The problem with predictions is not the knowing of what but the knowing of when.
I spectacularly failed to predict the current financial crisis having expected it to occur several years earlier. I'm also expecting that my 2007 prediction of when cloud computing brokers will appear (along with exchanges in computing resources, futures and swaps in resources) will also fail.
This became clear as soon as Amazon introduced reserved instances. The only thing preventing the formation of a trading market in EC2 instances, is that reserved instances are not transferable. As soon as that changes, we will quickly move from reserved instances and the spot price to a commodity exchange and the usual paraphernalia of derivatives to spread betting.
Whilst I predicted that "six years from now, you'll be seeing job adverts for computer resource brokers", it now seems likely to be earlier.