I think it's easy for each side of the debate to come up with examples of where crowdsourcing works, and where it doesn't work.
Mr. Freedmans example, while fair, miss some of the critical points James points out in his book for what crowds need to be effective - specifically independence, and a vested interest in the outcome.
We are strong believers in crowdsourcing. At Cambrian House we use the crowds rank our upcoming product ideas. We even use the crowds to then help us build them.
Anyone will always be able to come up with an example of how a crowd has been mislead (stock markets), but I don't again with Mr. Freeman when he says collaboration of crowds does not work.