If this becomes a trend, it's potentially revolutionary:  "Goldman Sachs's fund arm is developing a new global credit strategy for institutions that will rely on market prices rather than heavily-criticised credit rating agencies."

Any purveyor of information is vulnerable to the fact that the subjects of the information have much more at stake than the (many) consumers of said information.  You care more about the contents of your credit report than any individual bank.  In the case of the banks, this is counterbalanced by the fact that they have more money, and perhaps paradoxically, that they are fewer.  But in the case of things like bond ratings and search engine results, you always have to guard against the possibility that the subjects will find some way to bribe the intermediary.

The Goldman fund kills that incentive with transparency--if its ratings are bad, that will show up when it underperforms the market.  On the other hand, there's a real risk that its analysis will nonetheless converge with the ratings agencies, because of career risk for the managers.  If your fund drops when the market does, you still have a job.  If your fund drops at some other time, then no matter how good your probabilistic analysis, sheeplike investors will move their money elsewhere.  Like the sellout risk, career risk  seems to afflict most intermediaries that operate long enough.  (Including, natch, the government.)

Still, over the short term, it will be interesting to see whether investors prefer their black box to the black boxes at Fitch's and Moody's.