Hacker News Reader

I'm not sure telepresence is sufficient to win a war. People generally agree that when you have a human soldier standing on every street corner, you've won. I think a telepresence robot on every street corner would not have the same effect.

One reason is that even in war, most people are reluctant to kill enemy soldiers. But few would have much trouble sabotaging robots. So telepresence robots won't cause the civilian population to accept the new regime.

Despite some effort, I have not managed to cause people to form the sort of bond with a telepresence robot that naturally forms with people thrown together.

I think regime changes will be done from the inside more often than not in the future - it's been proven to be more effective (and cheaper) than sending thousands of soldiers and machinery to foreign lands...

Meanwhile, telepresent groups of drones and semi-autonomous turrets would be used to fight any localized resistance and protect important objects.

I think (and hope) we won't see any major conflicts like WW1 and WW2 in the future, since there's really no good reason for one (even Europe finally gets it, hence the EU), unless it's an alien invasion...

If the occupied society is too willing to destroy the guard robots, but would not hurt humans, why not give them both? Post a sign on each bot: "WARNING: if this guard is tampered with, its controlling human will be subject to electroshock." Or wire it up so if one bot is damaged it commands a nearby one to retaliate, either against the original aggressor or a random target.

Telepresence, at least in terms of overall kinetic impact, is going to easily trump bioenhancements.

Combat AI makes much more sense, since they can react faster than a human sitting behind a computer separated by a further thousand miles away. The latency will add up, and reaction speed is going to be a decisive factors between combat robots. Telepresence might makes much more sense if you can decide on a slower time frame such as strategic or tactical decisions but those will be increasingly be made by the AIs or the humans near them.

Agreed. I generalized, but you are entirely correct.

You can't beat the speed of light, or the time it takes nerve impulses to fire Especially once we see armed robots facing other robots, the people will quickly be shuffled out of the mix.

Having said that, my bet is that telepresence still rules to some degree. What you'll see is more strategic control. Just like with HFT, the tactical stuff will simply happen too quickly for meat bags.

It'll be interesting if we start seeing a "war of missiles" for land forces like we've seen play out in the skies. Ground-to-ground anti-robotic ordinance shouldn't be too difficult to produce. It would have a much different signature than other forms of ordinance. If the pattern holds, this would lead to stealth robots, which would be a very interesting development. Hoards of invisible, lethal, ground-based intelligent robots is a pretty freaking scary scenario.

Right now, most drones are fighting non-robotic targets, and doing it with an emphasis on stealth (for example, flying so high that forces on the ground can't see or hear them). This makes latency less of a factor, since the human operators can take their time in selecting targets, and are fairly confident that they won't face any sudden attacks.

As drone-on-drone combat becomes more common, you're definitely correct that more decisions will be made by the drones, without human intervention. I can imagine drones having a bedrock set of commands that don't require human approval (such as, "if another aircraft is shooting at you, drop what you're doing and shoot back"). Humans would still provide overall goals, such as "blow up that bunker", but the drones will have increasing autonomy in how that is carried out.

There's an interesting parallel to the Mars rovers here, precisely because of the massive latency, even though Mars isn't an actively hostile environment. I don't recall the source, but I remember reading about Curiosity having much more autonomy than previous versions. It's given the same bedrock set of commands ("don't drive off cliffs", "don't smash into large rocks", etc), and human operators give it a similar set of high-level goals ("drive towards that weird outcropping over there"). This lets it explore much more efficiently than driving a few feet, waiting on an Mars-Earth-Mars round-trip for instructions, and so on.

"they can react faster than a human sitting behind a computer separated by a further thousand miles away"

Maybe. The problem is identifying friend and foes. An AI might need more time than an actual person. Beyond that, losing a robot is not a big deal compared to losing a person, so the mistakes made in the field (ex. killing an innocent person who happened to get in the way) can be lessened because the fear of death is no longer there for the soldier.

> Beyond that, losing a robot is not a big deal compared to losing a person, so the mistakes made in the field (ex. killing an innocent person who happened to get in the way) can be lessened because the fear of death is no longer there for the soldier.

