I'm *not* sure that you have to vote for harmful intentional actions when harmful unintentional actions may fit the bill. For example, we have a neutron star (RX J185635-3754) zipping through space about 200 light years from Earth. The explosion that produced it was only about a million years ago. The production of neutron stars from supernovas or gamma-ray bursts from mega-supernovas produce enough radiation to fry most information preservation systems (e.g. DNA, computer memories, etc.) on nearby planets. So the development of advanced civilizations involves a *lot* of luck.
There is also the fact that Berserker's don't provide much unless you activate them on the basis of revenge. Otherwise it makes a lot more sense to send out harvester probes to bring material and energy resources back to an originating system (over the longer run the construction of end-point Matrioshka Brains or other ultimate computing resources are matter & energy constrained). It seems probable that such harvesting would be detectable and it makes much more sense to get the resources from the cheapest source possible rather than getting into a "debate" with another civilization. So if/when such harvesting is taking place we (in a more advanced form we will probably be able to observe it -- subject to light speed time delay constraints) and we will probably be able to plot developmental paths that would waste the minimal amount of resources (fighting over them rather than harvesting whatever can be obtained at the lowest cost). The Berserker idea *might* be valid at a later developmental stage of the Universe -- but I suspect we really don't know enough about physics yet to begin to make that case.
Robert, who may speak at TransVision, is also currently thinking about the potential for out of body processing power:
Think of something something like a pacemaker that connects to a external computer that is linked to a bunch of sensors that monitor skin moisture/resistance (driven by how much you are sweating), external temperature (that may be causing you to sweat), blood pressure, eyeglass cameras and listening devices that determine whether you are watching a scary movie, etc. The computer performs an integrative analysis of all of this data and sends out signals that allow the pacemaker to optimize blood flow.
Mental inputs & outputs will not be far behind. There is *way* too much money to be made by using a Wifi interface between the human brain and Google to improve memory and ultimately intelligence for it to be far behind. And the fact that the Google founders are going to be *really* rich soon makes it an obvious thing to work on.