[Ultra-Tech] What would naval warfare at TL10 look like?

[RogerBW]

Originally Posted by Crystalline_Entity

How likely is it that obtaining overflight rights from third-party/neutral nations would be problematic?

It's politically problematic enough in the real world that a sub-orbital troop transport was studied in 2002 specifically to get round the issue.

Ithacus and SUSTAIN

Originally Posted by lwcamp

Or you could put a big mirror on the aircraft.Then the friendly ship "shoots" the aircraft in its beam collector, and the beam is then re-directed toward the real target (or to another friendly aircraft, and so on, for longer chains before you reach the target).

You run into problems there, because the mirror typically has to be light, but the beam has to burn through an armoured enemy vehicle or missile. Making things shinier works up to a point, but you quickly run into diminishing returns.

[Daigoro]

Originally Posted by RogerBW

IMaking things shinier works up to a point, but you quickly run into diminishing returns.

A dielectric mirror can have 99.9999% reflectivity now at particular chosen wavelengths, so a perfect mirror for weapons-strength lasers at TL10 seems quite feasible.

[RogerBW]

Originally Posted by Daigoro

A dielectric mirror can have 99.9999% reflectivity now at particular chosen wavelengths, so a perfect mirror for weapons-strength lasers at TL10 seems quite feasible.

So you stick it on a plane. The plane flies through a layer of water vapour and the mirror picks up condensation. Oops. Or dust, or exhaust fumes, or…

I'm not saying this can't be done, but it's not at all easy.

[ericthered]

Originally Posted by Johnny1A.2

That'll depend on a few things. How powerful the belligerent that wants to do the overflight will be a big factor, as well as how powerful their opponent is and how likely they are to strike at them while in the neutral country's territory. Geography will be a factor, too.

If you're as big and powerful as the belligerents, you can say 'no'. If you're not...well, then things may be different..

World War I in particular saw a lot of that. Belgium is the most noted example. but probably of more note are Iran and Greece. Greece was a battle ground for 2 years before it legally joined a side, and underwent a civil war while troops from 8 or more nations (UK, France, Russia, Italy, Serbia, Bulgaria, Germany, and Austria Hungary) were conducting military operations in it.

Originally Posted by RogerBW

It's politically problematic enough in the real world that a sub-orbital troop transport was studied in 2002 specifically to get round the issue.

Yeah. Airspace is a politically finicky. The hardest part is usually the last leg, because letting someone use your airspace against your neighbor is pretty much an act of war. Most of the time you can get to the country you want, but you can't get quite the angle you want, and flying all the way across enemy territory to reach the far edge is only something you do once you have air supremacy.

Originally Posted by Daigoro

A dielectric mirror can have 99.9999% reflectivity now at particular chosen wavelengths, so a perfect mirror for weapons-strength lasers at TL10 seems quite feasible.

That has the small hiccup of making it possible for your enemies to armor against you with those same mirrors. It takes more armoring but its possible. And leads to a hilariously star-trek like discussion of beam frequencies that penetrate shielding.

[Anthony]

Originally Posted by RogerBW

You run into problems there, because the mirror typically has to be light, but the beam has to burn through an armoured enemy vehicle or missile.

There's dramatic differences in light intensity, though; you could (relatively) easily have a 1 meter mirror that reflects onto a 1 centimeter spot.

[Johnny1A.2]

Originally Posted by Andreas

There not being any AI worth anything like that, is not really plausible. Even today autonomus aircraft are feasible to some extent and it is perfectly clear that computers will continue to significantly improve for at least a while longer. Also, even in a scenario without any computer hardware improvements at all, being able to develop and test software over many decades (or even centuries, slow or no computer hardware improvements are after all likely to slow down tech level progression) would be a huge advantage.

Even a very conservative assumption of the capabilities of future autonomous aircraft should take these factors into account.

That isn't AI in the sense we're talking about. What you're talking about it simply automation. By itself it won't take humans out of the warfare loop. Your automated fighter-plane might be fully capable of carrying out the mission unmanned, but you'd still need either a pilot or a remote-pilot to decide if it should carry out the mission, and deal with changing circumstances.

What we're guessing about is both 'strong' AI in the SFNal sense, where the machine is basically a person itself, or what THS calls NAI, or other things along those lines that can take humans out of the loop totally. We have no idea how long that will take, if it'll happen at all, or what it's limitations will be.

For the thread topic, there's a universe of difference between a plane that can fly itself to the target unsupervised, use its weapons, and fly back...and a plane that can fly itself to the target, match or outthink the living minds trying to find a way to stop it, and above all else, make the judgement calls about whether to fire the weapons at all, when, and how.

