gun
New Member
Posts: 21
|
Post by gun on Nov 19, 2017 8:11:31 GMT
The bottlenecks on "AI" are two (or perhaps three fold): 1. The psychoneuro people still do not have a fucking clue what "intelligence" is in terms of the actual neural networks and organelles of nervous systems. I retired a couple years ago after 25 years in academia doing research in a cross-disciplinary position that straddles these fields, so while I wouldn't claim I'm completely up-to-date, I am almost certainly more up-to-date than 97% of "computer science people" and 99% of AI dreamers in general. Until the psych and bio people have some idea of the actual mechanisms of "intelligence" any prospect of the engineering people "creating" it is quite literally impossible. Bit like engineers creating "artificial compassion" or "artificial conscience" or "artificial pride!" 2. Even once the BASIC mechanisms of nervous systems and bodies which "comprise" intelligence are understood at a level that can inform computer science, there is no guarantee that any of the fundamental design elements of existing computers will be sufficient to "replicate" those biological mechanisms. Modern computers depend on a binary information system. It is almost assured that biological systems do no such thing, so adapting existing computer technology to try to create artificial intelligence could be more analagous to "adapting" pottery making technology for creating warp drives. 3. Lets suppose that the mechanisms of actual biological "intelligence" eventually get ironed out (they will eventually, I'd say no longer than 10,000 years, maybe only 1,000 . . . heck it could even occur in 500!) and the necessary adjustments to create proper cybernetic hardware and software are achieved pretty readily, the proverbial "libraries" of perception, analysis, decision making and reward-seeking/punishment-avoidance which are common to even one species will likely take the biological researchers another generation or three to sort out as they relate to the hardware/software designs. I'm a biomedical researcher and also studied AI at university. In the 2010's, there's been a boom in application of AI, but we are no closer to creating anything like how laymen imagine AI. Marketing buzzwords like AI, evolutionary algorithms, genetic programming, and deep learning get thrown around a lot in tech, especially among app developers, today, but there hasn’t really been any corresponding breakthrough in AI research, as all of these AI applications may suggest: it’s only machine learning. A less impressive term would be “optimisation algorithms”. In essence, the only “breakthrough” has been in starting to apply the ability of computers to make endless trial and error, but no logical inferences are involved, which is what I would associate with intelligence – anything but trial and error. Regarding points 2 & 3, they're valid (but I wouldn't expect "perfect AI" to take us more than 1000 years), but regarding point 1, it sounds like saying we can't harness gravity without understanding "what it is", not only "how it works". I don't think there's anything deeper to understand here: what intelligence really "is" will never have an answer. It's an emergent property of a bunch of neurons and high neuroplasticity and how it works is what it is.
The problem with your argument is, you don't need a human level AI to operate a war machine. You just collect previous battle data, put them in a computer and make it learn from simulations and battles. Deep Blue style, but more advanced. Certainly possible within maybe 50 years. Ever heard of ALPHA? It is an artificial intelligence that is often tested against Air Force veterans in simulations, and ALPHA seems to get slowly the upper hand.. I haven't read a lot about Alpha, but AI winning in any simulation today is only possible if the AI has a certain amount of omniscience and ALPHA most likely has perfect knowledge over all factors that possible could have any effect on it, as opposed to simulated senses. It's not necessarily more impressive than how there are bots that can outplay players in most video games, but developers intentionally dumb down the AI as to not make them impossible.
