Adam Cohen is wrong.
Seeding Initial Wealth: If you're jealous of the early adopters, become one. If we assume Bitcoins fail, who cares who adopts when? If they succeed, then it's still early. As for mining being random or useless, no, miners are paid for securing transactions. The way Bitcoins work, seeding initial wealth is a side-effect of the distributed transaction resolution system.
Built In Deflation: He repeats the idea that the set of early adopters is fixed. If Bitcoins last long enough for the deflation to happen, you can still be an early adopter since it's still very early. (And they actually won't deflate because the money supply will be inflated by things like fractional reserves, mortgages, and the like just like other currencies are.)
Lack of Convertibility: I don't think a currency has to be able to act as a store of value. In fact, use as a store of value to some extent conflicts with use as a means of exchange. But in any event, Bitcoins are already reasonably convertible and as they become more popular, exchanges will increase as well.
When Something Goes Wrong, It Will Die: Serious things have gone wrong with Bitcoins, including an exploited bug that let people create Bitcoins out of thin air. What happened? Everyone who cared about Bitcoins pulled together to fix the problem. But if you're really worried about this, the fix is simple -- don't use Bitcoins as a store of value, just use them as a means of exchange.