It's the basic game theory: you have to be prepared in wars you don't want to happen, exactly because you being prepared for them prevents it.
Preventing conflict by developing ways of measuring outcome of a potential conflict without actually acting it out is a technique that's been used by animals way before humans came along.
Perhaps you could clarify your point a bit?
Seems like you could build a drone that meets your other specifications. You can either spend the saved lbs on more armor and weapons, or make the aircraft lighter for even shorter takeoff.
My cynical view is that too many people wearing stars grew up in the cockpit, and are still fixated on human pilots. If a remote drone manager in Henderson can handoff a CAS asset to your control, why do you need a human pilot in the mix?
But another problem with American policy is a need to keep American casualties to a minimum, to the point of being the overriding directive.
So from a combat perspective, a good solution is light, low-cost planes that can fly from very rough airstrips close by, armed with a light cannon that can shred buildings, light armor, and other low-intensity cover, and fly slow enough to make in-the-moment decisions from visual inspection by a human pilot. But such a plane is a tempting target for inexpensive, widely available shoulder-fired missiles, or even old-fashioned flak. And then casualties go up. This isn't WWII anymore. America won't tolerate high casualty rates politically for problems that are not existential threats to our nation.
And a cheap low-tech plane solution also has to slot in somewhere between drones (expensive but effective), and the helicopter gunships that have had the mission since Vietnam.
So instead, we're going to keep using hammers to swat flies, because we might need hammers against Russia someday and we're rich anyhow.