt at ease with. He just needed to enable third-party developers to do it, and it would happen automatically. [my emphasis]

Leave a Reply to Anonymous Cancel reply

Your e-mail address will not be published. Required fields are marked *