That's really helpful. It predicts 534 Watts to get me (110kg, 0.4 CdA) to the 45kph limit for Class 3 California bikes.

The Grin Technologies motor simulator https://ebikes.ca/tools/simulator.html predicts (the "Load Line") 837 Watts for "MTB upright", 648W for "MTB, Tuck" and 570W for "Road, Tuck".

I do find that in an upright position it takes 100-200W of human power in addition to 750W power measured at the input to the motor controller. The controller's pretty efficient, but the motor (GMAC) is only 85% max, so there's maybe 570W at the axle. The US CPSC and California laws aren't clear whether the 750W/1 horsepower limit is power in or power out, but power in is what shows on the display, so I limit it to that. Also, even a mild grade will cause the motor to approach the 70 deg C limit I chose after 1 minute. I live in the flattish Santa Clara County, but I'd definitely want a mid-drive if I lived in a hilly area. A 250 or 350W Riese & Müller Delite https://www.r-m.de/en-us/bikes/delite/ I test rode needed a lot of human power to hit the 45 kph max, but it climbed hills in Larkspur like a goat.

750W will have no problem getting a recumbent bike to 30MPH, but don't count on it for an upright.

Oh, and to see what 700W from a human looks like, an Olympic track cyclist toasts a piece of bread: https://www.youtube.com/watch?v=S4O5voOCqAQ

Yep. Bradley Wiggins's 2015 velodrome hour world record was 440W (since eclipsed in distance but not power). This puts into perspective how cheap power is in the industrial age. His hour record was the equivalent power of 35 grams of gasoline. Or it's about 5 not great power tool batteries (eyeballing eBay listings). It's gonna be tough when we have a world built for nearly free energy and we don't have the free energy anymore.