Robot dog with machine gun hints of a dystopian future | Hacker News Reader

Robot dog with machine gun hints of a dystopian future

cybernews.com

29 pointsbmcn202026 comments

It is disquieting for sure, but from the first police robot with a sand bag shotgun and likely even before, we have been on this path. I imagine many people when they first saw "spot" and "big dog" thought to themselves that they were ultimately going to be a weapons platforms. Hell their (Boston Dynamic's) "cheetah" is a 10 times more terrifying than this "Russian demo" in my opinion.

In any case, if that Russian guy's back yard project scares you, you definably don't want to know about Ghost Robotics collaboration with Sword International.

I'm not sure it was ever feasible to prevent this escalation in unmanned warfare.

The Israeli-American firm Duke Robotics has built drones with guns for the better part of a decade. I believe they’ve already been deployed in Israel:

https://www.discovermagazine.com/technology/israeli-military...

Ya, those are the professional version of a second back yard job by an American teen six or seven years back who put a handgun on a cheap quad copter. The remote warfare cat" is as the say "out of the bag" for sure.

I think the only real debate is how autonomous are we willing to be with respect to these platforms.

Is this the video you meant?

https://youtu.be/MBBC-xL_MTg

If my math is correct, the Duke prototype (2015) predates it by a few years. Agreed with your point though, this genie’s out of the bottle.

This is just proof of principle of weaponized AI, but this is absolutely revolutionary. All the pieces are already here now for a Raytheon or some other company to manufacture robots that can conduct offense/defense. Put this in an urban hot zone with enemy detection, millisecond precision targeting, and tactical objectives (defend this entrance, clear that chokepoint), and you've changed the face of warfare overnight. I'd argue building out weaponized AI is an easier problem than self-driving cars. We've already normalized collateral damage in war so this doesn't need to work perfectly to be highly effective.

I agree, however I would like to point out that Collateral Damage has been the norm for all of the history of warfare. The reduction of collateral damage by smart munitions and rules of engagement is an extremely new thing and still a foreign concept to most militaries. It's essentially a luxury afforded to countries so powerful that they can expend resources to not destroy the entire country they are fighting.

Collateral damage hasn't been normalized, in fact the opposite has occurred. Much worse, the idea of a "Clean War" has been normalized. The only clean war is NO war.

Not in my opinion, collateral damage appeared as the fire lethality soared. I'd say the US civil War was probably the first time you could really find collateral victims. Once the fire became as lethal as the shock I'd say

This doesn't really make sense. Genghis khan massacred entire cities. Soldiers don't just go home when the enemy army is routed. Historically, the first thing they want to do is pillage and rape.

I guess to count as collateral damage they would need to at least pretend it was unintentional. Massacring cities that didn't surrender was a deliberate strategy.

Collateral damage before the ages of strategic bombing and artillery usually consisted of stuff like famine, fire, and exposure. Not that any of the militaries of the time really gave a flying fig. War back then was more along the lines of national scale murder rather than the extension of politics by force. Collateral damage is a new concept that originated during Vietnam. Before then it was just "War", "Burn the Heretic", and "Don't hurt the new slaves!"

doesn't handle recoil well but only a matter of time before it's figured out. Pack of robot dogs would be an intimidating sight, maybe the next battlefield innovation after UAV's to have game changing impact. Even 'as is', it would be useful for suppression fire

Cost? £75K for Boston Dynamics Spot, this knock off probably could retail for £10K. Far cheaper than a human infantryman

The recoil issue is heavily exacerbated by the way the weapon is mounted. The barrel and thus recoil force are ~12" above the center of mass, causing the recoil to try to spin the robot rather than just pushing it backwards.

If they mount the rifle upside down or sideways so the recoil is closer to center mass, I think the recoil would have a far lower effect. They could probably change that over fairly quickly.

