You simply await the development of nuclear isotope batteries (up to a five-order-of-magnitude improvement on chemical batteries) as a mature technology: https://apps.dtic.mil/dtic/tr/fulltext/u2/a433348.pdf

(And hope that the technology doesn't lead to a new era of arms proliferation with the use of non-fission primaries)

No need — regular hydrocarbons have a high enough specific energy to do the job.

The largest pterosaur, Quetzalcoatlus, had a wingspan of ~10 meters and an estimated mass of 200–250kg [1]. I don’t know how much power they could produce, but it can’t have been more than a few horsepower given they weighed less than a half a horse.

A 30cc petrol motor will easily output 2–3 horsepower, and weighs less than 10 kg. Combine it with wings from a hang glider (similar wingspan, weight ~30kg), and as long as the powertrain isn’t too heavy (compressor + pneumatic artificial muscles?), it seems possible to build something light enough to carry a human.

1. https://en.wikipedia.org/wiki/Quetzalcoatlus

Yep, there was I under the impression that this was more of a marginal endeavour than it really is; I was probably thinking along the lines of human-powered aircraft. I'd imagine that a lot of safety problems and requirements for a very skilled pilot could be reduced in the future with computer control/brain-computer interfaces and exoskeletons.

What are the odds that this went classified and is already working and being used in the field?