"That's the point: through 2020. Here is Kocherlakota's language::
Overall: I see it as possible and beneficial to adopt policies that would lead to 5-6% growth per year over the next four years - which translates into a level of output in 2020 that is about 15% higher than is currently anticipated.
The 15% would carry forward, so output in 2026 would be 15% higher than the baseline. This is more or less what Brad seems to be suggesting as an extreme. Friedman zips that forward to a permanently higher rate of growth, leaving 2026 output 37% higher than the baseline. The other 22% would be coming from some sort of supply-side response (you can't apply demand multipliers once you get to full employment.)
Kocherlakota's reasoning isn't that sort of supply-side response: he's assuming there is likely a backlog of good productivity-enhancing ideas that haven't been implemented because of lousy demand. Increase demand, the ideas get implemented, and there is a spurt of non-inflationary growth. Not unreasonable, but we can dispute the likely magnitude and length. 10 years of it is a long time for that to play out.