Basically, I believe that exploring the general questions about how mainstream views are generated in practice and what are the implications for their reliability is by far the most fruitful direction for people interested in increasing the accuracy of their beliefs across the board. Of course, if you have a particular interest in some question, you have to grapple with the concrete issues involved, and also a general exploration must be based on concrete case studies. But attacking particular mainstream views head-on may well be counterproductive in every sense, as I noted above.

This is fair; you've made your position clear, thanks.

That's assuming you have discussion partners who are knowledgeable, open-minded, and patient enough. However, such people are the most difficult to find exactly in those areas where you're faced with the Scylla of a deeply flawed mainstream and the Charybdis of even worse crackpot contrarians.

Agree in general. How about Less Wrong in particular?