Lanier seems to be looking for a way to improve software quality towards a goal of creating applications with tens of millions of lines of code.

I think the software universe is much more likely to evolve in the direction of tens of millions of applications (of perhaps a thousand lines of code each) that can all figure out how to talk to each other if they need to.

Even if that's not much more likely, it's the way I think the software universe *should* evolve -- not towards ponderously magnificent edifices with cracks in the foundation. Swarms at the edge of chaos, not crystalline skyscrapers.

If phenotropics can help to nudge us in that direction, then I'm all ears.