This is only a few days after the massive drone attack in Russia. Only a matter of time until we have drones smart enough to dodge bullets (or at least dodge out of where guns are pointing) while flying at breakneck speeds being controlled by AIs we don't fully understand.
The tech industry is working hard to bring about the Terminator future.
Oh man, can anyone imagine a non-Terminator scenario for this?
Update: I'm not saying people shouldn't develop this, we're never going to squash human curiosity. But when I see this kind of stuff, I'm deeply troubled by how bad actors (state and non-state) will use this.
I hope our security services are working hard on countering these potential threats.
Looks like most of the comments here are about the use as weapons and the possible dangers. I believe "Slaughterbots" is the canonical sci-fi video on the subject, and it appears to be aging pretty well. Unfortunately…
The actual race is also worth watching:
https://www.youtube.com/watch?v=yz2in2eFATE
The speed and flawlessness is quite impressive considering it is being resolved with what I imagine is noisy inertial data and a motion blurred CCD camera.
I remember being blown away by a TED talk were "minimum snap trajectories" are planned for quadcopters to fly through hoops and slots.
It's really cool to see this happening fully autonomously and at such high speed. I wonder if the use of AI means that the approach is fundamentally different, or if it uses the same principle of minimizing snap?
https://www.ted.com/talks/vijay_kumar_robots_that_fly_and_co...
FPVs are the most distinctive element in Russian-Ukrainian war, changing many traditional war dogmas. So far AI in this war wasn't heavily used, FPVs are controlled by humans at both sides. I feel uneasy with more AI tech improving FPVs. It's inevitable, but it'll be utilised by military operations to kill more people, to conduct terror acts, controlled from afar. It allows weak players to have more leverage which is not always a good thing, and certainly disrupts power balance. Think about Al-Qaeda bringing a truck of explosive drones to Washington DC and unleashing them at White House, all autonomous, with cameras guiding the way.
Interestingly, the URL for the embedded youtube video ends with the word "FATE"...
Looks like it had NVIDIA Jetson Orin NX 16 GB. No GPS, Lidar, motion capture so its vision only. 6s battery so 5 incher?
I’ve seen AI beat humans in simulations before, but doing it on a real track with the same hardware is honestly kind of amazing. What surprised me the most is they didn’t use any traditional flight controller. They just let the neural network handle the flying.
I’m really curious how this would perform in messier, less controlled environments.
Nearly two years from comparable to human to beat the best.
https://www.npr.org/2023/08/30/1196777528/an-ai-quadcopter-h...
ELI5? so, presumably if you put this thing in front of any starting gate it can navigate any course of similar gates?
or was it overfitted to this specific course?
This makes me wonder what the "best" vehicle for human racing would be like, if there was no requirement for a human driver.
Let's say your task is to move a human from A to B (by a pre-designated route) as fast as possible. The only conditions are vehicle weight, no outside radio contact and no damaging the road (assume each vehicle goes separately, so e.g. slip stream effects don't matter). You can rely on the human to drive or use AI, you can go on the ground or fly through the air, anything is allowed. What would be the best way to do this?
Let's not forget that this works solely for this particular racing setup. If you change a single gate, the AI they are using would not be able to adapt. Still fascinating, though.
Meanwhile, many defense companies are quietly watching this racing achievement far far away through their palantÃri orb researching who built that autonomous drone.
The sci-fi novel I wrote sacrifices commercial art considerations to portray the future of war more realistically. All the robots are one-shot, one-kill and all action takes place faster than human reaction time. The most humans do is approve battle plans. This is already the case largely in the Ukraine/Russia war, and now we're seeing that human FPV operators will soon be obsolete. Realistic sci-fi movies are going to be not have much action in the future. Things will have to be about careful political maneuvering, like in totalitarian country politics. You can't win against robots unless you are using other robots.
Bright futures for these engineers in the defense industry.
A few questions / thoughts:
1. I didn't see it stated explicitly, but I presume the neural net is on the far end of a radio link somewhere, not running on hardware physically mounted on the drone?
2. After viewing the FPV video on the linked page: how the hell do human pilots even come close to this pace? Insane (even assuming that the video they're seeing is higher quality than what's shown on YouTube – is it?)
3. The control software has access to an IMU. This seems to represent some degree of unfair advantage? I presume the human pilots don't have that – unless the IMU data is somehow overlaid onto their FPV view (but even then, I can't imagine how much practice would be needed to learn to make use of that in realtime).
This technological breakthrough is truly amazing, especially the fact that the drone can fly on an actual race track independently, without relying on human control. It's really cool. But honestly, as AI gets better at doing what we can do, even better in some cases, it makes me a little uneasy. Will there come a day when we truly become redundant, with AI taking over the work?
Since I can't access.
> Flying drones faster will be important for many economic and societal applications, ranging from delivering blood samples and defibrillators in time to finding people in natural disaster scenarios
Ah yes. No mention of the real big use case
I highly recommend Macross Plus for further research this topic: https://www.imdb.com/title/tt2330912/
This feels like a bigger deal than what Carmack is doing with an Atari controller robot.
So the drones in the Slaughterbots short film were depicted to be way too slow.
It looks like the reason it won is due to a human pilot crashing into the obstacle. Not exactly an unfair problem, but worth mentioning that it was close until human mistakes were made.
Similar tech from 2023: https://youtu.be/HGULBBAo5lA?si=D5uCJYAPD9iHA-p7
The drone has a camera and a IMU while the human has only the camera. How big is the advantage there?
So is the processing happening on the drone? Presumably not...
Even R2D2 doesn't pilot the X-wing itself.
Have the team published based on this work yet?
...why are we training skynet again?
Quite cool but this is the beginning of the end I’m afraid
> This is more than just a racing win. The smart lightweight AI that powered the drone could help all kinds of future robots, making them faster, more efficient, and better at...
...killing you, their target.
"Coming to a battlefield near you soon."
I man at this point, given what we know I'm sure someone smart can connect some dots and describe what's inevitable with 99% confidence just in the next year or two in terms of society right?
This is quite cool since past efforts in this direction have usually relied on crutches like outside-in imaging and positioning.
A few details I picked up:
* The drones are a spec drone across the league. It's a fairly large-footprint FPV racing drone (it's a 5" propped drone, but it's very stretched out and quite heavy) with both a Betaflight flight controller and a Jetson Orin NX onboard. Teams were only allowed an IMU and a single forward camera.
* It's unclear to me whether the teams were allowed to bypass the typical Betaflight flight controller which is present on the drone and use direct IMU input and ESC commands from the Jetson, or whether they were sending and receiving commands from the flight controller and relying on its onboard rate stabilization PID loop.
DCL is kind of a weird drone racing league since it's made for TV; it's mostly simulator based with, more recently, only few real events a year. The spec DCL drone isn't very capable compared to the more open-specification drones in racing leagues like MultiGP, in large part to keep the events more spectator friendly. This probably makes it more amenable to AI, which is an interesting side effect.