"Friendly fire" from remote-operated robots would be pretty hard for people to cope with.

'Robots killed my dad, a hero soldier who fought for this country' is a headline no-one wants to read.

FWIW, here are the causes of death of US soldiers in Iraq [1]:

  2573 Hostile action
  2496 IED
  665  Non-combat
  402  Helicopter accident
  361  Vehicle accident
  63   Unknown
  39   other-unknown
  31   Airplane crash

Over 800 are accidents, and IEDs are just enemy robots.

1. http://apps.washingtonpost.com/national/fallen/

I need to rephrase, I guess...

If we put robots in the field and leave our soldier at home / base using telepresence to control the robots, the soldier don't have the immediate fear of death and can take the moment needed to identify friend, bystanders, or foes. Friendly fire wouldn't be much of a problem if none of our soldiers are actually in the combat zone. Robot blowing up, bad but no headlines.

Exactly. "You can't safely return fire, so allow yourself to be killed" is something you can say to a robot, but not to a human.

The point of war these days isn't to kill the enemy, but to convince them to surrender and join your side. Faster combat reactions might kill enemy soldiers faster, but if that were the goal you would just drop bombs on them.

You get people to surrender by being reasonable and helpful until they shoot at you, and then fighting back fairly following the rules of war. You have to make being a soldier in your army seem like a better life than being a soldier in their army. Telepresence and combat AI have some value, but will not achieve the main goals.

I have a hard time seeing this. If one side can field robotic armies, what stops the other (poorer) side from deploying massive jammers (or lots of small ones)?

> You really don't want to be a freedom fighter in the year 2050. Or piss off a major world power.

You would think that would be the case today, with all our computers, fleets of carriers and aircraft, intelligence and satellites... and here we were having trouble with IEDs and roadside bombs in Iraq.

Please. All the engineering in the world may not may a human stop having a heatbeat -- but you are overlooking two factors that you really shouldn't in this day and age:

1) It doesn't matter if your target has a heartbeat when the seven civilians loitering in the area also have heartbeats (and you don't know which one it is) or when your targets heartbeat stops because he just blew himself up in the middle of a market square, and you didn't even knew there were enemies around. Or when your soldiers die because they accidentally kicked over a piece of trash that was wired up to blow an old soviet grenade.

The attitude that all we need is much better equipment is the same attitude the current leadership of the pentagon has -- but it couldn't be further from the truth, because compared with the weapons the iraq fighters and the Taliban currently have, our weapons are already much better.

What we really need is much better intelligence.

I don't disagree, but I think you're having some other conversation about eyeballs on the ground versus more and more expensive weapon systems. As a former marine, this is an argument in which I'm very much inclined to agree with you. God knows we don't need many more B-2 bombers to fight the Taliban.

Having said that, the point here is that massive, inexpensive, intelligent, and mobile systems will work alongside troops on the ground more and more as warfighting evolves. This will happen much the same as the dog, horse, or mule worked alongside fighters, only the robotic versions will continue to increase in functionality.

But in the end that's a process discussion -- how will things advance over time. My point was that just like in all other fields, our helpers will soon become our betters. Won't happen for decades yet, but you can be sure it's coming sooner or later. And at that point, just like every other human endeavor, some pimply-faced kid in a Starbucks can control just as much firepower as a four-star general. It will just be a matter of credentials and security.

Sort of agree. Improving the quality of soldiers is about the worst way to go about making the world a better place. It's like the ambulance and the bottom of the cliff, but worse.

Actually, we need better soliders -- specifically more teams to do raid like the one on Bin Laden.

It is just that we need much better intel to be able to do these kinds of raids.

Extrajudicial killing isn't something that needs to increase. Hunting down and killing those that have hurt and harmed us isn't the best way to deal with those who have hurt and harmed us.

Well, Dick Cheney, along with a large number of other people, does not have a heartbeat anymore. He was implanted a continuous pump.

Not to diverge the conversation, but that is no longer true... http://www.nytimes.com/2012/03/26/us/cheney-heart-transplant...;