Police are the only thing that would qualify as anything close to infantry, but I think that would be akin to saying cars are technically the same as tanks. Cars and tanks: all vehicles are "cavalry", in the same context as "infantry" is used in this thread. I only intended "extra vehicular", or "personal" combat and not to imply any institutions. Stealth in space is possible.. just really hard. I'd recommend Reading a certain blog post from matterbeam with the same name. I haven't read too much about stealth in space yet, but it seems to me that people simply say "wouldn't work, because reasons", then shrug and move on. I'm personally not convinced about the whole "omniscient probes" scenario. Has someone actually simulated this in KSP, to figure out how many probes you would need to cover the solar system, how good they would have to be, how expensive, how long it would take to establish such a network, how it would have to be maintained, what use you could possibly get out of it in peace-time, how easily it would be erected in war-time, and how easily these probes could be destroyed? I figure if someone could put a probe in place, someone else could put it out of place a lot cheaper. Because without this "omniscient" network of space probes, I don't see why you couldn't simply have a big shield on one side of your spaceship, facing your enemy and hiding the ship. Why would anyone have a 30kt of energy per second weapon handy at this point? You seem to be assuming that militarization of space will proceed in advance of financial exploitation of space. Why on Orbit would anyone pay to militarize space before there was money to be made there? First of all, the scenarios you've been describing were definitely not about early days of solar system exploration. If there are manned ore transports, space have already been exploited for quite some time. Even more so if there are destinations beyond low Earth orbit to send shipments. Even unmanned regular orbital shipments bound for Earth or Lunar orbits indicate that space financial exploitation have made a lot of progress. As for militarization of space, it has long since happened. ASAT missiles can deal quite handily with any satellite or spacecraft in low Earth and some can presumably reach geostationary orbit. These capabilities will likely advance ahead of commercial exploitation because if a company can launch a spacecraft to mine asteroids, so can the government, including rival governments. A cannon has also been tested in space, and there are a few satellites designed to take down other satellites, like Polyot series. That system was thoroughly tested in seventies and eighties and has been on combat duty till 1993. Additionally, right now any space capability is tightly controlled by the government in all nations that have them. For example, SpaceX cannot hire anybody who isn't a US citizen. This attitude will very likely persist for the foreseeable future, especially if commercial exploitation of space resources begins in earnest. So, no, militarization of space has happened long before space piracy will have a chance to become a thing.Point is, I don't see any reason to believe there will be World War II-style battleships in space before there are 1st century galleys. In other words, there will be not reason what so ever for any nation to spend quadrillions of dollars developing a space warfleet, when there's no war and nobody to fight. By the time to first crooks start "stealing cars" in style with the early 20th century "Mafia", there won't be modern military police rolling in with ARs, armoured SWAT cars and tear gas grenades. In Japan, small groups of bandits were using firearms before the big armies employed firearms and cannons en masse, to my knowledge. 500 years from now might be too much to predict with any certainty; OA has humanity in the nanotech age by then with at the least above-normal human intelligent AI. If we're talking about 50 years from now, or even a 100, the average body - or cyborg body at least - might be prevalent; and electronic warfare and communications warfare might be so proficient to limit teleoperation or telecommunication with drones or so. And then we have to ask what is to fight for? Cyborg body in 2117? *doubt* 2067 is too near-future for anything described in this thread to happen, for sure. Think more 2267 or 2567.
So let's circle back. Firstly, and I cannot stress this enough, space is already militarized and literally everything that is both in orbit and has dV is a goddamn kinetic weapon. Have y'all ever played vanilla Kerbal Space Program? Literally everything is a goddamn missile that can be rammed into anything else. The part about space piracy using cold gas missiles placed by hydrogen steamers is as crazy as it is crazy expensive. What space pirates actually do is hack anything that is remote controlled and has dV to use as a weapon to ransom vessels in transit with the threat of attack. And back to infantry; the only infantry I can forsee are actual swashbuckling space pirates. And they don't want the cargo, they want the ship, as undamaged as possible. And that means they are the only group of people who actually want to room clear their target's habitat module. They can't irradiated it with a nuke, because then the ship glows and is worthless; they can't blast it and totally wreck the habitat module, that's where all the computers and controls to operate the vessel are located and an uncontrollable ship is worthless. The cargo on a cargo ship is peanuts compared to the fission reactor, reaction mass, reaction tanks, radiators, and MPD, and will probably be spaced ASAP just to increase accelleration and dV. An actual ship possesses all the resources actually needed to continue to live in space without resupply or access to the precision spare parts needed to repair and replace all the shit that could break down in the deep black. And furthermore, a ship or ship parts are worth far more than their mass in gold to anyone else in space. There's always a ready buyer for ship parts who either dosen't want to or dosen't care to ask if a part is legal, they need reaction mass now, or a reactor now, or a radiator now, or would you rather wait another 3 months for an air scrubber system or whatever from Mars to arrive? TLDR: anything in space with dV is a weapon, and pirates are hackers who both repurpose orbital craft with dV as missiles to extort money and are literally the only people who would want to have an infantry incursion in your habitat module so they can rid their new spaceship of all this unnecessary meat cluttering it up. Most stuff in space today can't survive re-entry though, right? If you tried dropping the ISS on a major city, it would simply fizzle. I agree about spaceships being high-value targets for any kind of space aggression or piracy though. Especially in the Shattered Horizon scenario. I figure it's a bit like capturing ships instead of burning them down in the Total War games. And before spaceships are stuffed with point defense weapons (which would need serious motivation: we don't see coilguns and railguns lining the ISS today), it might not be that much harder to do.
