Who Should Obama Hire To Be America’s CTO?

There’s a discussion making the rounds of several groups I frequent (including the Enterprise Irregulars) as well as Techmeme about who Barrack Obama should get to be America’s CTO.  It comes in the wake of John Doerr saying he thought Bill Joy should get the job.  This was in response to a question supposedly sent indirectly to Doerr via New York Magazine writer John Heilemann.

If indeed Obama asked for a CTO, Bill Joy might be a good choice, but I don’t think the government has any business needing a CTO.  At best, an agency like DARPA might need someone with vision to decide where to allocate scarce research money, but even there, the job makes more sense to turn over to a group of people who are up to speed many different disciplines.  Sticking with the CTO theme for a minute longer, others suggested candidates such as Vinton Cerf, Ray Kurzweil, Danny Hillis (where is he now that the multicore crisis is upon us?) or Marc Andreesen.

My problem is I just don’t believe the US government needs a CTO.  I can’t imagine how that role makes any sense whatsoever because a technology visionary is by definition not going to have vision that is broad enough (find me one who spans everything the US government may want or need vision about).  An entrepreneur as some of these folks are (Kurzweil and Andreesen at least) will have had zero experience pushing the rope that is any high level government job where you have minimal ability to just mandate what you want, so that doesn’t work either.

Others suggested Bill Gates or Steve Ballmer.  I’m not sure anyone sees Ballmer as any kind of visionary, but even if he was (and Gates has been, though not recently), these two are CEO’s.  They won’t work for reasons similar to why the entrepreneurs don’t work–they expect far more control over getting people to jump on command than the job is likely to deliver.  Meg Whitman was briefly discussed as a political participant not too long ago and suffers the same problem.  It’s this mindset that causes Seth Godin to say of CEO’s:

Trying to convince a CEO of anything is a little like trying to convince a cop not to give you a ticket. It’s possible, but rarely worth the effort, given the odds.

To be effective, I have to believe this person needs to be someone who listens, who can work the system politically, who can manage outwardly and upwardly, and who is more of a conductor/orchestrator than great visionary.  After all, do we need this guy to invent the equivalent of the warp drive, or do we just want the government to screw up their IT projects a whole lot less and avoid damaging the tech industry and US innovation inadvertantly through foolish policies?

Richard Brandt wants to hire Nathan Myhrvold.  I can’t think of a worse choice, and I can’t imagine why Brandt (a journalist of some repute) things he is a good choice.  Why is Nathan the worst choice possible?  Because he is running perhaps the Mother of all Patent Trolls.  He is exploiting the system to a prodigious degree for personal gain at the expense of innovation in a way that I wouldn’t want any public servant to ever emulate.  The mere association with such a venture would kill any possibility I’d support him.

TechFlash, coming from Seattle, suggests Jeff Bezos or Ballmer, and even puts Jerry Yang (yikes, did I say Nathan was last, hate to choose between Nathan and Jerry right now!) on their poll of who to choose.  Bezos is a smart man who would fail because he is an entrepreneur and CEO for the reasons I’ve discussed.

One thing I think Doerr has gotten right in terms of the role of this person is to create programs that produce more technologists (specifically he wants more engineers and scientists).  But here again, I don’t think Bill Joy, some technologist, or some CEO is likely to understand that.  This is a person that would understand how large organizations behave and how to incent them to do the right thing despite themselves.

Richard Koman has the right idea about this when he says:

These, as well as Hillis and Joy, are ridiculous choices. The CTO job is a political job, a bureaucratic job. The person who succeeds in that job will be someone who can bring an entrepreneurial spirit into a government setting. They will have to familiar with the CTO positions at the whole range of federal agencies; they will have to know their way around Washington to some extent; they will know how to work with large, combative constituencies; and they will expect to be held accountable.

So what sort of person has all of the right qualities? 

1.  They have to be political animals, able to manage and motivate those who are not even in their organization. 

2.  They have to be able to deal with significant accountability, so they must actually produce results.

3.  They have to be able to get smart people of the technology geek persuasion rallied and all moving the right direction.

4.  They have to be able to execute hugely complicated projects that are over constrained and come out of it with real ROI’s and as few failures as possible.

5.  In all likelihood their organization won’t build any technology.  They must successfully manage a gaggle of outside vendors to do what’s best for their country and get over their own self-interests and desire for pork.

I’m surprised Ed Cone at CIO insight didn’t recognize this job description when he wrote about this story.

Aren’t describing the job of a corporate CIO rather than a CTO?  They do #’s 1 through 5 all day long.  I’d look for one that has been with a company with a particularly large scale and a reputation for having used IT to gain a real edge.  Check the GE’s, Dell’s, and Wallmarts of the world.  Heck look for Maynard Webb for than Meg Whitman if you’re going to look to eBay.  Look to Rick Dalzell (was the CIO) of Amazon rather than Werner Vogells (the CTO).  No others need apply.