Pmarca: The three kinds of platform you meet on the Internet
blog.pmarca.com
blog.pmarca.com
There's a great discussion of it brewing at http://avc.blogs.com
That's only true for Ning and Amazon's web services, not the other examples he gave. The whole point of Ning is to have custom social networks, so if you took the "custom" part away, then yeah, people would stop using it. Amazon's web services are just platforms, so if you took them away... they wouldn't have a product anymore.
For products that aren't solely platforms, you can take away the platform and still have something people want to use, even if it's a Level 3 platform. All it would take for Facebook's platform to qualify is for them to run third party code on their servers. Even with that, they could decide that the whole platform thing was a bad idea, and you'd still have a useful site. I haven't used Salesforce, but I'm assuming they had plenty of core functionality before they opened up their platform, and that would still be usable. Second Life could just hire a lot of people to build neat stuff for their world and close off the platform to outsiders. As long as your entire product isn't a platform, you don't have to support the platform forever. People might hate you for getting rid of it, but it's still doable.
"I believe that in the long run, all credible large-scale Internet companies will provide Level 3 platforms. Those that don't won't be competitive with those that do, because those that do will give their users the ability to so easily customize and program as to unleash supernovas of creativity."
If by that he means that Google, Yahoo, Microsoft and what remains of AOL will have products that offer a Level 3 platform, then sure, that sounds reasonable. But there are plenty of things on the internet that don't benefit from letting people add random new functionality within the product itself.
Take Flickr, for instance. Lots of people use their Level 1 platform to make nifty things. Most Flickr hacks integrate Flickr with some other product or show nifty visualizations with Flickr photos. How would letting people do that on Flickr itself benefit anyone? The data is all that matters in this case. No one who makes Flickr hacks thinks to themself, "Man, I wish my hack could run on Flickr's site itself."
It seems to me that Level 2 and 3 platforms are only really useful in games, where they're nifty but not really a new thing, and two other places: portal pages and social networks. In terms of what makes a Level 2 or 3 platform, social networks are basically glorified portal pages to begin with. They give you a box, and you stick your content in it while perhaps using the site's Level 1 platform to make your content more useful and interesting. If you're not a social network or a portal, you don't have boxes for third party developers to stick things in.
I may be setting myself up for a "640k should be enough for anyone" moment, but I really don't see it.
Well, it could be argued that someone who makes Flickr hacks will, from the beginning, only consider external hacks, because only those are possible... it's a different class of functionality to start with.
I think a rebuttal is in Marc's definition of a platform:
A "platform" is a system that can be programmed and therefore customized by outside developers -- users -- and in that way, adapted to countless needs and niches that the platform's original developers could not have possibly contemplated, much less had time to accommodate.
In other words, just because we can't see the benefits, doesn't mean there aren't any (that's pretty scary reasoning to base a business decision on, though).
This discussion of platforms reminded me of Steve Yegge's rant on software that's alive: http://steve-yegge.blogspot.com/2007/01/pinocchio-problem.ht... He only specifically mentions platforms in passing, but he's mostly talking about the same thing.