TVTropes Now available in the app store!
Open

Follow TV Tropes

Following

Sci-Fi Weapons, Vehicles and Equipment

Go To

PresidentStalkeyes Eats moldy bread and flies into windows from United Kingdom of England-land Since: Feb, 2016 Relationship Status: Do you like me? (Yes ⎕ Definitely ⎕ Absolutely!!! ⎕)
Eats moldy bread and flies into windows
#13826: Apr 16th 2020 at 12:08:44 PM

Can I just say - I've noticed that discussions of AI taking over combat roles traditionally performed by humans in the future always seem to provoke particularly furious debate around these parts; from a writers' perspective, is this because AI seems 'too good' and we don't want it to be a Game-Breaker that would eliminate potential stories?

Those sell-by-dates won't stop me because I can't read!
Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13827: Apr 16th 2020 at 12:30:27 PM

The reason We Will Use Manual Labor in the Future exists as a trope is because you want human characters for the audience to relate to. It's not nearly as exciting to watch Joe the AI Supervisor as it is Joe the Daring Pilot.

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
EchoingSilence Since: Jun, 2013
#13828: Apr 16th 2020 at 1:10:56 PM

This is also why writers make robot characters so human, easier to sympathize with.

DeMarquis (4 Score & 7 Years Ago)
#13829: Apr 16th 2020 at 3:08:24 PM

Its because people with some knowledge of how computers work get very frustrated with lazy cliched depiction of it you always get from Hollywood.

I'm done trying to sound smart. "Clear" is the new smart.
Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13830: Apr 16th 2020 at 3:23:37 PM

Well, if anyone were competent at programming software, including AI, then a lot of potential stories involving them would be impossible. It's the same reason that most plots exist: someone has to start the conflict, and most of the time that is due to being grossly incompetent, illogically malicious, or both.

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
Draedi Since: Mar, 2019
#13831: Apr 16th 2020 at 4:58:05 PM

Human beings being grossly incompetent and illogically malicious?

Reality Is Unrealistic

Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13832: Apr 17th 2020 at 8:08:33 PM

Well, okay, that's entirely valid, but typically fiction doesn't play on the kind of general laziness and errors that any human can be subject to. It gets magnified and compounded with problems that are completely illogical in order to turn it into a drama rather than an accident.

Think about it this way: in reality, the Space Shuttle Challenger blows up during ascent because the people in charge ignored warnings about a sealing system on the solid rocket boosters that was out of spec for the weather conditions and had already come very close to causing failures on previous missions. In fiction, there's a conspiracy to destroy the U.S. manned space program and someone has a car chase with a spy.

In reality, a self-driving car mistakes a piece of paper blowing around for a jaywalker and stops suddenly, spilling your coffee. In fiction, the car's talking AI gets a worm installed by the sinister tech guy and starts trying to crash into orphanages.

In reality, some people rob a bank, they leave evidence everywhere because they're not the brightest stars in the night sky, the police investigate and catch them, and there's a report on the evening news. In fiction, a half-dozen guys in clown masks blast their way into a bank (a mob bank, folks), murder each other as part of a complex scheme, then the mastermind (who is part of the heist, mind you) gets away in a parade of school buses.

In reality, when a computer breaks, it displays garbage on the screen and/or stops working. In fiction, when a computer breaks, it explodes.

In reality, when an asteroid is about to smash into Earth and you need to blow it up, you train astronauts to use mining equipment. In fiction, you train miners to be astronauts.


Here's another true story. In 2016, a SpaceX Falcon 9 rocket with the AMOS-6 satellite on board suddenly exploded during a routine pre-launch fueling test. SpaceX has been the center of a lot of drama, of course, and in the reporting that followed, people claimed that UFOs did it, or that a sniper (possibly hired by a rival company) shot the rocket. In fiction, one of those things would indeed be true. In reality, there was a completely unanticipated and never-before-seen failure in a helium pressure vessel within the liquid oxygen tank. Engineers identified the problem, proposed solutions, they were implemented, and that was that.

That's boring, though, which is why fiction dramatizes both the nature of the accident and its cause.

In fiction, someone runs breathlessly into the control room with an urgent, last-second warning which the arrogant launch director ignores, while people on the ground run desperately for shelter to avoid the impending disaster.* In reality, there's nobody within 3 km of the launch site because of standard safety protocols and the board remains green up until the instant of the explosion.

*Worth noting that something like this did happen much earlier. A Delta II exploded shortly after takeoff in 1997, and this was in the days when they had ground crew on-site in an armored blockhouse, so they were trapped inside by burning fuel and a number of cars in the parking lot were destroyed. This incident taught NASA to adopt its policy of clearing launch sites of people.


In real life, accidents happen all the time, but they almost always come from very simple mistakes and have relatively straightforward repercussions. My point is that if you want to write a story about a combat AI "going rogue" and murdering civilians, go ahead, but remember that you're writing fiction. Don't get too wound up in the realism of it, because then you have to ask yourself questions, like who programmed it so that it was even possible for it to fail in that way.

I mean, if someone is going to develop a combat AI in the first place, the very first thing they'll do after teaching it to move and shoot (or even before) is make sure it can identify non-combatants. There shouldn't even be a logical pathway that could lead to that happening, and you would perform millions of simulations to confirm this. No one person would have the kind of access to the development cycle needed to slip in "rogue code". (Actually, the first first thing they'd put in the robot is a remote kill switch, which makes this scenario even more absurd, but I digress.)

In fiction, you can ignore this and do what you like. In fiction.

Edited by Fighteer on Apr 18th 2020 at 11:32:43 AM

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
Belisaurius Since: Feb, 2010
#13833: Apr 18th 2020 at 6:20:02 AM

Always thought remote kill switches are something of a liability. Yes, you can make it neigh impossible to trigger from the outside but it's also a massive target for espionage and it's capture could decide the entire war.

It's also not like you can hide the kill codes perfectly. After all, the soviets somehow got their hands on the US's nuclear bomb designs.

Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13834: Apr 18th 2020 at 7:03:10 AM

Almost every rocket launched, including nuclear-tipped ballistic missiles, has a flight termination system that can be triggered by a controller on the ground. This will destroy the vehicle if it goes off course or suffers a malfunction. This system is "safed" (i.e., turned off) once the vehicle reaches its target orbit/trajectory. At that point, absent some minor terminal guidance, it is impossible to steer it onto a significantly different course. Many torpedoes and cruise missiles have similar systems.

(It is worth noting that Russia has not used flight termination systems on many of its rockets, and has paid the price for it on at least one occasion — a spectacular failure of a Proton rocket that hit the ground barely a kilometer away from the launch site and spilled toxic hypergolic propellant everywhere.)

The risk of the enemy learning how to blow up or turn off your weapons is judged to be lower than the risk of those weapons going off course and hitting an unintended target. This is particularly important when the weapon is autonomous and can guide itself without direct human control.

The U.S. employs both robotic and remote-piloted drone aircraft. All such aircraft have self-destruct explosives in the event they lose control or get shot down, to prevent the enemy from reverse-engineering them.

I absolutely guarantee you that any hypothetical AI combat robot would have such a kill switch. The risk of it not having one is far too great.

Edited by Fighteer on Apr 18th 2020 at 11:04:16 AM

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
Jasaiga Since: Jan, 2015
#13835: Apr 18th 2020 at 8:48:48 AM

Do we even need AI like Cortana? Like I don’t see the point other than “hey look at us we can do it,” than just having AI In extremely specialized roles.

MajorTom Since: Dec, 2009
#13836: Apr 18th 2020 at 3:42:46 PM

The difference between extremely specialized AI and a specific program or routine of computer logic that does the same thing is effectively academic.

The reason for AI is the ability to adapt and modify itself based on changing scenarios or information provided. Or so is the theory.

In non-AI software, even if The Dev Team Thinks of Everything the software is effectively rigid, unchanging without updates by the programmers. It will only have a set, finite number of outputs it can do and if something is fed into it that is brand new to that software it'll at best just spit out an (outdated) error message, at worst it'll slag your entire program and force a restart of it potentially jeopardizing data, runtime and more.

With an AI if you fed it an input that was brand new and accurate but unplanned or unaccounted for at the time of programming such as a new discovery in a field of science, the AI could adapt and modify itself to account for that without requiring a programmer's patch.

Of course, sometimes even that can be accounted for in rigid non-AI software.

DeMarquis (4 Score & 7 Years Ago)
#13837: Apr 18th 2020 at 5:51:37 PM

Forget Cortana—''Duplex'' is truly frightening. Here's a demo.

Edited by DeMarquis on Apr 18th 2020 at 8:56:39 AM

I'm done trying to sound smart. "Clear" is the new smart.
TairaMai rollin' on dubs from El Paso Tx Since: Jul, 2011 Relationship Status: Mu
rollin' on dubs
#13838: Apr 19th 2020 at 10:42:57 PM

A friend and I were having a convo about AI back in The '90s. The USAF had already started work with drones.

The problem with any AI in combat is the chaos. The game Harpoon told GM's to hold players accountable for their actions. Many a game had players fire bearing only launch missiles at neutral ships or give Unfriendly Fire to aircraft on their side. In Real Life "Blue on Blue" is still a huge issue, see the death of Pat Tillman.

He told me something that stuck with me, he brought up the Cyberpunk full 'borg Dragoon:

"Made even worse for 'Borgs. Many full body conversions have a human brain as a plug-n-play Wetware CPU. ... One conversion, the Dragoon, combines this trope with And I Must Scream. The cyberware and the drugs keep the thing (barely) controlled. It acts almost like a dumb robot."

— from "Cybernetics Eat Your Soul"

In the game, the Dragoon is so psychotic and Literal-Minded that it's sent into areas with only (or mostly) enemies and given instructions like "kill all enemies and then return."

An early AI would be sent where there would be no question at to what to shoot at: enemy bases and ports, massed formations on the ground or in the air, sent in for recon and told to run if confronted.

Eventually AI would start to get there - the "loyal wingman" following along with human pilots to engage only hostile aircraft. Unmanned ground systems that seek out their counterparts and ignore anything that's no hostile.

Humans have problems with shoot/don't shoot decisions, therefore most militaries would try to limit what an AI can do when it's armed and without a human in the loop.

Even the best system will fire at a bus full of nuns, a neutral aircraft or sink a red cross ship. Humans will do the same (and not just on the table top).

Edited by TairaMai on Apr 20th 2020 at 12:46:16 AM

I tried to walk like an Egyptian and now I need to see a Cairo practor....
EchoingSilence Since: Jun, 2013
#13839: Apr 20th 2020 at 2:01:47 PM

So, murderous killbot soldiers. Whether humanoid, starfish, tin can, or rolling treads... How would you design them and what locomotion systems would they use?

Hard Scifi or Soft Scifi, the purpose here is to build a robotic killing machine that would terrify any enemy forces having to fight it.

Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13840: Apr 20th 2020 at 2:11:34 PM

Well, any practical robotic killing machine would terrify people pretty effectively without having to be specifically designed for it.

Do you have any specific parameters? Are we talking dog-size, man-size, tank-size? I'm not exactly a combat vehicle engineer, but my first thoughts would be something that uses treads for locomotion and has turreted weapons attached to an armored core structure containing the control systems and power supply. Those weapons should independently target, of course.

Just going on coolness factor plus practicality, you'd want a minigun-style weapon for close combat and point defense, a heavy gun for defeating armor, and possibly missiles to kill airborne targets. It's really hard to give specifics, though.

You also have to consider the value of specialization vs. generalization. If you build to the role, you'd have a large variety of robots with different propulsion systems and weapons. If you go general, you'll want to fit as much capability into a single platform as possible.

At one maximum, you have a mega-tank the size of a building that does as much damage crushing things beneath its treads as it does shooting them with missiles, guns, nukes, or whatever. At the other, you have swarms of killer drones zipping around all over the place.

I think it was Elon Musk, or possibly Musk quoting somebody else, who proposed the idea of extraordinarily simple attack drones using an off-the-shelf drone body with a half-pound of explosives strapped to it and an iPhone using facial recognition. Program it with the face of the person you want dead and send a swarm of them into a building. This is obviously over-simplified but the basic concept is sound enough. Forget suicide bombers: suicide bomber drone swarms.

Edited by Fighteer on Apr 20th 2020 at 5:15:06 AM

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
EchoingSilence Since: Jun, 2013
#13841: Apr 20th 2020 at 2:14:15 PM

This robotic soldier must be under 3 meters in height, can be used in urban or outdoor scenarios. Only things I'm gonna specify as I wanna see how imaginative the people here can get.

Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13842: Apr 20th 2020 at 2:15:50 PM

Suicide bomber drone swarms, as I edited into my post above. Absolutely terrifying. Program the image recognition AI with the people you want dead, set them loose in a geo-fenced area that all friendlies are out of, go in twenty minutes later to mop up the remains.

Edited by Fighteer on Apr 20th 2020 at 5:22:39 AM

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
DeMarquis (4 Score & 7 Years Ago)
#13843: Apr 20th 2020 at 3:03:16 PM

Spider designs. A swarm of three meter metal spiders going in and out of buildings will disturb some people.

I'm done trying to sound smart. "Clear" is the new smart.
Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13844: Apr 20th 2020 at 3:13:07 PM

A swarm of three-meter metal spiders going in and out of buildings will disturb most people.

FTFY

Edited by Fighteer on Apr 20th 2020 at 6:27:59 AM

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
TuefelHundenIV Night Clerk of the Apocalypse from Doomsday Facility Corner Store. Since: Aug, 2009 Relationship Status: I'd need a PowerPoint presentation
Night Clerk of the Apocalypse
#13845: Apr 20th 2020 at 4:08:59 PM

Basically a mini-tank with some form of heavy weapon and sensor suite mounted on it with articulated tracks so it can go upstairs.

Robo-snipers that can fly to a position and sit and wait for a target with systems for image, acoustic, and thermal/IR pattern recognition and a robust customizable target library. Sort of like that Fire Ant Drone only firing a single targeted shot and can fly away on its own after the fact.

Who watches the watchmen?
DeMarquis (4 Score & 7 Years Ago)
#13846: Apr 21st 2020 at 5:12:10 AM

"A swarm of three-meter metal spiders going in and out of buildings will disturb most people."

Imagine them scuttling up and down walls, and across ceilings.

I'm done trying to sound smart. "Clear" is the new smart.
Imca (Veteran)
#13847: Apr 21st 2020 at 5:16:14 AM

TBH small spiders are scarier then big spiders, since the scary aspect is just how many fucking places they can hide in.

MajorTom Since: Dec, 2009
#13848: Apr 21st 2020 at 5:59:10 AM

^ And sometimes the more venomous spiders in an area can be the smaller ones. For example in my neck of the woods, tarantulas and wolf spiders can get bigger than my fist but the most venomous stuff is either the Western Black Widow or the Brown Recluse, both much smaller spiders.

Fighteer Lost in Space from The Time Vortex (Time Abyss) Relationship Status: TV Tropes ruined my love life
Lost in Space
#13849: Apr 26th 2020 at 5:42:40 AM

Reviving this conversation a bit, the problem I have with three-meter metal spiders is that they would not be stable in locomotion. They'd be too heavy to traverse walls without massive claws that would leave gaping holes and possibly destabilize weak structures.

Terrifying, yes, but "swarming up walls", no.

"It's Occam's Shuriken! If the answer is elusive, never rule out ninjas!"
ericshaofangwang Messenger of the Daemon Sultan from the Void between universes Since: Jul, 2017 Relationship Status: Having tea with Cthulhu
Messenger of the Daemon Sultan
#13850: Apr 26th 2020 at 9:47:39 AM

Aye on that one. Something that requires such mobility would need to be a good deal smaller, though that wouldn't be as terrifying. Of course, as said previously, one could always stick explosive payloads on them and watch the fireworks from far away...

Edited by ericshaofangwang on Apr 27th 2020 at 12:48:35 AM

This is the internet. Jokes fly over in private jets, and sarcasm has bullshit stealth technology.

Total posts: 19,725
Top