|
|
|
Post by bigbombr on Nov 19, 2017 10:20:09 GMT
Stealth in space is possible.. just really hard. I'd recommend Reading a certain blog post from matterbeam with the same name. I haven't read too much about stealth in space yet, but it seems to me that people simply say "wouldn't work, because reasons", then shrug and move on. I'm personally not convinced about the whole "omniscient probes" scenario. Has someone actually simulated this in KSP, to figure out how many probes you would need to cover the solar system, how good they would have to be, how expensive, how long it would take to establish such a network, how it would have to be maintained, what use you could possibly get out of it in peace-time, how easily it would be erected in war-time, and how easily these probes could be destroyed? I figure if someone could put a probe in place, someone else could put it out of place a lot cheaper. Because without this "omniscient" network of space probes, I don't see why you couldn't simply have a big shield on one side of your spaceship, facing your enemy and hiding the ship. About stealth in space: link to the mentioned blog post. IR stealth is possible but has significant drawbacks.
And as soon as there is significant space transportation and/or infrastructure, there is a need to regulate and monitor it, the same way as with air traffic. You don't want an aircraft crashing into your city, and you don't want a spaceship smashing into satellites and space stations.
So how hard would intercepting small, plentiful sensor platforms be? Keep in mind that the vast majority of these would be civilian, and many nationalities/corporations would have their own. So you can't attack all of them willy-nilly without a good reason/probable cause. And space is big, taking down sensor networks build up over decades will take a while.Point is, I don't see any reason to believe there will be World War II-style battleships in space before there are 1st century galleys. In other words, there will be not reason what so ever for any nation to spend quadrillions of dollars developing a space warfleet, when there's no war and nobody to fight. By the time to first crooks start "stealing cars" in style with the early 20th century "Mafia", there won't be modern military police rolling in with ARs, armoured SWAT cars and tear gas grenades. In Japan, small groups of bandits were using firearms before the big armies employed firearms and cannons en masse, to my knowledge. Space combat will likely emerge as an extension from planet-side combat (ASAT missiles, recon/communication satellites, the occasional K-rod or laser platform) but gradually might develop into interplanetary conflict. But it would never resemble WW II style naval combat the same way modern combat doesn't resemble phalanx combat from 200 BCE. It's more likely to revolve around diffuse and redundant sensor- and weapon platforms. It will never resemble anything like "1st century galleys". Space is a different environment. On Earth, computers are expensive, humans are cheap. So low-tech pirates are a thing. In space, humans are more expensive than computers. So no manned rockets with machine guns slapped to them holding cargo ships ransom. Space isn't forgiving for cobbled together equipment. I expect little to no space pirates for the same reasons I expect little to no underwater pirates (narco subs don't have an equivalent in space because stealth in space is hard, and makes spaceflight even more expensive than it already is).
Most stuff in space today can't survive re-entry though, right? If you tried dropping the ISS on a major city, it would simply fizzle. I agree about spaceships being high-value targets for any kind of space aggression or piracy though. Especially in the Shattered Horizon scenario. I figure it's a bit like capturing ships instead of burning them down in the Total War games. And before spaceships are stuffed with point defense weapons (which would need serious motivation: we don't see coilguns and railguns lining the ISS today), it might not be that much harder to do. Good luck trying to board the ISS if they don't want to get boarded. If they rig one of those robotic astronauts or EVA suit propulsion units with remote control, they can smash your approaching boarding craft to pieces. If accelerated to a few hundred (or even dozen) m/s, random debris makes for sufficiently lethal projectiles against current spacecraft.
