[Space] Fighter-to-ship ratio: what is it and why?

[Daigoro]

Was this the thread that was discussing encryption methods and randomness generators?

Apparently, lava lamps help with that-
https://sploid.gizmodo.com/one-of-th...-wa-1820188866

[Johnny1A.2]

Originally Posted by Ulzgoroth

Funnily enough, computers are substantially more capable of handling lots of disparate considerations than humans.

No, computers can't make judgement calls involving disparate considerations, or simple ones, at all. Current-day computers are excellent at processing information, and totally incapable of doing much with it that isn't simply a displaced human choice.

There are many specific things about your scenario (which largely have no place in this discussion since they don't actually exist in space or air combat) that computers may currently have problems assessing well - but the multiplicity is a point for computers over humans, not the other way around.

Not when judgement calls are involved.

As I said, I simply picked an infantry setting for a convenient example, the same limitations and logic apply elsewhere.

Now, if you have true 'strong AI', then computers can make judgement calls, and the question changes to 'reliability, trustworthiness, and loyalty'. But we have no slightest idea of how to create such an entity right now. It's just as much a technological handwave as any other aspect of a space setting.

[Johnny1A.2]

Originally Posted by Rupert

Star Trek didn't originally, and given the power of the weapons and apparent need for large ships to be able to power decent defensive screens, it's always seemed to me to be more comfortable without them.

The reason space fighters became popular in the 1970s is that was about when popular fiction caught up to the change in wet-naval warfare, the shift away from battleships to aircraft carriers. If Trek had originated post about 1975, it probably would have had fighters.

Treknology might actually produce space fighters that made sense, for that matter.

Hollywood esp. likes space fighters for the same reason they like Old West gunfights (which rarely ever happened the way the movies portray) or one-on-one sword fights. It's individual, a one-on-one engagement, it makes for good drama.

[Johnny1A.2]

Originally Posted by warellis

Why do you guys talk so much about fully autonomous craft when we know the weapons firing phase will be kept under the control of humans due tonfears about killer AI?

I have no such confidence, on an open-ended timescale.

[Johnny1A.2]

Originally Posted by Anthony

Sure they are. It's just a very very dumb AI.

No, it's a very dumb automaton.

We run into a problem because 'artificial intelligence' is one of those terms that keeps getting redefined the mean something else. The SFnal sense of it as 'artificial person' is actually what the experts originally tended to mean when they said it, modulo details.

Over time, the professional use of the term morphed into 'expert systems' and specialized programming applications, as the ideal of the SAI predicted by Moravec and others kept receding into the future. (At one point in the 1970s, Moravec confidently predicted human-level AI by 2000.)

Now it's used to mean a mess of different things, even by the professionals, none of which has much to do with actual conscious intelligence.

Which is also why it's unlikely we'll see the human element removed entirely from combat any time soon.

[Ulzgoroth]

Originally Posted by Johnny1A.2

No, computers can't make judgement calls involving disparate considerations, or simple ones, at all. Current-day computers are excellent at processing information, and totally incapable of doing much with it that isn't simply a displaced human choice.



Not when judgement calls are involved.

As I said, I simply picked an infantry setting for a convenient example, the same limitations and logic apply elsewhere.

Now, if you have true 'strong AI', then computers can make judgement calls, and the question changes to 'reliability, trustworthiness, and loyalty'. But we have no slightest idea of how to create such an entity right now. It's just as much a technological handwave as any other aspect of a space setting.

The only definition I can deduce for 'judgement calls' is 'things a computer can't do', which is begging several questions at once.

[sir_pudding]

Originally Posted by Rupert

Star Trek didn't originally, and given the power of the weapons and apparent need for large ships to be able to power decent defensive screens, it's always seemed to me to be more comfortable without them.

Dominion War era "fighters" in Trek are more like patrol or torpedo boats anyway. They are long range, relatively large, and don't have carriers.

[Fred Brackin]

Originally Posted by sir_pudding

Dominion War era "Fighters" in Trek are more like patrol or torpedo boats anyway. They are long range, relatively large, and don't have carriers.

The thing that appears to have made them at all viable was the debelopment of very small photon torpedoes. That gets to the core of the issue. Can a little ship hurt a big ship? If the answer is "yes" then using little ships probably looks like a good idea (though these small attack ships may not be very "fighter" like).

In TOS the answer was a definite "no" and that was the case in Lensman and neither had little attack ships.

[acrosome]

Originally Posted by Johnny1A.2

No, computers can't make judgement calls involving disparate considerations, or simple ones, at all. Current-day computers are excellent at processing information, and totally incapable of doing much with it that isn't simply a displaced human choice.

Actually, no, that's wrong.

Well, first, your "displaced human choice" argument. There is no computer about which you cannot claim that, so it's moot. It's an attempt to define your way to winning the argument. The whole point is that humans (or committees of expert humans) program a series of 'best practices' into the algorithms, and the computers execute them better than any human could. So that's exactly what we're talking about.

More to the point, take medicine as just one example- there have been computers programmed with diagnostic algorithms and they dramatically out-performed human diagnosticians. And, as a surgeon, I can almost see the day where I'll be replaced with a robot; at least right now the robots are more like waldos that I control. But full automation probably won't be widely implemented in my lifetime since it's a much harder problem than just diagnosis; it won't be much longer than that, though. (It actually has already been done, experimentally, in simulated cases using pigs.)

And, of course, as has already been pointed out ALPHA beat the pants off of the human pilots against whom it was tested.

In every field where this has been tried the result has been the same. The computer had better 'judgement' than the human. Or, at least equivalent but much faster judgement than the human, which also has it's benefits in combat.

All you need humans for are decisions that aren't majoritively data-driven. In this argument, that's "war or not war?" A human must be the one to decide when to engage in hostilities, but once that decision is made the computers will be much better at executing the hostilities. In a situations less than total war, yes, clearly humans will have to be rather more involved in deciding political questions, such as "how much collateral damage are we willing to tolerate", "is that freighter a valid target", "etc.

[sir_pudding]

A human pilot is only exercising "displaced choice" too, that is why he needs rules of engagement, and authorization for use of force.