[A]s close students of economic history know, the greater threat to freedom on the Web would be for the government to decide which companies are good and which need to be broken up or punished
Crovitz has it right. The "big is bad" myth is an enduring myth of antitrust--and one that undermines the important purpose that antitrust law is supposed to serve.
The fundamental problem is not "big," but "concentrated power." The motivation to abuse power, once concentrated, is overwhelming. That is true whether it is centralized in the public or the private sector. Those who possess concentrated power tend to favor slow, managed change over revolutionary innovation.
In the tech sector, the appropriate question has correctly focused on leverage, rather than either on size or dominance. While it is true that Microsoft attempted to leverage its OS monopoly to ensure that the Internet would develop through a MS lens, the problem was its leveraging tactics, NOT its OS monopoly.
The biggest challenge to the appropriate development of the information economy remains antiquated IP laws, not concentrated power. These laws create artificial bottlenecks that interfere unnecessarily with the deployment of marvelous innovations. Many of the most damaging leveraging tactics that we have seen over the past decade remain entirely legal uses of IP (at least as IP law is generally understood).
At the same time, however, the debate between those who favor draconian enforcement of no-longer-sensible laws, and those who would eliminate IP entirely, is facile. IP laws are critical to motivating innovation. But not all IP regimes were created equal. Our goal, as a society, should be to purchase maximum innovation with minimal interference to competition. As things stand, we are doing a poor job.
The Internet, like all bits of infrastructure, requires enough rules of the road to prevent it from descending into anarchy. IP reform is the right way to go. Antitrust law, other than in specific and unique circumstances, is not.