Actually, I've just gone throught it a bit more rigerously and for a given mass and road speed the reduction in total kinetic energy of the bike and rider by switching in isolation to wheels that are around 2/3ds the diameter (and 2/3ds the mass) comes in at less than 2% compared to 700c wheels (using ballpark values).
Something that was previously lost on me until I looked at the spreadsheet a bit more is that while smaller wheels weigh less, have a smaller radius of gyration and as such have a much lower polar moment of inertia (for the example above it falls by around 70%, so to accelerate the smaller wheel at a given angular rate would only take 30% of the energy), for a given road speed of course the smaller wheels needs to rotate at a higher speed; which largely offsets its reduction in inertia.
All the above is for masses typically encountered on an adult bike. If we cut the bike mass by a third and rider mass by a half to simulate your younger self the difference in total system energy between the two wheel sizes rises to a bit over 3%.
Interesting stuff, arguably
I'd certainly be interested to give one a go.. although I'm not sure I've even seen one in the flesh before tbh, so the chances are probably pretty minimal!