Human input in Trek spaceship battles
- Thread starter Anton Dolinsnky
- Start date
Sorry if I misrepresented you, I just thought it was an odd thing to say.
That really isn't a very good analogy. A better analogy would be a parent using genetic engineering to make his kid superior in every aspect to his natural parents, ensuring that he had the best genes, the best senses, the greatest intelligence, the greatest confidence.[/QUOTE]
I'm trying to figure out how that is not awesome.
At any rate, people practice eugenics every time they choose to conceive, or not to conceive. Genetic engineering is the next step forward in the control of our reproduction, but not at all the beginning of it. Genetic engineering just gives us better tools to do what we've been doing since time immemorial--trying to make the most successful offspring we can, so that our genes persist all the way down the line.
Well, technically, wholesale genetic engineering wouldn't, and spells the end of biology-based replicators as the fittest lifeforms on our planet. What genetic engineering--and to a greater degree AI--offer is a chance for our ideals and knowledge to carry on into the future, because eventually our DNA won't--heck, in the cosmic long run, can't.
Timo,
You can think something through and be wrong...
Myasischev,
Don't worry, I forgive you
I understand using gene therapy or genetic modification (with a person's consent of course) to treat hereditary disorders, birth-defects, and such. But I don't think it's a good idea to start trying to use genetic engineering to try and go around breeding genetic superhumans.
Nothing lasts forever...
You can think something through and be wrong...
Myasischev,
Sorry if I misrepresented you, I just thought it was an odd thing to say.
Don't worry, I forgive you
I understand using gene therapy or genetic modification (with a person's consent of course) to treat hereditary disorders, birth-defects, and such. But I don't think it's a good idea to start trying to use genetic engineering to try and go around breeding genetic superhumans.
Nothing lasts forever...
The important part is that I spent the better part of a year talking to experts in the field and paying actual money to get my hands on real research journals on the subject just so I would know what I was talking about. There's something to be said for genuine research as opposed to "I saw 'Terminator' on DVD the other night and this is what I think will happen."
ProtoAvatar
Fleet Captain
About human/AI input in spaceship battles.
This depends heavily on the type of AI available.
If AIs are smarter then humans, they won't need humans at all.
If the Ai is as smart/creative as a human and thinks 1000 times faster, then the AI should control the ship at in all situations in which time is of the essence - the AI will think all situations and options through, unlike humans who will have to rely on the moment's inspiration.
If the AI is not as smart/creative as humans but thinks 1000 times faster, then this AI should control all more or less conventional reactions/actions of the ship in battle/other emergency situations. Humans should be able to propose/command a certain tactic, though.
The same can be said about any other potential advantage an AI could have over a human brain:
If the exact memorization of a large volume of information is needed in a situation, then an AI should completely replace humans in that situation, if it's as smart as a human. If it's below humans, the AI should perform the functions that require the memorised information, and humans should do the rest.
Etc.
This depends heavily on the type of AI available.
If AIs are smarter then humans, they won't need humans at all.
If the Ai is as smart/creative as a human and thinks 1000 times faster, then the AI should control the ship at in all situations in which time is of the essence - the AI will think all situations and options through, unlike humans who will have to rely on the moment's inspiration.
If the AI is not as smart/creative as humans but thinks 1000 times faster, then this AI should control all more or less conventional reactions/actions of the ship in battle/other emergency situations. Humans should be able to propose/command a certain tactic, though.
The same can be said about any other potential advantage an AI could have over a human brain:
If the exact memorization of a large volume of information is needed in a situation, then an AI should completely replace humans in that situation, if it's as smart as a human. If it's below humans, the AI should perform the functions that require the memorised information, and humans should do the rest.
Etc.
What do you think all that "attack pattern alpha" business is about? The Conn's entire job is probably to program--and keep track of--a limited set of pre-determined maneuvers and call the right ones into action for the right situation. Otherwise, to be blunt, there just aren't enough buttons on a helm station to realistically pilot a galaxy class starship, especially in combat.
And then there's the extremely verbose and frantic business on the Defiant when Eddington fried their computers; would would normally be done with two or three keystrokes took five minutes of manually calling out thruster sequences and system checks.
I believe there is always going to be a human (sapient) being's finger on the button. If not a firing button, then something labeled 'enable' or 'weapons release'. It would not seem to be a part of the federations core philosophy to have a computer make the judgment about ending intelligent life. The machine would figure angles, speed, power requirements, a being would be required to make the ethical decision to destroy and kill.
ProtoAvatar
Fleet Captain
The problem is - the AIs from trek were shown to think 1000s of time faster than humans.
Their role, as shown, is too limited. There should be no need for a weapons officer or for a helm operator during a fight; indeed, using organics for this should be contraindicated (human/organic reaction speed is to slow for this) - the computer should take over those functions.
The captain's role should be only to propose/order the execution of innovative tactics/strategies to the AI.