I also suspect that they'll end up chambered in pistol or intermediate calibers. The ammo is lighter and more compact, as are the firing mechanisms, and I don't know if these things can orient themselves accurately enough to hit something at 100 yards. Something like this that can move fast with an integrated P90-like weapon would be extremely powerful and terrifying.

Actually, the robot in the video appears to retail for US$2700: https://shop.unitree.com/products/unitreeyushutechnologydog-...

You could probably build two of these, small arms included, for that 10K GBP (if you did it on the cheap).

I'm of a mixed mind about robots in warfare. On the one hand, robots instead of people means less potential lives lost, at least on one side. On the other hand, it makes combat less personal, so you don't see or feel the loss and devastation up close, which might lead to more of it if you have little or no direct emotional attachment. If you can just send robots to do your killing then it's ok because they're bad guys and no good guys get hurt. Unfortunately history has shown that it's not so simple to just not fight.

May be a bit unfair, but to me this looks like a gun taped to a robot. Not really impressive except for the shock value on the internet. More aesthetics than function or anything revolutionary.

Ya, and mounted on 12 to 18 inch risers to boot. No wonder recoil is a "bit of an issue". It really looks more like a back yard job done by someone who did not want to damage their personal rifle.

Your vision is too limited.

Priced at 2.7k bucks, you can buy ten or twenty of these for important outposts.

Then while the human army is dug in deep inside some sort of concrete wall and fortification, they send fifteen of these to fight.

While these are easily killable, let's say fifteen takes out two human and injures three more, and they die.

Now the unit in full health and rest comes to play. They already have an advantage. A huge one.

If both sides do not have these, then it is a huge disbalance in power.

Let's just forget robots killing humans. Let's say twenty of these are deployed on a yard. They are fully armed. Humans will not even want to go in front of that wall. Why risk losing the most important asset (a human) to gain an advantage where none of the enemy humans are harmed?

Either both sides should have these, or one side should invest in better EMPs, and other methods. And it should start now.

Also remember, this is the one they are _showing_ you. You can't even fathom what they _have_.

Suspect the usefulness will end up being dictated by energy management concerns. Balancing mobility & runtime. The logistics of swapping batteries / charging every hour, how well you can make them "camp" somewhere & wake up, needing very close "home bases" where they can charge and batteries are stockpiled etc.

> Let's say twenty of these are deployed on a yard.

If they are moving around, you need either a human swapping batteries more or less continuously (~20 battery changes / hour) next to a pile of lithium (maybe 100-200kg for 20 units) and a generator of at least 10-20kw. I guess you would want them to auto-charge off a vehicle or fixed equipment for it to be practical.

Seems like you could make something work but IMHO they seem like less of an obvious "slam dunk" when you consider the fiddlier aspects of operating them.

I don't think there is any suggestion that a platform such as this has no practical uses. I think their point was that this particular implementation looks more like a motivated amateur rather than a professional manufacturer.

The key factor for me is the change in the substrate of production. Rather than a military strength being in many ways bottle-necked by the rate and lag of human creation, it can be rebased on the manufacture rate and costs of robotic sentries. Even if they don't have parity with human soldiers, a dramatic shift in costs and manufacturing logistics may still tip the balance on the calculus.

This is literally just a gun strapped to an RC toy. It has about as much combat utility as a Roomba with a knife on it.

As with all technology, any obvious military use will be taken and exploited to the max. How long do you think it took to exploit sticks, rocks and fire as weaponry? The discovery of making hands into a fist was weaponized way before even then.

There is no technology in history that has ever been exempt from weaponization, and it is unfeasible to think there ever will be. We can't not see an object's weapon potential, for life itself dictates the existence of adversaries.

Remember, this is the one they are _showing_ you. You can't even fathom what they _have_.

Ghost robotics have this: https://www.engadget.com/robot-dog-gun-ghost-robotics-sword-...

There's an episode of Black Mirror with AI killer robot dogs that is quietly terrifying.

This is the episode that got a colleague to quit watching Black Mirror: "too depressing"...

Reply on news.ycombinator.com