Recently Hangzhou, Zhejiang based Unitree announced a humanoid robot that costs just 39,999 yuan, or 5,900 USD.
It is capable of doing complex movements like hand stands, cartwheels, punching, lying down prone, and standing up again on its own. It weighs 25 Kilos, and is about as tall as a smaller sized human.
I think this is an area we should be paying very special attention to. AI is getting all the hype, but it’s unlikely to have a big effect on the outcome of a war. Being able to mass produce soldiers though? That’s a game changer.
These robots would work in any terrain a human can once water proofed, and could be remotely piloted by human soldiers. Retrofitting them with weapons systems would be simple, and they could have armor plating added on so they could just stand under heavy fire and be fine. You’d need higher caliber rounds to take them out. (Exactly the things that the US is floundering to secure metals to make since China controls so much of the rare earth industry).
I don’t think it’s hyperbole to say that these could be the equivalent of the invention of the musket. If WW3 happens sometime in the next decade i expect the world to be shocked as it becomes clear war will never be the same again. It’s like a countdown has begun where everyday we get closer to the moment one of these is first used in a peer conflict, and an arms race begins. One China already seems to be winning before it even starts.
Very interesting. Some thoughts:
A difference between drones and robot soldiers is that drones are really not new technology. Radio control of various types of flying machines has been around, in one form or another, since World War I; and drones have been used, albeit in a limited function, in World War II and most wars since. What makes them, in the 21st century, suddenly an important weapon is the miniaturization of camera technology (so the drone operator can guide the craft as if he were in the air), and the development of things like FPV control systems. But the remote-controlled flying machine is in fact old and tested technology. Fibre-optic drones are in some ways even simpler, though the system of relaying control signals is exotic; for the concept of a craft connected to its operator by a cable is a reversion to pre-radio technology.
A machine which can walk and balance like a human, on the other hand, is much more complex and “experimental” – the more so if it is equipped with machine learning. People have been trying to build such things since complex clockwork was developed, and probably before, and the result has always been unsatisfactory; simply because walking and balancing on two legs is a much more complicated matter than it appears. The human brain, as probably the world’s most complex and astonishing computer – and the only “machine” which makes tools in its own likeness – manages it all. Whether humanoid robot soldiers are possible depends on whether we have finally been able to turn the corner and make a machine with something like our own balance system and situational awareness, something that can’t easily be tripped or defeated. (I can imagine “tiger traps” of the kind used by the People’s Army of Vietnam during their war with the US imperialists being quite effective; maybe even being built into city streets). Otherwise, something like a small remote-controlled tracked vehicle seems to me more likely.
There is a race to develop swarm technology, so I think like AI robot locusts are more likely than human robots. Or they will just do cyborgs.
Weird how we are just now entering a kind of 1950s sci-fi level of battlefield tech.
I am of the opinion that AI controlled ones arent the way things will go. I think they’ll be remotely piloted by humans perhaps in a sort of VR setup where when the operator moves the machine mimics their movement. Atleast for now. Maybe with some machine learning for targetting assistance at most. But a fully software piloted one seems pretty far off to me. A mixed system could work fine though. Like the machine learning layer translates the human inputs into outputs for the robot making small adjustments where needed. But it wouldn’t have the tactical and combat ability of a human fighter so it doing it alone wouldnt be very effective i don’t think.
So that way most of the actual “thinking” is being done by a human brain somewhere else. Then the machine learning control mechanism just takes that and uses it to know what to do.
There are already drones with AI (mostly for final targeting and perhaps targeting acquisition) and I really doubt there won’t be a massive push for full automation of drones and humanoid bots as soon as it becomes feasible.
In the future, qnything less than fully automated resource extraction, production and war fighting seems like defeat to me.
fun fact, the algorithm used to target facebook ads is the exact same one used to determine targeting criteria for drone strikes during the gulf war - its how you had drone operators targeting civilians for shitting in a certain way as culturally sunnis/shi’ites shit in different ways.
examples of this in a modern context -
https://www.theguardian.com/world/2024/apr/03/israel-gaza-ai-database-hamas-airstrikes
Yes, I can see that too. The AI in that scenario would be kind of like the motions we perform “automatically” once we’ve learned some skill like playing the piano – while the conscious mind is focused on other things like “what interpretation do I want to give this passage,” etc.