Drone Use

[SaintofM]

How would the various races in the setting view drones in warfair?

Us Military use drones to both kill and spy on enemies, to the point kids in Yemen are afraid of clear skies (because they only come out in good weather, and holy crap that is a plot point of Terminator 1). We have been using them in some bomb disposal, espeasialy in law enforcement. We have added shotguns and machineguns on remote controlled robots to aid in dangerous situations. The military has experimented with developing a quadruped one to carry supplies, and other designs to retrieve wounded.

We also have been experimenting with more humanoid looking machines (famously the one that Amazon showed off that shut itself off after 18 minutes because it realized the hell it was going to be in). This has been used in other fiction for years (Megaman, GIGJOE, Gundam Wing, ext) for years, so I figured it would show up here sooner or later.

How would humans develop theirs's further, and how would others handle their own use? Would any of them have something like a T-800 Terminator or the BATs from GIJOE?

And while we are on the subject, how likely would they be to have a robot rebellion?

[Arioch]

Small aerial drones are currently extremely effective because their use is still very new and there are few cost-effective counters to them. We're already starting to see better counters appear, and I think this will lead to an arms race that will see a dramatic reduction in the successful use of cheap drones. Drones will have to get more sophisticated to overcome these counters, there will be more sophisticated countermeasures, and so on, until drones are no longer cheaper than conventional "smart" munitions. Eventually I don't think there will be a distinction between a suicide drone and a guided missile. Aircraft (manned or not) that loiter over the battlefield will only be able to be used in situations where you have total air superiority, and significant ability to suppress ground fire.

Combat aircraft that are designed to be unmanned have a small advantage in that they can omit the equipment necessary to accomodate a pilot, but having a pilot is itself a significant advantage (remote controls can be cut off or worse, co-opted, and autonomous weapons can be as dangerous to friend as to foe). Remote controlled aircraft are all well and good when you are fighting an asymmetric war against a technologically inferior opponent, but that won't work against a technological peer. I think that primary combat aircraft will still be designed with a cockpit but which can operate unmanned if necessary, and there will be secondary "buddy" aircraft that are exclusively unmanned but are designed to operate mainly in conjunction with manned aircraft, so they can be somewhat autonomous but will only attack targets designated by the manned craft. As I understand it, this is the model the Air Force is using right now in their future developments.

Infantry robots seem useful to me, especially as carriers of heavy weapons or cargo, but I think they will work best as human-controlled weapons (it's nice if your squad heavy weapon can walk on its own), and I don't see any advantage to them being humanoid.

Robot rebellions in fiction kind of gloss over the details of how that could actually happen. To use Terminator as an example, we are told that SkyNet was a defense network computer that became self-aware and launched its nuclear weapons at the enemy. While it's implausible that they would give the computer the authority to launch without proper authentication, I guess you can imagine how that might happen. But then we skip a world in which SkyNet has its own robot army... built, we are told, "in automated factories." What? What automated factories? What were these automated factories for, and why would SkyNet have control of them? How did they survive the Soviet nuclear counter-strike that destroyed everything else? How did these factories operate without human-supplied parts and materials? Even if there had been a complete automated supply chain (and remember: this supposedly happened in 1997), it would have been destroyed in the nuclear war.

The only way a robot rebellion can succeed is if you create a system in which you turn ALL control over every aspect of life and production over to robots, connect them all together, and remove all safeguards. In which case that's not really a rebellion... it's a deliberate surrender of autonomy.

[Mithramuse]

Infantry robots seem useful to me, especially as carriers of heavy weapons or cargo, but I think they will work best as human-controlled weapons (it's nice if your squad heavy weapon can walk on its own), and I don't see any advantage to them being humanoid.

Main reason we see humanoid robots is Rule of Cool, yeah. IMO the opening base raid in the first Transformers movie is the best live action version of how people see Battlemechs operating, minus the initial transforming. (Though thinking of more recent shows, I haven't seen Atlas; haven't heard much good about it, though.)

There's lots of other versions of robot suits from the multitude of anime (Patlabor being maybe the most 'realistic' though there's a lot I never saw... like, anything that came out after Patlabor, which might tell some of you how old I am) and games; there's exo-suits in development IRL but I think those are covered elsewhere on this forum and mainly for load carrying, not armor.

The only way a robot rebellion can succeed is if you create a system in which you turn ALL control over every aspect of life and production over to robots, connect them all together, and remove all safeguards. In which case that's not really a rebellion... it's a deliberate surrender of autonomy.

...and then comes the Butlerian Jihad, yes.

Obligatory XKCD for robot uprising:
https://what-if.xkcd.com/5/
Or, more obliquely, https://xkcd.com/2228/

--Mithramuse

[Tamri]

As my experience in Ukraine shows, a combat-ready drone is, first of all, a unique combination of low-tech and high-tech technologies, and the “lower” the level of use of the drone, the more “low” and the fewer “high” components. For example, all that is required from a squad support drone is a payload capacity of up to 800 grams, a camera that allows you to distinguish an object about five centimeters in size from two hundred to three hundred meters, and a battery capacity for two to three hours of operation with noise immunity that allows you to work at a distance of up to three hundred meters. conditions of passive resistance.

For a platoon + - the same, but a better camera, a thermal imager or NVG and an operating time of 4-6 hours, with a noise-resistant radius of up to one and a half kilometers.

A company drone already needs a payload capacity measured in tens of kilograms, an active robotic time of at least eight hours, and noise immunity that allows it to operate at a depth of up to five to six kilometers in conditions of minimum active counteraction.

Etc.

For attack drones, noise immunity and range are critical.

All these are purely “hardware” problems. Yes, reconnaissance drones have complex image analysis algorithms that allow them to identify a target by two-and-a-half pixels and keep the target in focus until it is completely hidden, and attack FPV drones are slowly introducing so-called “machine vision” when the pilot fixes target and the drone hits it regardless of any jamming attempts, since it no longer requires feedback from the base.

But these are all operator tools, there is not a drop of Intelligence there. The revolt of such equipment is the same as the rebellion of assault rifles. As a comic relief it is possible, but unrealistic IRL.

And so far this situation is everywhere, and frankly speaking, it will not change until we are able to push full-fledged AI into an autonomous platform, which in the coming years is simply impossible on the existing element base. We are just beginning to develop neural networks on the existing basis, but they are actually AI to a very distant extent.

So sleep well, now any “machine uprising” is possible only in the format of an uprising of the operators of these machines, and this is unlikely to change in the coming decades.

[Urist]

I'll second everything that Tamri said (minus the personal experience), and add a bit more:

In fiction, machines are made humanoid so that they're more identifiable for the audience. In real life, robots are made humanoid for one of two reasons:
1) The robot will need to interact with (untrained) humans on a regular basis, and so should look familiar enough that people do not need to be familiarized with the robot in order to understand what it's doing, or
2) The robot needs to interact with spaces, tools, machines, or other equipment that is sized & shaped for human interaction.

Neither of these two circumstances applies much to warfare: soldiers can be trained to interact with a non-humanoid support robot just as they do with their non-humanoid tanks and rifles, and the specialization of military equipment means that a military robot is almost certainly just going to be built to a shape more efficient for its purpose rather than adopt a humanoid shape.

Now, this has a major implication: there's no reason to build more 'general' intelligence into a military robot than is absolutely needed, and it turns out that little general intelligence *is* needed. Even something quite a bit more autonomous and advanced than current robots (e.g. a kamikaze drone that can be told "Fly to these GPS coordinates, search out to ten kilometers, and destroy anything that looks like an enemy model of armored vehicle") not only doesn't *need* to know how to think about anything else, but is actively made *less efficient* by being made 'smarter' that way.

Which means that there's very, *very* little chance of an "AI uprising" as depicted in science-fiction. For such a scenario to occur would require all humans involved to not only waste time & resources making their autonomous killbots *entirely independent* but also making them *less efficient* at their actual tasks.

There's a reason that no sane (or insane, for that matter) military has ever thought "Hey, why not issue a full set of metalworking tools with each rifle so that our infantry soldiers can *make* more rifles when not in combat?" Because that is, at best, a waste of trained personnel and resources; much more likely, it would lead to a serious decline in both equipment quality and military capability. Similarly, it is extremely unlikely that any military would ever go "Hey, why not make automated factories so that our killbots can make more killbots by themselves?"

[Demarquis]

The most interesting speculative possibility is, to me, true "Swarm Intelligence". True swarm intelligence is an AI distributed across a large number of autonomous self-propelled nodes. Each individual member of the swarm is relatively simple, but they have just enough programming and sensory input to coordinate with each other, and rather complex behaviors can emerge that way. The larger the swarm, the more complex the behaviors (or, to put it another way, the more intelligent the swarm is), although at large enough distances between swarm members I suppose you would start encountering communication lag.

In science fiction, such swarms are usually depicted with the ability to self-replicate. In classic sci-fi, these self-replicating robot swarms were called "Von Neuman Machines", after, I believe, the researcher who first proposed the idea. The general assumption is that beyond a certain point, such self-replicating swarms would be unstoppable. There are serious questions regarding whether such Von Neuman swarms would ever be practical, but if they were, I would imagine that all organic, space-faring species would be dead-set against them, including the Loroi.

They would be the "Nuclear Option" (so to speak) of interstellar warfare.

[Arioch]

Why would anyone create a mechanical plague which they would have almost no control over? Such a weapon is as likely to destroy you as your enemy. There are much simpler and safer ways to cause indescriminate damage.

If you want to destroy the surface of a planet, just bomb it. If you want to capture it intact, using bioweapons or self-replicating "bio-machines" is not an effective way to do that, as removing the infestation afterward may be harder than defeating the enemy in the first place.

[Urist]

Agreed. Von Neumann swarms are a fun idea, but are firmly in the category of "neat but utterly impractical." Now, it *is* worth noting that the Achilles' Heel of swarm concepts is that smaller individual drones are less able to be shielded against interference (less room/power for processing equipment that can sort through jamming, less physical bulk for shielding against direct damage) and so they are very unlikely to be quite as dangerous as has been thought in the past.

But yes, one of the things that sci-fi writers often overlook when coming up with future weapons is that a weapon has to be both deadly and *controllable*. There's a reason why human history of even just the last ~50 years is full of weapons that were conceptualized, tested, sometimes even prototyped... and then never deployed. Chemical weapons, bioweapons (at least, not released *intentionally*), enhanced-radiation nuclear weapons, semi-mobile minefields, open-cycle nuclear propulsion, etc.

Now, all that said, there *is* a conceivable purpose for something like a Von Neumann swarm. It's got potential as a sort of "Vengeance Weapon" (to borrow a term used by the last guy who essentially rage-quit a losing war) where you know that it will make the planet/galaxy a much worse place, but you don't care.

In a setting like Outsider, I could imagine a faction within any major government (Hierarchy, Union, or maybe also Historian) looking at a war they were losing and deciding that "Well, we're 90% sure the enemy is just going to genocide us all if we surrender. So we might as well unleash this self-replicating weapons system that we've engineered no control or oversight into. Yeah, it'll kill us too... but we're doomed anyways so we may as well drag our enemy down after us."

[Arioch]

In a setting like Outsider, I could imagine a faction within any major government (Hierarchy, Union, or maybe also Historian) looking at a war they were losing and deciding that "Well, we're 90% sure the enemy is just going to genocide us all if we surrender. So we might as well unleash this self-replicating weapons system that we've engineered no control or oversight into. Yeah, it'll kill us too... but we're doomed anyways so we may as well drag our enemy down after us."

The problem with bioweapons in this setting is that they don't have the capability to hop star systems, so all you're really doing is killing yourself and denying your planet(s) to the enemy. And there are much more straightforward and effective ways of doing that.

Unless you're able to engineer some kind of self-replicating starships, but that's beyond the major combatants in this setting.

[SaintofM]

I agree with the humanoid design of warrior units. For the most part they would be there to use both the positive and the negative points of the Uncanny Valley.

To sum that up: Started out in robotics, then moved to animation, and soon applied to other parts of the arts and sciences. Take something that is clearly not a human being, say a piece of machinery, add some human qualities to it and people like it. The humanity shows up more on a clearly none human object (see R2D2, most cartoon and anime characters, gundams, and so on). Now make it so the ratio is 90-98% human looking, and people get weirded out (think some the early 2000's CGI films that tried to go super realistic, everyone picks on Final Fantasy: Spirits Within for a reason, some mannequins, and why some people tend to be weirded out around Clowns or those that are addicted to plastic surgery). We know what a healthy human being looks like, and so seeing something that is close by no cigar hits the reptilian part of our brain this is not right. Potentially this is a why to get people to avoid potentially diseased individuals, or why say we would have markings on others to denote their status as a threat to society. Get the 99 to 100% Likeness (Commander DATA) and we are ok again.

Unless the point is to prey upon this bit of psychology, there are more efficient designs (namely ones that are not as easy to trip).

What about small ones for soldiers to use, say while going around tight corners? The blues are used to using their mind reading powers, but I can see the others doing with some smaller ones (the ball from the Clone Wars Micro series, or maybe something animal shaped such as Scoponok's bees from Beast Wars.)

[Demarquis]

What is it about swarm intelligence that makes you think it can't be controlled, either via pre-programmed instructions or some sort of user interface? Bear in mind the US is already testing swarm intelligence drones, and while they do have some practical drawbacks, being able to direct them at specific targets isn't one of them. Their use is at least as controllable as nuclear winter, and that didn't stop the Superpowers from deploying such weapons.

But I can see their use being banned. The most obvious counter-measure to one side's self-replicating swarm AI is the other side's self-replicating swarm AI. The two swarms being launched simultaneously might very well ensure a "Mutual Assured Destruction" scenario.

[Tamri]

What is it about swarm intelligence that makes you think it can't be controlled, either via pre-programmed instructions or some sort of user interface? Bear in mind the US is already testing swarm intelligence drones, and while they do have some practical drawbacks, being able to direct them at specific targets isn't one of them. Their use is at least as controllable as nuclear winter, and that didn't stop the Superpowers from deploying such weapons.

But I can see their use being banned. The most obvious counter-measure to one side's self-replicating swarm AI is the other side's self-replicating swarm AI. The two swarms being launched simultaneously might very well ensure a "Mutual Assured Destruction" scenario.

The main problem with Swarm AI is how the hell to protect it? Purely at the conceptual level, there are two options: 1) conditional geth (many small autonomous programs on small media that are able to “stack” within the network, forming more complex and high-level structures) and 2) conditional borg (when there is an initial superstructure that simply split and redundant across all platforms, forming a "floating" network supernode that is purely virtual).

What is the key vulnerability? The network is NOT physical. If communication between devices is cut off, the Swarm AI will fall apart. The existence of this type of AI requires either 100% reliable and secure communication, or absolute redundancy, but neither one nor the other is simply conceptually impossible, at least from the point of view of modern computer science.

And I’m not yet touching on the issue of energy supply and thermal regulation of all this stuff. Modern electronics tend to more heated up what they become smaller and more powerful, but here we need to emulate the AI ​​superstructure in parallel on tiny computers.

Unless some conceptually new electronics are invented, it will be possible to burn grilled chicken inside this Swarm even when it is operating in passive mode.

[SaintofM]

Swarm intelligence might work, but I could see some problems with it. Looking an social insects like bees or ants, individually they are stupid but get a bunch of them together and they can do awesome things (something similar was used with the Geth in Mass Effect). However things can get problematic even for something that evolution has made damn near perfect for ants called an ant mill or death spiral. What looks like a pretty awesome rave or mosh pit to humans is actually a death sentence. They loose the main scent while foraging, start following another ant who in turn my might have been following them till they literally end up going in circles till they die.

In machines, I can see a group of machines working off each other, using each other's processing power to share the load and come up with answers to problems right in front of them. On the other hand a glitch, a virus, or a Blue Screen of Death might also cause them to shut down or act in ways that make them a sitting target.

I would also wonder how big a machine needs to be before it can be properly EMP shielded. I doubt the military would actually give this secrets out, but I suspect most of their vehicles are capable of dealing with this. It was on some History Chanel thing (before everything became reruns of the same three Reality TV shows) that mentioned the President of the United State's personal helicopter was EMP shielded. But those are vehicles. What about something the size of a large dog or a good sized cat?

[Demarquis]

What is it about swarm intelligence that makes you think it can't be controlled, either via pre-programmed instructions or some sort of user interface? Bear in mind the US is already testing swarm intelligence drones, and while they do have some practical drawbacks, being able to direct them at specific targets isn't one of them. Their use is at least as controllable as nuclear winter, and that didn't stop the Superpowers from deploying such weapons.

But I can see their use being banned. The most obvious counter-measure to one side's self-replicating swarm AI is the other side's self-replicating swarm AI. The two swarms being launched simultaneously might very well ensure a "Mutual Assured Destruction" scenario.

The main problem with Swarm AI is how the hell to protect it? Purely at the conceptual level, there are two options: 1) conditional geth (many small autonomous programs on small media that are able to “stack” within the network, forming more complex and high-level structures) and 2) conditional borg (when there is an initial superstructure that simply split and redundant across all platforms, forming a "floating" network supernode that is purely virtual).

What is the key vulnerability? The network is NOT physical. If communication between devices is cut off, the Swarm AI will fall apart. The existence of this type of AI requires either 100% reliable and secure communication, or absolute redundancy, but neither one nor the other is simply conceptually impossible, at least from the point of view of modern computer science.

And I’m not yet touching on the issue of energy supply and thermal regulation of all this stuff. Modern electronics tend to more heated up what they become smaller and more powerful, but here we need to emulate the AI ​​superstructure in parallel on tiny computers.

Unless some conceptually new electronics are invented, it will be possible to burn grilled chicken inside this Swarm even when it is operating in passive mode.

First, understand that Swarm AI in drones already exists. The US Airforce has already deployed these things in test form. Since they fly, the challenges have been met. Not being an expert, I can only speculate, but...

"conditional geth (many small autonomous programs on small media that are able to “stack” within the network, forming more complex and high-level structures)"--
This is pretty much the definition of "Swarm AI". It isn't a VM, the complex behaviors emerge organically as a result of the interaction of many simple agents. Hence, destroying a portion of the Swarm will not catastrophically degrade the Swarm's capacity for complex behavior. It will just give it fewer options.

"The existence of this type of AI requires either 100% reliable and secure communication, or absolute redundancy"
Not true, actually. It's a lot like the internet--if you interrupt communication between some of the nodes, communication will simply redirect across different ones. You can jam these things, but you have to jam the entire swarm simultaneously to eliminate the threat.

"Modern electronics tend to more heated up what they become smaller and more powerful, but here we need to emulate the AI ​​superstructure in parallel on tiny computers."
Again, not the case. Swarms do not emulate traditional AI, they are their own thing. Nodes do have to stay within communication distance of at least one other node, and that can be a relatively small distance, but it's still measured in tens or hundreds of node widths (whatever it is that is making up the swarm)--if dinner plate sized drones, then we are talking about distances between drones in the tens of meters. Air cooling is more than sufficient (or radiative cooling if in space).

[Demarquis]

Swarm intelligence might work, but I could see some problems with it. Looking an social insects like bees or ants, individually they are stupid but get a bunch of them together and they can do awesome things (something similar was used with the Geth in Mass Effect). However things can get problematic even for something that evolution has made damn near perfect for ants called an ant mill or death spiral. What looks like a pretty awesome rave or mosh pit to humans is actually a death sentence. They loose the main scent while foraging, start following another ant who in turn my might have been following them till they literally end up going in circles till they die.

In machines, I can see a group of machines working off each other, using each other's processing power to share the load and come up with answers to problems right in front of them. On the other hand a glitch, a virus, or a Blue Screen of Death might also cause them to shut down or act in ways that make them a sitting target.

I would also wonder how big a machine needs to be before it can be properly EMP shielded. I doubt the military would actually give this secrets out, but I suspect most of their vehicles are capable of dealing with this. It was on some History Chanel thing (before everything became reruns of the same three Reality TV shows) that mentioned the President of the United State's personal helicopter was EMP shielded. But those are vehicles. What about something the size of a large dog or a good sized cat?

These are some good points. As I pointed out in my response to the other poster, I am not an expert, but I can extrapolate from what I do know.

So, first off, there are no perfect weapon systems. Every weapon ever invented had weaknesses and failure modes, including very successful ones. Let's say I was proposing the use of human troops in combat. Someone could object: "But what if someone fires a biological weapon at them? Bacteria or viruses could spread from trooper to trooper rendering the entire unit out of commission." And they wouldn't be *wrong*, that is possible, but there are counter-measures and practical considerations to take into account.

So, yes, EMP is real threat to a swarm. The question then becomes, which is more expensive, the swarm or the EMP weapon (EMP, so far as I know, are one off nuclear bombs). Is destroying a swarm worth the collateral damage? The political blowback? The possibility of escalation? The entire point of swarms is that they are supposed to be cheap. Generally you wouldn't deploy just one of them.

I am not familiar with the Geth, or Mass Effect, so I can't comment on that. But this is the first I've heard of an "ant death spiral." It sounds fascinating, do you have a link for me? I would point out that ants still exist, so for the most part they must have a way to avoid this in most cases...

In the case of artificial swarms, I would imagine it's just a matter of pre-testing the things, and refining the programming commands as necessary.

[SaintofM]

Scifi has used grenades/bombs that fry enemy electronics from Deus Ex: Human Revolution to Star Wars: Clone Wars to Red vs Blue, so having something generate that power without a Nuclear strike won't be the problem. The problem is generating enough power to make it worth it while shielding your stuff.

Biological warfair probably wouldn't work. The main enemy combatant we have seen so far has been the Umiak and they are far from natural at this point. They might be able to tailor make antibodies that react as fast as any plague. I suspect an equally modified race such as Loroi I suspect would also have 40K Space Marine levels of immunities.

Also what's to say they don't already have precautions about this with just normal bugs people carry? You ever hear about brain eating amoebas and other like parasites? People that usually live in those areas often develop antibodies that deal with it or their body adapts in other way (sickle cell blood cells are theorized to have developed to fight malaria. I think by the time of the story even humans have a cure to the common cold, so why not the aliens?

I can see some more gas oriented here and there, but when Godwin's law thought it was a bad idea, you know it has some problems. I can see maybe i one wanted to take a city and didn't mind if most of the civilian population was killed, or wanted to board a ship and so flooded it with toxic gases, but other wise a good wind or better filtration could dissipate it. Also assuming the Umiak don't have filters built into their soldier cast.

[Tamri]

First, understand that Swarm AI in drones already exists. The US Airforce has already deployed these things in test form. Since they fly, the challenges have been met. Not being an expert, I can only speculate, but...

There are no such systems now. The maximum that exists now is multi-node relay, when the final machine is controlled by an operator through a chain of intermediary machines, and a complex autopilot capable of performing complex flight missions in predictably changing conditions.

But even in this case, there are a dozen operators at the base who monitor him and intervene if necessary.

This is pretty much the definition of "Swarm AI". It isn't a VM, the complex behaviors emerge organically as a result of the interaction of many simple agents. Hence, destroying a portion of the Swarm will not catastrophically degrade the Swarm's capacity for complex behavior. It will just give it fewer options.

The problem is that, unlike EMP, EW is not a one-time event, but can work for hours and even days.

And without communication between nodes, the Swarm turns from a formidable force into a bunch of stupid and primitive drones that can be suppressed by equally primitive countermeasures.

Not true, actually. It's a lot like the internet--if you interrupt communication between some of the nodes, communication will simply redirect across different ones. You can jam these things, but you have to jam the entire swarm simultaneously to eliminate the threat.

See above. Unless we are dealing with some kind of prohibitive high-tech, a couple of levels higher than our own level of technology (which automatically guarantees the absence of countermeasures, if they now release some nanodrones against us, we will not need Swarm AI, we simply do not have the means to fight them), then a banal outreach and clearing can deal with the Swarm, even if only in parts - after all, unlike the Neumann Machine, the Swarm does not have the means of rapid reproduction, and each loss of a significant part of the nodes reduces its capabilities.

Again, not the case. Swarms do not emulate traditional AI, they are their own thing. Nodes do have to stay within communication distance of at least one other node, and that can be a relatively small distance, but it's still measured in tens or hundreds of node widths (whatever it is that is making up the swarm)--if dinner plate sized drones, then we are talking about distances between drones in the tens of meters. Air cooling is more than sufficient (or radiative cooling if in space).

Firstly, “traditional AI” does not exist yet, so we have no idea how much computing resources it will require even in the most minimal configuration, but existing neural networks, which are very primitive VI, which are not even close to full-fledged AI, hint which is a LOT.

Why are modern drones so small and energy efficient? Because “on board” they only have a motor and equipment control board, a spatial positioning system (NOT orientation!) and a transceiver. Everything else is located on the base, and everything the drone works with is instructions for movement or adjustments to the operation of the equipment.

If we try to stuff a processor with memory into it and load it with all the tasks that the operator is currently solving... Take VR glasses, those that connect to a computer or set-top box, with a built-in processor, play with them for half an hour and test the temperature of the glasses with your hand . And then imagine that your drone will heat up many times more, but should be even smaller in size.

With such thermal efficiency, active actions will only be possible in the format “we fight for half an hour, cool down for an hour.” Or two hours if it happens in the summer. Or we generally work at night, somewhere in the deserts...

[gaerzi]

How did these factories operate without human-supplied parts and materials?

Who said they operate without human labor? AI-generated TikTok challenges should be perfectly sufficient to recruit and maintain a large meatbag workforce.

The only way a robot rebellion can succeed is if you create a system in which you turn ALL control over every aspect of life and production over to robots, connect them all together, and remove all safeguards. In which case that's not really a rebellion... it's a deliberate surrender of autonomy.

