Basically, the OEMs are in a high fixed-cost business. But unlike steel, where the costs tend to be at the plant level, in autos they are at the firm level. Fiat's Marchionne is explicit that that is why wants Chrysler (and even more, Opel, GM's European operations). Without sheer scale, Fiat cannot undertake the R&D on new powertrains (such as plugins) and materials (a unibody car is a complex mix of steel and aluminum alloys, made using newish technologies such as hydroforming and with all sorts of complexities in welding or otherwise joining dissimilar metals). Without such clout, Fiat will have a restricted dealer network, in a narrow set of markets, limiting its pallet of vehicles types, and reducing the chance that a bit of sales here and a bit there of niche products will make the underlying vehicle platform a profitable proposition.
This is not to say that GM didn't have excess manufacturing capacity for any near-term volume of sales. This is not to say that GM didn't have unsustainable legacy costs. This is not to say that multiple model lines hadn't been muddied, and that cutting one might be the only way to clean up the mess. [See my earlier post on downsizing.] Some pruning is appropriate. Jerry Flint's argument is that what we see is not pruning, and will kill the tree.