Granting the possibility of the latter, then there's the issue of reliability, that is, will the AI system open the pod bay doors when told to, or abort the mission if the President orders it, or not launch the mission of its own initiative.

We're guessing on the reliability issue, too.

[Johnny1A.2]

Originally Posted by RogerBW

So you stick it on a plane. The plane flies through a layer of water vapour and the mirror picks up condensation. Oops. Or dust, or exhaust fumes, or…

I'm not saying this can't be done, but it's not at all easy.

Actually, that suggests a tactic. If ground/sea based lasers make air overflight prohibitively dangerous or impossible, you might see belligerents using bad weather as a 'cover', if it provide enough protection from the lasers downstairs. Fly above the storms, so to speak.

If the plane-mirror trick can be made to work, a counter-tactic would be the mount the big sea-lasers on submarines, which surface to fire, then dive again if things start to turn against them. You might still have missile-subs, but along with them there might also be laser-subs.

[Andreas]

Originally Posted by Johnny1A.2

That isn't AI in the sense we're talking about. What you're talking about it simply automation. By itself it won't take humans out of the warfare loop. Your automated fighter-plane might be fully capable of carrying out the mission unmanned, but you'd still need either a pilot or a remote-pilot to decide if it should carry out the mission, and deal with changing circumstances.

What we're guessing about is both 'strong' AI in the SFNal sense, where the machine is basically a person itself, or what THS calls NAI, or other things along those lines that can take humans out of the loop totally. We have no idea how long that will take, if it'll happen at all, or what it's limitations will be.

For the thread topic, there's a universe of difference between a plane that can fly itself to the target unsupervised, use its weapons, and fly back...and a plane that can fly itself to the target, match or outthink the living minds trying to find a way to stop it, and above all else, make the judgement calls about whether to fire the weapons at all, when, and how.

Granting the possibility of the latter, then there's the issue of reliability, that is, will the AI system open the pod bay doors when told to, or abort the mission if the President orders it, or not launch the mission of its own initiative.

We're guessing on the reliability issue, too.

You don't really need a pilot, remote pilot or strong AI for that. Not only is it already possible to automate such considerations to some extent (such as having the autonomus aircraft automatically cancel the mission if the target doesn't look as expected), we have already proven willing to extensively use weapons which can't make such distinctions (such as cruise missiles).

As for outthinking living minds trying to stop it, those kinds of tactical decisions is not something which you usually need strong AI for. Things like taking efficient evasive maneuvers is well within our means to automate.

The reliability issue is not really more of a problem for such aircraft than many of the currently used missiles. Also reliability of software is fundamentally something which can be reliably solved if you have time for extensive testing. You don't really need any advances in computer hardware or softeware development tools for that.

[lwcamp]

Originally Posted by ericthered

That has the small hiccup of making it possible for your enemies to armor against you with those same mirrors. It takes more armoring but its possible. And leads to a hilariously star-trek like discussion of beam frequencies that penetrate shielding.

The main problem with this is that
[1] Dielectric mirrors work best when the light is coming from a certain direction. So which direction is it you want to armor against?
[2] You can use big mirrors to direct the light while the beam is still wide and of low intensity. Then the mirror focuses the beam to a much smaller and higher intensity spot on the target, of a high enough intensity that the mirror fails.

Luke

[Johnny1A.2]

Originally Posted by Andreas

You don't really need a pilot, remote pilot or strong AI for that. Not only is it already possible to automate such considerations to some extent (such as having the autonomus aircraft automatically cancel the mission if the target doesn't look as expected), we have already proven willing to extensively use weapons which can't make such distinctions (such as cruise missiles).

Yes, you do need strong AI, or a human, for that. Cruise missiles are not analogous to planes or ships, they are analogous to bullets.

As for outthinking living minds trying to stop it, those kinds of tactical decisions is not something which you usually need strong AI for. Things like taking efficient evasive maneuvers is well within our means to automate.

Which helps how? Outthinking the foe is about far more than evasive maneuvers...and even there blind automation is soon going to come up short against minds.

The reliability issue is not really more of a problem for such aircraft than many of the currently used missiles. Also reliability of software is fundamentally something which can be reliably solved if you have time for extensive testing. You don't really need any advances in computer hardware or softeware development tools for that.

Lacking them, you still have to have pilots, remote-control operators, or humans in the loop, or you will lose the war.

The reliability issue is about whether you can trust a machine that can think and want and decide for itself to do what you want rather than what it wants.