Newtype Alpha,
You stated in a previous post that the BGM-109 used a large amount of A.I. in its design to enable it to navigate itself around, find and attack its target.
How would you say the A.I. on the BGM-109 would compare to the actual intelligence capabilities of say...
- An insect
- A fish
- A frog
- A crocodile
- A hawk/eagle/seagull
- A dog or cat
- An orangutang
- A chimpanzee
- A human
You stated in a previous post that the BGM-109 used a large amount of A.I. in its design to enable it to navigate itself around, find and attack its target.
How would you say the A.I. on the BGM-109 would compare to the actual intelligence capabilities of say...
- An insect
- A fish
- A frog
- A crocodile
- A hawk/eagle/seagull
- A dog or cat
- An orangutang
- A chimpanzee
- A human
Newtype Alpha,
You wrote in a previous reply that the BGM-109 Tomahawk used A.I. in it's design to enable it to locate and attack it's target.
How would you say the A.I. compares to the actual intelligence capacities of creatures such as
- Insects
- Fish
- Amphibians
- Crocodiles
- A hawk/eagle/seagull
- A dog or cat
- An orangutang
- A chimp
You wrote in a previous reply that the BGM-109 Tomahawk used A.I. in it's design to enable it to locate and attack it's target.
How would you say the A.I. compares to the actual intelligence capacities of creatures such as
- Insects
- Fish
- Amphibians
- Crocodiles
- A hawk/eagle/seagull
- A dog or cat
- An orangutang
- A chimp
None of the above, since the BGM-109 is somewhat better at NAVIGATING than just about all of these things, but isn't really smart enough to do anything else.
That's the thing about intelligence: it isn't an all-or-nothing proposition. A cruise missile that can find and attack a particular building in a particular city somewhere along a thousand-kilometer flight path can do this job better and more reliably than most natural organisms; this is because machines are really good at doing simple-minded, well-defined tasks and repeating those tasks the same way every single time. The more things you program the AI to do, the "smarter" it becomes, but in a variety of operations AIs are already thousands of times faster and more precise than humans.
In space, you could easily program a starship to execute a series of attack maneuvers against another vessel given sensor inputs of the target's relative velocity, distance, heading and status. That kind of programming is painfully simple; video game designers have done it for decades. In this case, a game AI placed in control of an actual starship could easily outmaneuver and out-fight another starship with a human crew (as was effectively proven in The Ultimate Computer) but it would lack the ability to do more subtle things--thinking on its feet, as it were--that humans are good at.
Even in air combat, it's an inconvenient but widely recognized truth that unmanned aircraft COULD be controlled entirely by AI if someone so chose. The main reasons nobody's done it yet are political and operational, not practical.
...There are some practical obstacles, too. A Nintendo control chip hooked to yer average F-15 could probably win air fights with eight pins tied behind its back, but it would have real trouble landing the plane afterwards. Air is not a benign environment, and controlling an aircraft with precision is a demanding job that calls for some "fuzziness" and, instead of mere rapid reactions, rapid choices of reactions. Air combat doesn't require much precision, but takeoffs and landings do, which is why the attrition rate of drones is pretty horrific today: remote pilots and AIs alike keep crashing them at landings when they, more or less literally, don't feel the wind through the seat of their pants.
Automated naval combat should be a breeze in comparison. Then again, it doesn't require split-second reactions as much as it requires calm reasoning, clever ploys, and good control of the strategy and politics of the situation.
Then again, automated space combat should be even easier. Orbital mechanics are trivially easy to calculate compared with aerodynamics, and space is a benign environment... But one might argue that Trek space combat is somewhere between the real deal and the classic naval engagement, with lots of leeway for slow but serious thinking as opposed to rapid reacting. The very existence of shields helps a lot: first hit doesn't kill, like it would in virtually any realistic space combat scenario imaginable. Even the tenth hit leaves you time to ponder your countermove.
Timo Saloniemi
Automated naval combat should be a breeze in comparison. Then again, it doesn't require split-second reactions as much as it requires calm reasoning, clever ploys, and good control of the strategy and politics of the situation.
Then again, automated space combat should be even easier. Orbital mechanics are trivially easy to calculate compared with aerodynamics, and space is a benign environment... But one might argue that Trek space combat is somewhere between the real deal and the classic naval engagement, with lots of leeway for slow but serious thinking as opposed to rapid reacting. The very existence of shields helps a lot: first hit doesn't kill, like it would in virtually any realistic space combat scenario imaginable. Even the tenth hit leaves you time to ponder your countermove.
Timo Saloniemi
Modern flight AIs are being programmed to do exactly this. AFAIK, the X-47 prototype is currently being tested on full-automatic landings on aircraft carriers.
Airfield landings are supposed to be easier; I've heard from a few sources--but am unable to confirm--that the autopilot of the Airbus A380 is capable of landing the aircraft even in bad weather.
The F/A-18 can already do the takeoffs.
