Resolution limit of the eye
I'm not sure it's the number of megapixels in the eye that determines its optical performance. The detectors are a bit flakey, but the processor is very fast and well tuned.
An absolute limit to performance should be diffraction from the pupil, which could be up to 5mm behind sunglasses. This corresponds to a best possible resolution ability of order (wavelength/diameter) ~ 1/10,000 radians, 20 arcsec, 1ft at 10,000ft, or 50ft from low earth orbit. (1 arcsec is about an inch from 30,000ft.)
However, to see a bird you don't have to resolve it as such, as long as it's dark enough to see as a black dot on the sky background. I suspect this is the factor that determines when they could be spotted, rather than the angle subtended. Of course, even if they can be seen, seeing is one thing and avoiding is another.
To confirm little things with high contrast are visible: stars are tiny (~1 arcsec with atmospheric refraction, ~0.1 milliarcsec without). Your eye can't perceive their extent, but they can certainly be seen.