An excellent analysis, and although I don't agree with all of his points, I must concede he has changed my mind on others. I think his diffusion model is essentially correct, although I feel one aspect about it is self-defeating--even if a civilization is strongly motivated, psychologically or culturally, to explore and expand, it is unlikely that it's colonies will preserve those attributes for the long periods of time required to establish themselves and flourish on the new worlds. In addition, the new planet, no matter how hospitable, will still be somewhat unsuited for the colonists who did not evolve there. This means the probability of any Creole* culture to retain its exploratory drive, and the resources to support it, would decrease over time and space, not remain constant.

The other alternative, that suitable planets are not required, just suns with energy and comets and asteroids for building materials, (Oneill-type installations) certainly increases the number of potential colonies, but also reduces the motivation to start them. Why continue expanding when any star, even the nearest one, will do?

And though the advantages of directed energy communications devices over broadcast methods is obvious from an engineering standpoint, I feel they will mostly be used to communicate with the "mother planet" or other known colonies. I still feel the most likely method to contact strangers will be broadcast beacons. The Galaxy may be criss-crossed with high-bandwidth communications beams, but it is highly unlikely we would be lucky enough to intercept one of them, and even if we did, stellar proper motions would soon move us out of the line-of-sight. And to deliberately beam optical lasers or similar tight beam devices at stars that were selected as likely candidates for technological civilization would be very hit-or-miss. Civilizations are highly separated in space AND time. The probability of picking the right star to contact at the right point in it's technological development is negligible.

As for the self-replicating probe idea, not only do I find this economically unfeasible (babies are cheaper than robots, no matter how advanced your technology is) but it seems suicidal to me. von Neuman machines need not only be self-replicating, they also have to be able to evolve to meet new conditions and unanticipated obstacles. Indeed, even efforts to prevent them from doing this would have no guarantee of success. Sooner or later, your von Neumann hordes would be wandering around the galaxy with agendas of their own. This scares me.

It might make a good science fiction story: the only provocation that might make an interstellar war worthwhile, or even necessary, would be if you found out one of your neighbor civilizations was seeding the galaxy with autonomous intelligent evolving self-replicating machines. Sooner or later, you would get something like Fred Saberhagen's Berserkers.

*Creole - This term does not apply just to the Cajun residents of New Orleans, it is a word meaning "the new culture arising in the colony". The term exists not only in French, but in at least one other language, Spanish: "Criollo". I suppose the nearest English translation would be "Colonial", or in the case of the US/Britain experience, "Continental".