Fears of humanoid robots mostly come from science fiction, like the Terminator. Things that are like us, but stronger, faster, smarter, make for frightening villains. It goes way back. Robots, supervillains, demons, demigods, etc. 2/x
-
-
Show this thread
-
But there's no reason killer robots need 2 legs and 2 arms. Skynet creates Terminators for a specific purpose: infiltration. They look like people so they can sneak into human bases, or travel into the past and ask people about Sarah Connor. We have intel operatives for that 3/x
Show this thread -
Hunter-killer ground robots will have four legs, not two. Better balance, and can run a lot faster. Like this one
4/xpic.twitter.com/vkL7mkBpc9
Show this thread -
Whether on treads or legs, weapons-carrying robots will be smaller and lower to the ground than humans. Makes for a smaller target for enemies. (BTW, in tests these models hit targets with perfect accuracy from distance. Unlike human snipers, they don't breathe) 5/xpic.twitter.com/7XWcnhUesw
Show this thread -
And, of course, many killer robots will fly rather than walk or roll. Can move faster, and don't have to worry about ground obstacles. There's the big missile-firing drones you've heard of, which can stay in the air for 24-36 hours. Can be more selective about when to fire. 6/xpic.twitter.com/eBGNq9beut
Show this thread -
And then there's small combat drones, like the Switchblade, a "non-line-of-sight" weapon that carries a grenade-sized charge and can kamikaze into targets. 7/xpic.twitter.com/7Oit2pxHbL
Show this thread -
But if we really get into a war with the machines, the biggest thing we'll have to worry about is probably drone swarms. Humans have better situational awareness than drones, and will for a while (never say never). But robots can coordinate more efficiently. 8/xpic.twitter.com/dTu6LlORst
Show this thread -
Early drone swarms used a mothership to issue commands. But that means if the central drone malfunctions or gets damaged, the whole swarm stops working. Now swarms use "emergent intelligence," like ants or bees. Individuals are dumb, but the group, effectively, is smart. 9/x
Show this thread -
Each member of the drone swarm--like each ant or bee--follows simple rules in relation to each other. Follow the ant in front of you to get food. Don't get too close or too far from another drone in the swarm. From these simple individual rules, group intelligence emerges. 10/x
Show this thread -
Drone swarms have immense military potential, for both surveillance and combat. Also many dangerous implications. The US Navy's already experimenting with a drone swarm called LOCUST. You can see (partial) test videos. 11/xhttps://www.youtube.com/watch?v=qW77hVqux10 …
Show this thread -
Terrorists, insurgents, and criminals will make use of drone swarms as well. In January, a small swarm controlled by Syrian rebels attacked a Russian base. And last year, a gang in Colorado used a drone swarm to obstruct FBI agents as they tried to rescue hostages. 12/x
Show this thread -
So while humanoid robots are often the bad guys in movies--Terminator, Ultron, the Will Smith version of I, Robot--the real killer machines of the future will be various shapes and sizes, with a superhuman ability to coordinate group actions. Feel better now?
13/xShow this thread -
I can hear my publisher
@ibtauris in my head: "You wrote a thread on killer robots and you didn't plug the book? Plug the book!" Ok fine. "Drones and Terrorism"--Covers some of the stuff in this thread and more unsettling awesomeness. Available now. (END) https://www.amazon.com/Drones-Terrorism-Asymmetric-Warfare-Security/dp/1784538302/ref=sr_1_1?ie=UTF8&qid=1522173339&sr=8-1&keywords=drones+terrorism&dpID=61VwSgYN1OL&preST=_SY291_BO1,204,203,200_QL40_&dpSrc=srch …Show this thread
End of conversation
New conversation -
-
-
Great thread. IMHO, I don't think its so much about the 'physical' robots, as it is, the 'AI' that is controlling them. That's really where the 'killer' part is going to come/not come from.
-
Thanks. And I agree. A big bottleneck right now is lack of "explainable AI." With machine learning, computers can teach themselves to get really good at certain tasks, but can't explain their learned decision-making to humans. For chess that's fine. For firing a missile it's not.
-
Have you watched Chris Paine's 'Do You Trust This Computer?' Highly entertaining.
-
I haven't, but I'll check it out. Thanks for the recommendation.
End of conversation
New conversation -
-
-
That's what they WANT you to think.
- End of conversation
New conversation -
-
-
Two questions: 1. Why would robots need to jog? 2. How do these people feel about
@realDonaldTrump taking over? -
1. Because it's cool? Otherwise ¯\_(ツ)_/¯ 2. Also ¯\_(ツ)_/¯
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

