It's interesting that this write up is not as negative as it could be. For a phone service some of the failures he mentions are pretty bad. Is Arrington pulling punches because he's already been positive?

Arrington claims people expecting Grand Central to do what is says are putting too much faith in a beta product. What you can get away with has a lot to do with your offering. I'm not going to stop watching justin.tv because it's down (like right now, completely dead). But I'm not sure if beta is a good enough excuse for me to miss calls from clients. Especially now that beta is ubiquitous and virtually meaningless. Gmail certainly would not be off the hook for losing my emails. Is it really just caveat emptor for beta products as he suggests?

I would say that the answer to your last question should be a resounding "yes". The whole point of the "beta" label is to alert potential users that the product is not completely stable, but is in some usable state. The problem is not so much that the service was not 100% reliable, but that the NYT readers expected it to be and relied critically upon it.

As most Linux and 'BSD users have come to learn, "beta" means just that: not ready for critical use. The user gains some privileges, like being able to use "bleeding-edge" software or to get a valuable service for free, and in return they sacrifice the stability and reliability that comes with time-tested software; all that as they actually are an integral piece to testing that software. Without such a beta stage, the whole rapid development movement would fail as projects get mired in extensive internal testing stages.

Your GMail comment especially strikes me, as I hadn't considered that before. Perhaps they grew much faster than anticipated and are still working on stability issues, and the "losing my emails" scenario may not be far off the mark. By attaching the "beta" label to their service, they can probably avoid any liability that may arise if such an event did occur.