I refuse to work on anything that compromises my morality so that includes your case, gambling sites, user tracking (and advertising platforms) and particularly weapons systems etc.
Note: I have no issue with people who do, it's not on me to judge them.
Modern general usage aside, it's always meant 'overwhelmingly impressive' rather than 'good'.
It's all coming, the question is who gets it first and how accurate it is.
If you think about it rationally, this is actually not a bad thing. As MAD shows, if there is a potential for armed conflict, it's better from the casualty standpoint if one or both of the sides have overwhelming firepower. Then either the weaker side, or both sides will choose not to engage because the "victory", even if achievable, would come at too high a cost.
The only real problem I see with this is if this tech makes its way to the "bad" folks, like ISIS. When any idiot can somewhat reliably hit targets a mile out, things get problematic.
Another interesting thing about this is that with today's science and technology I could cobble a pretty good version of what I described together in less than a year, hardware and all, for under a million dollars (assuming I already have multispectral dataset, if not, add another million). Which tells me that pretty much every even somewhat developed nation state already has their own version.
That, or they don't because target identification and having authorization to fire are not problems that will be so easily solved.
We generally try to avoid friendly fire. The device you described sounds like an indiscriminate killing machine. There's a reason we gave up area-denial weapons like mines. They kill indiscriminately and create huge political problems that mean you can win a battle and lose the war.
And in this case there's no "indiscriminate killing machine", because soldier pulls the trigger - the weapon does not fire on its own. And that's the way it ought to be.
>Not even faces, _any_ part of a human. Butt sticks out from a shallow trench? Butt gets shot off. And it won't be limited to visible spectrum either, so it could largely "ignore" camo. And it could be augmented by similar AI-based person detection from a drone.
These two statements don't seem to reconcile. The second one has major things wrong with every single statement there.
>Soldiers in the field do not receive "authorization" for every shot they fire.
You might be surprised. For one, on just how detailed the rules of engagement for a particular operation can be. For two, outside of imminent threats, there's also quite a bit of direct authorization going on. Especially for sniper teams.
I don't see why you'd think that. Just because the scope highlights everything you can shoot doesn't mean that it will fire the weapon. We are not (yet) at the levels of accuracy where such decisions could be made by a machine, except for area denial.
(And if you're running out of ammo, you'd need to add more mines too)
A gun is already an indiscriminate killing machine. Having an 'only pewpew when on target' setting doesn't change this.
Now, using this tech to set up an automated sentry turret is another matter (and something I'm still flabbergasted isn't commonplace, given the number of people in any developed nation with the skills and resources to build one in their shed.)