This really is not my conclusion. I'm not nearly that smart as you well know. But here goes:
Many electronic currency models born out of supercomputing already show this is very doable. Still a theory, which is fine, but now I'm seeing more stuff about the practicalities of how the quirks in a hybrid cash/electronic system can be managed so it reminds me a lot of that moment where models about blockchain currencies suddenly became reality and how that moved very quickly from there. I don't expect this to move that fast, but it is something to watch.
How?
To far oversimplify a lot of very beautiful equations: "monetary" policy in an ideal electronic currency system would be freed to use its full quantitative forecasting power to right-size the amount of "money" in the system in real time. Need a small influx of "money" here or there, you don't rev up the actual money printers and flood the banks with trillions in currency (that's insane frankly) that then depreciates the money the bank already had and inflates the amount of that currency needed to buy real world stuff. You give the economy the money it needs, no more, no less. If the zero bound is broken, it is resolved in milliseconds in this system instead of having people lose their life savings and jump out of windows.