Be Either an App or a Platform, Not Both
ceklog.kindel.com
ceklog.kindel.com
Perhaps what he's getting at is that you need to provide value to the customer, or they won't buy. You can't rely on a 3rd party developer to do it for you, for no reason. But they might if you provide value to them. Which makes them the customer (and you a dev tools provider). If this is what he means, I don't see it as fatal to "building a platform". It just means that if you want to build a platform for 3rd party users, you'd better offer them some value in order to attract them. e.g. access to customers, data, IP, publicity, credibility, other resource, some cool proprietary technology, or even ease of development (dev tool). It had better be a real benefit that they couldn't easily (or better) do themselves.
Seems he's railing against some misconception that I haven't heard of, and my lack of that context and assumptions creates holes in his for me.
BTW: how is the first example he gives - the first IBM PC - an app? A PC on its own is useless. It might be bundled with some apps, but they aren't why you bought it. (Hobbiest computers were different: they were apps for people who bought them to hack). Of course, the "IBM" made it an exception: instant publicity, credibility, customers, ISVs, marketplace. But tellingly in favour of his position is the iPhone: so obviously a platform today, yet the original iPhone had no appstore. You couldn't put apps on it. It was sold for the built-in functionality: a phone, web-browser etc.
There's an old-fashioned term summing this up: killer-app. It means an application of a technology or platform that makes it worth buying. e.g. For apple IIe, it was word processing and spreadsheets; for the Mac it was desktop publishing. For game consoles, it's launch titles.
counter e.g. game consoles; eBay and Viacom (for 3rd parties not devs); android; heroku (and other PaaS)