Of course, the compilers needed to take this source code and recast it into executable form are expensive, and the expertise required to operate them somewhat rare. But the result is a real, bona fide weapon of mass destruction. Really mass destruction --- far wider than you could achieve with any chemical attack. More perhaps, even, in a worst-case scenario than a "mere" isolated tac nuke. Of course, these are less discriminate weapons than nukes, which don't stay around for months killing people thousands of miles from the point of initial release. No matter who releases this, they would be killing their friends. But serious young men driven by some wild ideology might not let that stop them, any more than the Sept. 11th hijackers were bothered much by all the Muslims working in the World Trade Center. So, he has a legitimate concern.
Now, to make a nuke, you need specialized knowledge and equipment --- and highly enriched uranium or plutonium which takes billions of dollars worth of infrastructure to create. (Before World War II, physicist Niels Bohr said that it was impossible to make a uranium bomb because you'd need to cover an entire country with enrichment plants. Later, on his first visit to Los Alamos, he claimed vindication: "I told you it couldn't be done without turning the whole country into a factory. You have done just that."). To create the virus, you need specialized knowledge and equipment made of cheap materials, which fits in a building that doesn't look all that much different from a well-equipped medical lab. And the genomic equipment has plenty of legitimate uses (unlike bomb tooling), and it's getting cheaper all the time, while the expertise gets ever more widely distributed. So Kurzweil (and his co-author, Bill Joy) argue that the source code for the virus --- its DNA sequence --- should be kept secret.
But that tactic has its limits too. Techie blog BoingBoing titled its post on the Kurzweil/Joy opEd, "Kurzweil and Joy call for genomic Manhattan Project". An odd title --- most of the time, when people call for a "new Manhattan project", they're talking about throwing massive resources at a problem. But plenty of money is being thrown at genomics anyway. What Kurzweil and Joy are thinking of is the secrecy. And the real Manhattan project shows the limits of what can be achieved in that field. Obsessive as the Manhattan Engineering District was about security, their innermost sanctum, the section doing the most sensitive parts of warhead design, let in a serious young man driven by a wild ideology --- Klaus Fuchs.
Bohr's quip is reported in ch. 15 of Richard Rhodes's "The Making of the Atomic Bomb" --- by way of Edward Teller, of all people, on p.500 of my edition. As for Fuchs, the best account I know of is in Rhodes's follow-up book, Dark Sun.