Believe it or not, both the Soviets and the U.S.A.F. experimented with different configurations of armed orbital space platforms. The Soviet conclusion with the Almaz experiments was that the human element really wasn't needed at all, since anything along the lines of ship-to-ship combat could be handled entirely by computer (even computers of 1970s era technology) or ground control input.
Space exploration has similar features, considering the extent to which we use probes to reconnoiter Mars and the outer solar system. But here, unlike combat, is alot more "fuzziness" in the task of exploration and AIs would only suffice for investigating very specific types of information as a prelude to manned exploration.
Answering questions like "Is the atmosphere breathable" or "Is that gaseous dichironium or are you just happy to see me?"
Quite so; it's what I'd call a "remaining obstacle", along with certain issues on maintaining contact with independently operating air combat robots and guiding them through contested and congested airspace.
To be sure, the single cannon on the Almaz stations wasn't a "combat" weapon as much as it was a "demolition" weapon. It couldn't be steered much without steering the entire station, and was basically only useful in "combat" against unarmed opponents, that is, hunting down and destroying helpless enemy spacecraft that couldn't even maneuver out of the way. But had the gun been mounted in a manner that would have allowed it to actually defend the station against incoming threats, total computer control would probably have been the only way to go. No fear of friendly fire incidents, either...
In space offense, one doesn't need split-second reflexes today, as there is virtually no space defense. That in turn may be because space defense does require those split-second reflexes, which today couldn't be provided even by a computer (because situational awareness in space is close to zip - the radars and lidars that might be used for the purpose aren't up there yet, and might not be all that easy to build).
Timo Saloniemi
I'm not sure I'd entirely agree. What happens if the computer system is damaged in some way? Why should the captain be expected to pick up all the slack if the system can't work right? The computer may indeed have better response time than an organic crew member, but if we're mainly talking in terms of "pre-programmed" tactics that's the only place the advantage comes in. More realistically, with the captain always trying to be innovate or resourceful, he's better off with fellow organics unless the computer is really advanced.
Newtype Alpha
Other than navigation, how would you compare it to the following organisms I listed in the previous post?
Newtype Alpha
I thought the F/A-18E/F's could do this too (if not the earlier F/A-18A/B/C/D's)
You do know commercial airliners have had auto-land capability for decades, and they're specifically designed for landing in poor-visibility conditions.
There are cross-wind limits and such, of course, and maybe the A-380 is better in that respect, I don't know the specifics.
Correct
Depending on how powerful the shields were.
Yes, I do recall reading about some of those.
What were the Almaz Experiments?
The reason you'd want a human in the loop is not because of the capability of the system, it's in the case a problem occurs, a person can take over and control everything in the manner desired.
Even in Star Trek they used space-probes...
You mean that episode with the blood-sucking cloud?
CuttingEdge100
Other than navigation, how would you compare it to the following organisms I listed in the previous post?
Newtype Alpha
I thought the F/A-18E/F's could do this too (if not the earlier F/A-18A/B/C/D's)
You do know commercial airliners have had auto-land capability for decades, and they're specifically designed for landing in poor-visibility conditions.
There are cross-wind limits and such, of course, and maybe the A-380 is better in that respect, I don't know the specifics.
Correct
Depending on how powerful the shields were.
Yes, I do recall reading about some of those.
What were the Almaz Experiments?
The reason you'd want a human in the loop is not because of the capability of the system, it's in the case a problem occurs, a person can take over and control everything in the manner desired.
Even in Star Trek they used space-probes...
You mean that episode with the blood-sucking cloud?
CuttingEdge100
You mean other than the one and only criteria that means anything? That's a little like asking "other than size, how would you compare boulder to a rock?"
Guidance computers are single-purpose computers: they do one thing, and they do it EXCEEDINGLY well. Biological systems are general purpose computers: they do a million things, and do all of them just well enough to avoid being killed 51% of the time. In the end, any task that you could train an animal to do, you could program a machine to do that task better; the advantage of animals is that they can be trained to perform various tasks whose exact parameters are not well understood, which is a way of adapting a general purpose computer to narrow down on a smaller number of priorities.
The MOST GENERALIST of all organisms is, in fact, human beings. That's not saying much except that humans can do just about anything reasonably well. It also means humans cannot do anything EXCEPTIONALLY well, because we are generalists and can only focus our talents so sharply. This is why--task for task--an AI will always be able to outperform a living organism on whatever task it was designed to do, simply because the AI--unlike the organism--is optimized for that specific task and can perform it in the most efficient and most effective way.
I don't think they can land automatically, but it is a feature the Navy has been working on for future upgrades.
I wasn't aware that auto-landing systems could land in bad weather, though (not just bad visibility, I mean high crosswinds and rain specifically). This is why I'm having trouble confirming this from aviation sources; there's some confusion, apparently, as to whether it's a newfound ability to land in crappy weather or the ability to land AT ALL.
Similar threads
