I just learned about bitcoin a few days ago, but really find it fascinating. I read OP's linked article, but I fail to be dissuaded of bitcoins potential, usefulness and security. Here is why:
Peer to Peer?
The author points out that its impractical to keep record of every transaction ever done on every single node, a point which the creators of bitcoin acknowledge. The author fails to point out that bitcoin uses a merkle tree to essentially compress transaction data while maintaining security, making it possible to avoid nodes having to contain every transaction. He does note that scalability would probably lead to a tiered system, in which there are supernodes that do have the capacity to maintain all transactions, while other nodes would not. The author likens it to "commoners" and "lords" in a british peerage system. This is not an accurate analogy, since the barrier for entry to become a "supernode" equivalent or "lord" in the system he describes is very high or impossible for most, while in bitcoin that barrier would remain relatively low. See bitcoin wiki on "scalability".
Anonymous?
The author points out that bitcoin is not perfectly anonymous. I don't feel this is a requirement for a useful currency, but regardless, bitcoin has very good anonymity compared to anything else. I wouldn't discount its usefulness simply because someone set the bar at perfect anonymity.
Secure?
security and reliability is probably the primary concern of some new form of currency, but i dont feel the author, nor do i have a good grasp of how bitcoin achieves its level of security. I think his claims here are a bit overblown. He talks about block forking attacks and how they have the potential to disrupt transactions and accounts, but does not acknowledge the scaling difficulty of the encryption required to find blocks, or how that relates to processing capacity of the entire network. With out understanding the rules nodes abide by when addressing forks either, its hard to make conclusions as to how the entire network will behave in these instances. See bitcoin wiki entry on "how bitcoin works" for more information.
He has an interesting point about asymmetry of good nodes vs bad nodes. Good nodes must be constantly on(using power and money) as a majority of a network to maintain security, while bad nodes can enter at any instant to disrupt things, implying that a cost benefit analysis would favor bad nodes. I feel this is addressed not only by coin generation in block finding but also by transaction fees, which again the author does not mention. These two ideas have the potential to put cost benefit in favor of the good nodes, if necessary at all.
While interesting, i dont think the authors points are very good or informed, i much preferred "Computer Scientist" on the fatwallet forums counter arguments to bitcoins here:
http://www.fatwallet.com/forums/finance/1090435/?start=192