Hacker News new | past | comments | ask | show | jobs | submit login

Don't worry, it will happen. Just like how we developed ABC weapons and it took the world wars for us to realize that we are insane and have to try to take our meds and calm the hell down; so will we have to go through a phase of robotic warfare, which hopefully does not end up like the Terminator story line, before we realize that we're kind of stupid and need to knock it the hell off for a while.



It isn't clear how this ends, though. Robots are going to enable a lot of anti-social behaviours that were previously limited by personal risk during the execution and the odds of getting caught.

People tend to worry about "robot armies" because we aren't very good at imagining the future, and I'm sure we'll see those. We already are in terms of drones.

But a much bigger issue as costs drop and availability increases is going to be freelancers. This is going to be particularly problematic because robots, like computers, are going to be too useful to limit very much in civilian applications. Any dangerous, unpleasant, ugly job is going to have a robot for that.

Including assassination and robbing banks. Although there will also be robot security guards to counter such activities.

Nor do robots on the battlefield have any of the issues that limited CBN (Chemical/Biological/Nuclear... which is the way I learned it) weapons. None of those are enormously useful in actual combat. Chemical and biological weapons are difficult to control and deploy, and the "Davy Crockett Nuclear Rifle" aside, nuclear weapons are difficult to use for anything other than killing cities, which is rarely a strategically or tactically good thing to do. I'm not saying they're useless, merely that they aren't the be-all-end-all.

Military robots, though, pretty much are. There is every reason to expect that they will be used very widely, just like drones. Once upon a time drones were surveillance only. Now it's routine to weaponize them. Soon enough it'll be routine to send them out on autonomous interdiction missions: the economic temptation is just too high. Why send troops to cut off an area when you can have a 24/7 patrol of small, cheap drones that'll kill anything larger than a rabbit that tries to cross a virtual Maginot Line?

I'm not saying we shouldn't do any of this--robots are too useful to not develop. I am saying that the robot apocalypse, when it comes, will look a lot weirder than our 20th century fantasies.


  Why send troops to cut off an area when you can have a 
  24/7 patrol of small, cheap drones that'll 
  kill anything larger than a rabbit ...?
That sounds like an automated smart minefield, which can be mobile. I'm sure we will see that soon, because the usefulness of such things (if they work) will be too hard to pass up.

I hope that Philip K. Dick's Second Variety [1] is not the way that ends up, though.

1: http://en.wikipedia.org/wiki/Second_Variety


Given that the only time they were used in a war it was directed at cities, many people believe that nuclear weaponry is cumbersome, strategic, and massive in effect. This is hype; barring terrorism, the most likely use of nukes in the future are will be attacking data/ energy infrastructure on the surface and in space. One does not need multiple megatons to render whole satellite constellations useless from a relatively long distance. Nukes, while sexy on film as city leveling fire pillars, are at their most useful as a disparity leveler in terms of force count.

This is all conjecture, of course. I think/hope nukes really come into their own as mining and propulsion resources.


The positive part is that along the way they will have to solve a number of very tough engineering problems to be able to realistically perform the tasks you outline. For example to my knowledge the efficiency of robotic movement is still far below that of humans, the same goes obviously for image recognition, autonomous planning, energy storage, etc..

I think as long as the US manages to scare the american public into funding those huge military projects, because they need to be protected from someone, be it communists or terrorists, and not actually use the results all that much for military purposes the world at large will be fine and eventually benefit tremendously.


This is significantly overhyped. Terrorists can already do huge amounts of data anonymously. E.g. putting bombs in the mail, dumping poison in the water supply, etc. And the vast majority of bank robbers are not that well planned or organized.

Plus you can just turn on a cellphone jammer and protect the entire area from robots.


They will send a robot to steal my bitcoin wallet?


Once perfected I can see the going into firefighting and policing duties. For the most part,aside from employment issues, this will result in positives -fewer people hurt fighting fires, more people rescued from previously inaccessible places and robots can brute force people into arrest while being shot at -without having to shoot back or pre-emptively --again to people's benefit


Don't be so pessimistic. I'm sure we can use these kinds of robots for our benefit --not just sending them to war for us.


Since it strangely difficult to look up, I'll leave this here for reference:

The acronym ABC stands for Atomic, Biological and Chemical. It's a cold-war term that has since been updated twice — first to NBC (nuclear), and then to CBRN (adding radiological). http://en.wikipedia.org/wiki/CBRN_defense


and now CBRNE (explosives)


<sarcasm> Soon to be CBRNEP (pointy)

Because clearly explosives or pointy sticks are just as bad as nuclear or biological weapons. </sarcasm>

PS: I thougt about saying heavy or fast, but that can still be scary. http://en.wikipedia.org/wiki/Kinetic_bombardment



A http://en.wikipedia.org/wiki/http:/Father_of_All_Bombs is less dangerous than you might think. Lots of energy sure but it’s poorly directed and there is no poison afterward.