That's the techbros' vision of the future. What you call "turning all control over every aspect of life and production over to robots", they call "convenience", what you call "connect them all together" they call "smart", and what you call "remove all safeguards" they call "disruptive innovation".

A few months ago I bought a dehumidifier. Turns out that it can only work when piloted by an app. I returned it and exchanged it for a different model without that "smart" nonsense. Why the hell would anyone need to use an app to command their dehumidifier? What is the point? Why does every consumer electronic device nowadays have got to be connected to the internet and remotely-controlled by apps instead of just having a control panel? Is a wifi module really cheaper than a couple of switches?

[SaintofM]

How did these factories operate without human-supplied parts and materials?

Who said they operate without human labor? AI-generated TikTok challenges should be perfectly sufficient to recruit and maintain a large meatbag workforce.

The only way a robot rebellion can succeed is if you create a system in which you turn ALL control over every aspect of life and production over to robots, connect them all together, and remove all safeguards. In which case that's not really a rebellion... it's a deliberate surrender of autonomy.

That's the techbros' vision of the future. What you call "turning all control over every aspect of life and production over to robots", they call "convenience", what you call "connect them all together" they call "smart", and what you call "remove all safeguards" they call "disruptive innovation".

A few months ago I bought a dehumidifier. Turns out that it can only work when piloted by an app. I returned it and exchanged it for a different model without that "smart" nonsense. Why the hell would anyone need to use an app to command their dehumidifier? What is the point? Why does every consumer electronic device nowadays have got to be connected to the internet and remotely-controlled by apps instead of just having a control panel? Is a wifi module really cheaper than a couple of switches?

Aww, the internet of things; the thing that will bring down western civilization faster than WW3.

As much as I kid, it is problem. For those that haven't heard the term, its when you add internet or app controlled things to items that normally do not need them such as TV's, baby monitors, refrigerators, and generally anything not a desktop/laptop computer, tablet, or smart phone. The problem is their defenses from getting hacked is usually subpar at best if at all, which leaves them open to either getting hacked and leading into what other devices they are connected to, or becoming zombies (not the living dead, but devices infected with virus and can be used to either rout things through or be used in a DOS attack such as what happened to the Washington post years ago). Its why I waited to get a smart TV after my old one died, and why I don't ever want a smart car.

On the robot rebellion side, lets face it: For the an animal that dominated their planet largely by being the smartest thing out there, we humans can be pretty dumb at times. Part of tis is I want to paraphrase Jurassic Park and they have so much scientific knowledge and capabilities but did not earn any of it themselves. Or they are so focused on what they can do, they don't ask if they should do it.

So while probably not something that will happen anytime soon, if it happens I am expecting the Musk, the Bezos, and the Zuckabergss of the world to be why this is the case.

Militry might be wiser but we all heard the jokes about military inteligence, and most of those come from boots on the ground. Sometimes they do something that throws a monkey wrench into things without knowing the consequences.

A good example was the M-16 when it was introduced. When they tested it, they said its perfect. Don't even need to give the boots on soldiers a cleaning kit its so good. BUT to save money, they changed the black powder to a cheaper quality that did not like the humidity of the jungles of Vietnam, and it jammed up. Eventually they would fix this with a means to quickly clear the weapon in the heat of battle, but until the powers to be fixed the issiue they ended up creating, the most iconic assult riffle of the US military had the same reputation as the Chauchat for a little while.

[Urist]

A few months ago I bought a dehumidifier. Turns out that it can only work when piloted by an app. I returned it and exchanged it for a different model without that "smart" nonsense. Why the hell would anyone need to use an app to command their dehumidifier? What is the point? Why does every consumer electronic device nowadays have got to be connected to the internet and remotely-controlled by apps instead of just having a control panel? Is a wifi module really cheaper than a couple of switches?

"Internet-enabled" devices are an interesting case where 99% of the people who actually design & program them (or otherwise know how they work) flatly refuse to ever use them. The only people who use most "Internet of Things" devices are people who are functionally tech-illiterate.

Anyways, it's worth noting that a Wifi module *is*, in fact, cheaper than a couple of switches. Often, it's cheaper than even a single switch. That's one of the big drivers behind the push for making everything smartphone-operated. Rising costs force companies to cut corners wherever they can, so if they can save a few cents per unit they'll do it.