|
|
|
Post by Enderminion on Nov 19, 2017 18:44:16 GMT
a network of solar polar orbital sats can see "down" on the plane of the solar system, you now only have to scan a fraction of the sky. a full sky scan takes four hours
|
|
|
Post by omnipotentvoid on Nov 20, 2017 15:32:04 GMT
[...] Police are the only thing that would qualify as anything close to infantry, but I think that would be akin to saying cars are technically the same as tanks. Cars and tanks: all vehicles are "cavalry", in the same context as "infantry" is used in this thread. I only intended "extra vehicular", or "personal" combat and not to imply any institutions. [...] Thats hardly a good way to think about this. By this definition, anyone with a weapon intended for a fight, but not in a vehicle, is infantry. It means that, if I were to carry a knife with the aim to use it in self defence if necesarry, I would be considered "infantry". Further more, would a bouncer at a club, armed with a taser, not also be infantry? The context in which you use infantry in this thread is so loose, that it makes discussion practically impossible. I readily agree that a discussion about infantry in space should be about more than just the millitary branch of infantry, but it should not be as wide as personal combat. Also, you misinterpreted my analogy in the same sense. A car is to a tank (perhaps millitary wheeled vehicle would be a better analogy) what infantry is to armed personel in generall. Not all armed individuals are necessarily combatants (like competition shooters) and not all combatants are necessarily armed. Police are armed and trained very differently (this is true even for special units, like anti terror units) compared to millitary units. This is because the goals and limitations of the combat environment are generally very different from actuall combat. A very good example of this in modern terms are expanding projectiles. They are banned from military use, but are highly prevelant in law enforcement and are even mandatory for hunting in some cases. Another good comparision is police marksman rifles vs. millitary marksmans rifles. The reason, then, why I would exclude police from this discussion, is because the way a police force operates, and thus, the technology it uses, are more dependant on the legal, ethical and economical systems that surround it, rather than the necessities of combat. Predicting the relevant legal, ethical and economical systems is basically impossible and doesn't allow for a generall discussion (as any arbitrary technologies and tactics could be used by selecting apropriat systems). For this reason, I think a discussion about infantry in space should focus mainly on those combatants, for whom the primary motivation behind technology and tactics is what is possible (and viable to achieve their goals), rather than what is permitted (while the millitary is regulated in use of technology, they are only restricted in technologies and tactics, that cause significantly more harm to combatants and noncombatants than others for only little benefit).
|
|
|
Post by Enderminion on Nov 20, 2017 17:45:59 GMT
Infantry in open space is not gonna be a thing, on planets though it relies on Ortillery either not being used or people being stupid
|
|
|
Post by newageofpower on Nov 21, 2017 3:41:06 GMT
Note that even with empty hands, Mikasa is definitely military grade firepower. I'm not sure if we'll count her as Infantry, though...
|
|
|
Post by EshaNas on Nov 21, 2017 15:36:37 GMT
500 years from now might be too much to predict with any certainty; OA has humanity in the nanotech age by then with at the least above-normal human intelligent AI. If we're talking about 50 years from now, or even a 100, the average body - or cyborg body at least - might be prevalent; and electronic warfare and communications warfare might be so proficient to limit teleoperation or telecommunication with drones or so. And then we have to ask what is to fight for? Cyborg body in 2117? *doubt* 2067 is too near-future for anything described in this thread to happen, for sure. Think more 2267 or 2567.
We already have bioelectric-mechanical prosthetic tech; along with humanoid robots. I doubt a full body cyborg is that far off. The main problem is not engineering the body but powering it: both the servomotors, their batteries, the biological hormones and energy converters to keep a person at a standard level of activity and stability. But that stands for a lot of things, our battery and energy tech is plateauing while our needs continue to rise.
|
|
|
Post by Enderminion on Nov 21, 2017 15:43:14 GMT
Torso Nuclear Generator
|
|
|
Post by tangentialthreat on Jan 2, 2018 8:05:06 GMT
Neutrons kill everything.
You could threaten to turn on your gigawatt NTR near an enemy ship. Surrender now or everything dies of neutrons. Also lithium-6 foil is unrealistically effective in-game.
You irradiate the fanatics and you also irradiate their machines. Sophisticated circuitry and sensors are corrupted at the atomic level. The mechanics of this are rather opaque (even by CDE standards which is part of why it's not fully in the game yet) but radiation strongy favors older chips where individual components are larger. In the real world it took many, many attempts to get an exploration robot into Fukushima without losing processors or cameras.
Neutrons don't stick around that long either. You'll spawn some activation products but in a few hours or days at the most it will be safe to board. Beware mechanical traps, and malicious but stupid steakknife roombas controlled by 1980s chips. Radiation-hardened murderbots are hopefully only a small part of the target ship's mass fraction and you'll win eventually.
