I'd say using a commercial component that's been left for dead is much worse.
The current title reads like: "this is what you owe me if you want me to look at you." Well, ok, but I don't feel that way. I'll use smaller, humbler projects. But I do agree that in doing so I must be prepared for the day that it is no longer updated and the bugs aren't being fixed. I'd certainly feel more comfortable with projects that met most of the lists' criteria, and would be willing to bet more on them for the long haul. But I would still use smaller projects.
Some things (couchdb) seem riskier than others (php), and some very small projects (php-on-couch) seem quite risky indeed. However, I could pick up and go if php-on-couch stalled, I could probably manage if couchdb stalled, but I would be in trouble if php did.
You get what you pay for. And your best insurance against the project being completely abandoned is to have the source code.
As someone whose tried to use open source whenever possible I can tell you there's no feeling worse than using an open source component that's been left for dead