It will be left to the Obama administration to sort out whether, or in what way, the government might advance new billions to GM and Ford. As Tom Friedman writes this morning, past failings of management at these companies are well known: they failed to support healthcare reform in the early 1990s, and they are now choking on health management costs; much more important, they decided to profit from SUVs, that is, from truck bodies and cheap oil, and foul the atmosphere. The smug attitudes of GM executives in particular--so I thought during my days editing at the Harvard Business Review--gave you some idea of why the Reds shot Kulaks.

Nevertheless, a couple of million jobs are at stake, at least until other global auto makers can take over some of this plant capacity. Should new public money believe in, as Secretary Paulson put it this morning, the sustainability of American car companies? I can see the macroeconomic arguments in favor, and I don't need Michael Moore to imagine the misery of new shut-downs. But before we spend more on this management, we should see that their worst failure is actually recent, not in some fading past. I mean their failure to remain competitive on the most basic principles of design and manufacturing, using global, peer-to-peer information platforms; principles Volkswagen Group, Toyota, and Nissan-Renault, have exploited, and GM and Ford have half-exploited, even as profits mounted from SUVs and endogenous quality improved.

The key to making all manufacturing profitable these days is lowering transaction costs for any particular product development program. Look at Sony or Samsung or Apple or Honda. You want to have the capacity to experiment often, design for many niche customers, and hold on for the grand-slam. You don't want that much riding on each try: you want (if you'll pardon another sports metaphor) to transform your manufacturing processes from football into basketball, that is, create the capacity to go the basket many times a game, not work your way to the end-zone only a few times a game.

This means fully exploiting peer-to-peer information and components sharing. You want to enable design and production managers in each brand unit to source components from every other brand unit (something like the way "open source" software designers source and integrate chunks of code). You want to decentralize vehicle design, to enable new vehicle introductions in shortened cycles (to be competitive, these days, you have to be under 36 months), and profit from production runs of, say, 50,000 vehicles rather than 200,000 vehicles, which is the American standard.

Skoda, which I know particularly well, thrives today because Bohemian designers have learned to exploit access to all of VW Group components. They create cars especially for niche markets; their latest, the Roomster, breaks-even on about 60,000 units, using components from the whole of VW Group. And Skoda is over 20% of VW Group profits today. (No, their competitiveness is not just cheap labor, which is more expensive than Korea's.)

Would Ford be dying if Ford designers, world-wide, had had access to Volvo and Mazda engineering and components years ago? (The flip: would Chrysler be dead if Daimler allowed it access to Mercedes components?) When designers use information technology to integrate components from federated sources, they have the capacity to introduce new vehicles more quickly, and make money on individual vehicles from comparatively few units sold. This has been true for consumer electronics for many years.

One key to making shrewd public policy, in other words, is understanding how products and services are actually assembled. Innovation--the driver of growth--is usually a matter of integrating in unique ways bits and snips of components, information, code, data, etc., from federated sources, much like a car (or a blog post, for that matter). The financial instruments that imploded were, after all, chunks of information--in the case of mortgage-backed assets, misinformation--derived and syndicated from federated lending institutions, integrated and bundled in novel ways, and distributed instantaneously over electronic networks--what Bill Gates called (too sunnily, obviously) "frictionless markets."

There are obvious dangers here, but there is no going back: what we have here is a template for assembling pretty much every high-tech product or every high-tech manufacturing process that produces a low-tech product. The same template works for delivering professional services. Peer-to-peer networks are crucial; nobody is as smart as everybody.

Too often, talk of "stimulus" is lazy, outdated, or vague. The government cannot save GM and Ford by giving them money and waiting for (or even mandating) advanced hybrids. Products don't succeed this way; the Chevy Volt will need to be a part of a family, its components quickly wrapped into a new Saab, or some new California sports-car wrapped around an I-Phone. And each of these models will face stiff competition from current global competitors.

Government--more precisely, governments--will have to see these things, especially if they are going to help auto makers. There is a need for a tariff regime that would allow components to come from wherever they are designed and, or, competitively produced (the EU was crucial for Skoda getting components from Seat in Spain or Audi in Germany). There is a need to make R&D programs in crucial areas (batteries, fuel cells, etc.) more transparent and less burdened by potential IP battles, so that university labs and free-lance inventors can participate--programs no single company can afford any longer. There is a need to jumpstart a grid and social network for electric cars, as Israel is doing (or at least pretending to do).

Most important, there is a need to stimulate innovation by entrepreneurial companies in the supplier base: by making patent protection harder to come by (therefore, less of an obstacle to start-ups), creating new ways for companies to share intellectual property (like IP exchanges), or creating incubators and tax breaks for new businesses.

So. Can US companies use government funds to create things customers will actually want? If so, are there things all manufacturing companies need in common, like roads and bridges? If not, should rescue be forthcoming, or should Toyota (which is not exactly a Japanese company anymore) be allowed to take over GM's capacity in the hope it will reinvigorate its plants and suppliers? This would not be unprecedented, after all: during the generation before the year 2000, it took about 15 years for a third of the Fortune 500 to be selected out, that is, fail or be acquired; since 2000, it took about 4 years.

The most compelling challenge remains the competitiveness of ordinary businesses--especially what we call, justifiably, knowledge-based entrepreneurial businesses. Perhaps 60,000 new businesses a year started up in the US during the 1960s; during the past decade, the annual number was closer to a million. We have to understand how and why barriers to entering businesses have fallen, how and why investors find new businesses (much more quickly than before), how companies survive using information platforms to design and market new things. Financial capital is not the only capital that flows or is regulated and managed. So is intellectual capital--and this is the more important kind.

Correspondingly, engineering innovation is not only driven from the center, but from myriad peripheries: from product development teams in companies, and the entrepreneurial start-ups they tap into. And there is a role for government in managing up a national "network effect"--improving the quality and scale of knowledge sharing on the platform. Facilitating even marginal improvements in the exchange of knowledge will have an enormous impact on how we live. It will give a whole new meaning to the term "roads and bridges."