The downside is that many of the ship's systems are dead, but the air handler might have been programmed for revenge anyway so cooking its brains was a good thing. You just get a hull, some fuel, and a lot of scrap computers and high-nitrogen fertilizer.
|
|
|
Post by Kerr on Jan 2, 2018 8:35:52 GMT
Neutrons kill everything. You could threaten to turn on your gigawatt NTR near an enemy ship. Surrender now or everything dies of neutrons. Also lithium-6 foil is unrealistically effective in-game. You irradiate the fanatics and you also irradiate their machines. Sophisticated circuitry and sensors are corrupted at the atomic level. The mechanics of this are rather opaque (even by CDE standards which is part of why it's not fully in the game yet) but radiation strongy favors older chips where individual components are larger. In the real world it took many, many attempts to get an exploration robot into Fukushima without losing processors or cameras. Neutrons don't stick around that long either. You'll spawn some activation products but in a few hours or days at the most it will be safe to board. Beware mechanical traps, and malicious but stupid steakknife roombas controlled by 1980s chips. Radiation-hardened murderbots are hopefully only a small part of the target ship's mass fraction and you'll win eventually. The downside is that many of the ship's systems are dead, but the air handler might have been programmed for revenge anyway so cooking its brains was a good thing. You just get a hull, some fuel, and a lot of scrap computers and high-nitrogen fertilizer. or just use missiles armed with neutron bombs.
|
|
|
Post by tangentialthreat on Jan 2, 2018 16:43:50 GMT
I wish those were in the game too, and I believe nukes would be terrifyingly cheap with no price controls, but the best death-threats are ones that cost you pennies to implement. Now the hard part would be hostage rescue or capturing a specific humanoid alive. You have a few minutes between exposure to space and permanent brain-death, so you might just start ripping the ship apart with explosives and atomic-powered anime cyborgs to rip the walls off. I've played enough games with SWAT in the title to be somewhat optimistic about breeching and clearing a room before everybody without a suit dies, especially if you have futuristic murder-bots helping you. Anyone you want to keep goes in a large Ziploc baggie, like a Walmart goldfish. What I don't see happening is capturing both the ship and the crew fully intact. I also like to think that the people saying MAD and logistics would stop sane people from doing much conflict in space are correct, although making humans stop trying to destroy eachother will probably require a couple of paradigm shifts and some significant edits to what it means to be human.
|
|
|
Post by newageofpower on Jan 2, 2018 18:19:33 GMT
I also like to think that the people saying MAD and logistics would stop sane people from doing much conflict in space are correct, although making humans stop trying to destroy eachother will probably require a couple of paradigm shifts and some significant edits to what it means to be human. Even very smart humans are not always rational. Plus, necessity breeds innovation, and conflict creates necessity at amazing pressures. War tends to drive progress at an incredible pace.
|
|
|
Post by thorneel on Jan 2, 2018 21:52:58 GMT
Now the hard part would be hostage rescue or capturing a specific humanoid alive. You have a few minutes between exposure to space and permanent brain-death, so you might just start ripping the ship apart with explosives and atomic-powered anime cyborgs to rip the walls off. I've played enough games with SWAT in the title to be somewhat optimistic about breeching and clearing a room before everybody without a suit dies, especially if you have futuristic murder-bots helping you. Anyone you want to keep goes in a large Ziploc baggie, like a Walmart goldfish. Meh, spray everyone with neutrons to incapacitate them, pull them out of the now brittle, radioactive husk and give them anti-radiation medicines. You do have anti-radiation medicines, right? You wouldn't turn enemy combatants into "walking ghosts" that can only hope to kill you (and your civilians) as much as possible before dying a horrible death, right?
|
|
|
Post by Enderminion on Jan 3, 2018 0:16:38 GMT
Now the hard part would be hostage rescue or capturing a specific humanoid alive. You have a few minutes between exposure to space and permanent brain-death, so you might just start ripping the ship apart with explosives and atomic-powered anime cyborgs to rip the walls off. I've played enough games with SWAT in the title to be somewhat optimistic about breeching and clearing a room before everybody without a suit dies, especially if you have futuristic murder-bots helping you. Anyone you want to keep goes in a large Ziploc baggie, like a Walmart goldfish. Meh, spray everyone with neutrons to incapacitate them, pull them out of the now brittle, radioactive husk and give them anti-radiation medicines. You do have anti-radiation medicines, right? You wouldn't turn enemy combatants into "walking ghosts" that can only hope to kill you (and your civilians) as much as possible before dying a horrible death, right? it they're brittle, nope not a chance, restore from backups
|
|