Cluster bombs’ are generally a bigger threat vs people. (ex: football stadium) Bunker busters are more useful vs. hardened structures. (Command center)

I have heard a suggestion that the world trade center would have likely survived a hit from a Father of all Bombs. Though the building would likely need to be condemned afterward.

PS: Aparently they are vary useful for cave systems though.


Orginally Weapons of Mass Destruction was specifically defined under US and International law (by treaty) as NBC Weapons only (Nuclear, Biological, Chemical.) Gradually, the definition has become a bit wobbly, with some attempting to include any weapon that could be used in a terror attack (domestic law enforcement) and those seeking to include portions of their enemies arsenals (e.g. SCUD missiles) under the term WMD.

"In the Butler Review, the official UK report on intelligence on Iraqi weapons of mass destruction, the definition of WMDs included Ballistic missiles with a range greater than 150 kilometres and related major parts, and repair and production facilities.

Thankfully the term WMD has fallen out of favor, but the point the Butler Review makes clear is that it is an inherently Orwellian term. Its Orwellian not in the sense of dystopian, but rather that it is a deliberately ambiguous term used to obfuscate the underlying reality.

It is the politician's version of "hook up." Completely descriptive while so ambiguous that it could be describing anything at all.

Creating terms like WMD, CBRNE, NBC, only serves to undermine the underlying weapon. Like most people, I believe ordinary explosive ordinance deserves far less attention than rogue nuclear weapons.


>before we realize that we're kind of stupid and need to knock it the hell off for a while.

There's nothing stupid about mechanized warfare. Why waste soldiers when cheap robotic soldiers are feasible? If anything, its just a natural advancement of warfare. If we don't do it, the Chinese and Russians will and that will give them an advantage over the democratic nations of the world, and as we have seen both nations are either in the middle of annexations or seem to be planning more.

The real question, to me, is why are we bothering with the human form-factor? A insect-like robot or something with wheels or tank treads would outperform legs. Ideally, a robot soldier should be able to move in a way that outperforms people easily. A bit like how the robots in Elysium are so much stronger and faster than humans.


"If we don't do it, the Chinese and Russians will and that will give them an advantage over the democratic nations of the world"...

I apologize, but I'm now reading this comment in the voice of Gen. "Buck" Turgidson from Dr. Strangelove.

"Mr President, we cannot afford a roomba gap!"


>why are we bothering with the human form-factor?

it's a human world. doorways, windows, cars, sidewalks, elevators

> A insect-like robot or something with wheels or tank treads would outperform legs

task specific robots will exist and outperform for the predetermined task, but the focus is on replacing humans and being able to complete human tasks, whatever the future task may be.

open doors designed for a human? open a window like a human? climb a tree like a human? Check the mailbox like a human? burn a grilled cheese like a human? Do ______ like a human? rather than build for a task, build to match the form that we know can complete these tasks, and viola! you've replaced the need for a human without designing for and testing all tasks that could ever be requested of it


I understand your point, but a spider-like robot could run up stairs and go through windows and doors better than a human-like one. Or a four legged robot with a lot center of gravity and twice the traction. Or novel designs like the robots from Interstellar.

I mean, our drones aren't in the shape of men using gliders.


> I mean, our drones aren't in the shape of men using gliders.

They are in the shape of manned aeroplanes, though -- they need to land on carriers and runways, they need to refuel, and they sometimes need to have weapons mounted on them.

I agree that we can generally improve on the human form in some cases -- "eyes at the front of a head with limited rotation" is an easy fix, for example -- but most things are going to be loaded with tradeoffs, not just functionally but in how their public use might become accepted.


What's the best form factor to be able to go up stairs, or places humans typically go, and to then recover a body from a building or similar?

That seems like it would be legs + arms in some arrangement.

Powerful wheels could perhaps get you up the stairs (arms + legs can scale and climb, wheels don't do so well there), but legs can fold, and a robot could in theory fold its legs to a significantly smaller size to fit through various spaces. You can't fold large'ish wheels very well without completely disassembling the structure.


Probably a spider-like robot, with 8 arms/legs would be the ideal. I could also see a dog-like robot with four legs which give twice the traction with a low center of gravity as well. Or novel designs, like what we saw in the Interstellar movie.


Agreed on form factor. But I struggle with the idea that making kill bots are going to a) save lives and/or b) be seen as a positive advance by any one in le future.

Drones make us far more enemies than they kill. US soldiers have gone the way of the US cop (all utility belt; no substance) and the casualty counts being low is a bad thing. It is the equivalent of a person to high to receive signal from the nerves in their extremities telling them to quit whatever activity is causing the pain.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: