He has a point there. Transportation, heavy machinery, power grids, cruical infrastructure, healthcare, economy and so on all increasingly run computer systems with questionable security. Lives and world order depend on them. Are we really going to draw the line at voting? (edit: expanded on this point a bit)
Trust on the scale of several years is a difficult thing. People go from lovers to archenemies in that time. But yes, some long term projects do need stability and long term support to become fruitful. You could introduce constraints to support this, or trust the voters to take it into account.
But even more important than time is the granularity of trust. Just as in everyday life I trust different people to different degrees in different domains, I would love to have a voting system that reflects this with regards to political representatives or activists.
If you can change that byte across even a small percentage of voting machines, you can leverage that into control of the country's government.
All of the examples you listed are continuously running, auditable, highly heterogeneous systems that can have multiple parties cross-checking because there isn't the absolute need for the actions of the participants to be untraceable back to them, the way it needs to be with voting.
I work with government often. I am often outside begging to be given access to help my community with one civic project or another. Interactions that take hours in open source communities, take months to coordinate with City staff.
I'm the last one to advocate taking the brakes off the car, but there are many many many parts of government where decisions require several years only because everything takes years to do. I think there is a perspective, if you step back to imagine this world, where we as citizens can participate and make the turn-around quicker. Not everywhere. Many government functions require care and caution and sober second/third/fourth thoughts.
But many of these actions need not take years, if we could re-imagine something better. And right now, the lifeblood of government runs like molasses, not because each slow process has been intentionally made slow, but because the processes and practices and hierarchies are structurally slow at their core. That's not clever slowness. That's a calcifying fossil.
I personally think giving citizens more decision-making access will keep civil servants on their toes, keep them learning, and speed up the cycle time in many needlessly slow areas of government.
Can we start with the city? Bill de Blasio (bless his heart) won 282,344 votes in the 2013 Democratic Primary. What's worse is Lhota won his primary with just 32,236 votes. https://en.wikipedia.org/wiki/New_York_City_mayoral_election...
I hate to agree with President Obama on this topic but I have to agree with him. The first problem we need to tackle is not to fix the electoral process. The first problem we need to tackle is getting people to actually care enough to show up and show up in large enough numbers so we can effect change.
Sure it might get harder with increase scale, but I don't see it as a huge problem.
Recently they tried out e-voting, with what I consider predictable results:
http://www.swissinfo.ch/eng/voting-with-a-click_hacking-fear...
Also, I agree with ianstormtaylor in that I don't think that what you're proposing would necessarily be good. You need a political system that you trust will promote wise decision makers, and then you need to not micromanage them. If the people who get elected can't be trusted to make good decisions on a scale of several years, then the political system needs to be corrected, not expanded.