Regarding the energy costs of mining, the current lowest-economically-feasible (calculated by the ability to breakeven from mining) ASIC is 1.5W/GH and the current best is 0.6W/GH - the calculations you link are based on 650W/GH so off by a factor of 400-1000+. It's probably not that hard and I don't think anyone has really plugged in the numbers into figuring out power costs / transaction, but as mentioned in a previous comment, it's probably less carbon intensive than how money is currently shipped around.
The flip side of all this mining compute power is that it's economically undesirable to use zombie networks for mining vs other more profitable ventures.
A bunch of your issues against Bitcoin revolve around anonymity/lack of regulation, but I think the reality of the situation is very different from what you lay out.
It's almost impossible to get anonymous bitcoins: mining connected to a mining pool on Tor, or exchanging cash via LocalBitcoins, again w/ Tor. These are dependent mostly on the (questionably) anonymous properties of Tor and cash. If you are acquiring Bitcoin via an exchange like Coinbase, Bitstamp, etc you are pretty much required to hook up your bank account/provide identification, and reported to FinCEN or whatever your local regulatory agency is. What we're seeing now (like the China regulation you started out with!) is increasing regulation of BTC/local currency exchange.
Bitcoin-to-Bitcoin transactions, by protocol, are completely traceable as every single transaction is stored in the blockchain forever. As we've seen, once there are known addresses or transactions it's trivial to unwind or do network analysis to identify wallets, especially for government/regulatory agencies (or those engaging in huge dragnet surveillance/digital intrusion operations, ahem).
The same techniques used to bring down something like Silk Road and things like existing CP operations will be just as effective against other illegal markets. These techniques are actually aided by the aforementioned traceability aspects of Bitcoin and probably one of the reasons that LE/govt is supportive of Bitcoin adoption vs truly anonymous P2P cryptocurrencies like Zerocoin.
I suppose that Bitcoin could be used for large scale international money laundering, assuming one can easily spend BTC or convert BTC into currency sans regulation (again, unlikely anywhere except where currency conversion wouldn't be a problem in the first place) but you'd already need to have an infrastructure for moving/distributing a lot of off-the-books cash and the wherewithal to create a huge mining operation (which would still be traceable without a lot of opsec) and I don't think Bitcoin is any more conducive to that than any other global digital currency would be.
To me, Bitcoin-like (P2P cryptocurrencies) do have issues - can Proof of Stake work better than Proof of work? (ppcoin) Can PoW be useful computation? (see primecoin, but imagine if mining/transaction verification was linked to say Folding at Home/protein discovery) Can cryptocurrency be linked in a way that's internal and resistant-to-gaming to something like energy input or usage or can it otherwise be stabilized when bootstrapped?
I think the positive thing about Bitcoin is both in driving a big interest in analysis of the fundamental qualities people want in a currency, and providing a huge-step-up in terms of a platform/framework for creating/testing out different cryptocurrencies.
To me, Dogecoin points to the future - a world where it's ridiculously easy to instantiate and use semi-private Rushkoffian barter currencies, which nicely sidesteps a lot of the unsolved macro-economic issues.
]]>