Which brings us to the one thing that will kill the Lego building block internet; it will be all about service and how service is perceived by the end user. Would anyone have trusted Facebook Connect four years ago and would anyone trust a MySpace connect now? If a service provider (and if you are providing a widget or an API, you are a service provider) fails, then it becomes a problem for the entire ecosystem that is build around that service. If the service becomes unreliable, then people will flee the ecosystem. Even a hint that a service provider is not reliable will cause adoption issues.
Right now developers are looking only at the big systems, Facebook, Google, and Twitter (even with its track record Twitter has been compelling) to build their eco systems around. Much like people built applications based on the Linux, Apple or Windows eco systems, people are focusing on the major systems because they have a longer staying power and a better chance of survival over time. When someone is building something on your API, widget or infrastructure, being 99.999% available is critical, anything less means that the developer or the company quickly loses credibility. Companies need to address the SLA issues first, and then the Lego building block internet might be something that many companies can participate in.
Interesting discussion about the perception of trust regarding third-party API’s such as Facebook Connect & others. He’s dead on regarding need for a focus solid SLA issues first before rapid adoption as an ecosystem or social layer to the web.