HN Theater @HNTheaterMonth

Progress takes time, and ambulatory motion is a finicky, delicate nightmare that only seems simple in humans and animals thanks to millions of years of evolution. CAM was created in the 1960s[0], but BigDog[1] and other recent Boston Dynamics quadrupeds that aren't designed for human passengers are much faster[2], but still awkward.

Also remember these are expensive prototypes designed to test things other than speed. Making them faster only guarantees that they break themselves harder when they run into a wall or trip over a cliff or something.

[0]https://www.youtube.com/watch?v=ZMGCFLEYakM

[1]https://www.youtube.com/watch?v=W1czBcnX1Ww

[2]https://www.youtube.com/watch?v=3OKZ_n8QW4w

>I think Neal Stephenson described a skateboard with hundreds of tiny feet, moving faster than wheels on any surface. I want to see that.

I am neither Neal Stephenson nor a physicist but I'm pretty sure the friction caused by the increased surface contact of "hundreds of tiny feet" versus wheels would still make the wheels faster and more efficient. Geckos and things can stick to vertical surfaces and move quickly because they're lightweight. Anything capable of sticking to an arbitrary surface and supporting human weight is going to be slow. And if moving downhill, wheels don't really need to do work - gravity does the work. Something with "hundreds of tiny feet" either has to slide or "walk."

⬐ timonoko

I disagree. Centipedes and other many-legged critters can move very fast. Every leg is just a tiny lightweight thing with its own balanced dynamics, just a part of a wheel basically. First thing that came to my mind is a plastic brush with flexible back. If you move a rollers along the back the bristles extend downwards and move backwards in a walking motion, just like a sea urchin does. Scale that and you have a walking machine moving amazingly fast. Tiny bristles yield and conform to any surface.

⬐ magduf

>Centipedes and other many-legged critters can move very fast.

Can they?

Humans have legs too, but they aren't very fast at all. I can easily outrun any human on my bicycle.

The main advantage for legs is dealing with rough terrain. I can hike over/through extremely rough terrain that I would never dream of taking a bike on. But for speed, nothing beats wheels on a smooth surface. You can't have both.

⬐ Qworg

CWU's "whegs" are much in this vein: http://biorobots.case.edu/projects/whegs/

⬐ cr0sh

I could also see Brook's Genghis platform being extended to "centipede" multiples of legs and probably working well.

https://en.wikipedia.org/wiki/Genghis_Robot

EDIT: fixed mistake - put "Dante" but meant Genghis...

⬐ timonoko

I found such a brush with flexible back (from a broken sauna brush). Amazing discovery: If you move the said roller forward along the back the bristles make the walking movement backwards. Need to make room for the Nobel prize now.

⬐ krapp

You're not taking inertia and friction into account.

Centipedes appear to move very fast because they're small relative to us and lightweight - at scale, they don't move very fast. But if you take a centipede and scale it up to the size of a train (square cube law notwithstanding), it would just be a very slow train-sized centipede. To make it faster, you would have to make its legs longer, because legs are levers.

⬐ MagnumOpus

> Something with "hundreds of tiny feet" either has to slide or "walk."

IIRC the Neal Stephenson version was tiny feet mounted on wheels - not walking or sliding but rolling. But better than rubber pneumatics at evening out curbs or other obstacles.

⬐ dmix

Sounds more like legs than feet...

⬐ PhasmaFelis

For the record, Stephenson's skateboard smartwheels used a bunch of little pads mounted on telescoping spokes instead of a solid wheel. The board had a terrain-scanning laser mounted in the nose, and pulled the spokes in and out so that the wheel was perfectly shaped for the contours of whatever it was about to roll over; you can ride a smartwheel board down stairs as smoothly as you would a ramp. But they were still very much wheels. There was no walking motion involved.

⬐ cr0sh

To me, I envisioned them as being a wheel made with a bunch of shock-absorber leg-like "spokes" with feet.

They seemed similar to old-tech Pedrail wheels:

https://en.wikipedia.org/wiki/Pedrail_wheel

I think something like them has been made - but instead of the spokes ending in "feet" - they instead were encircled by a rubber deformable tread; Tweels are another interesting style of this technology, and simpler mechanically:

https://en.wikipedia.org/wiki/Tweel

> > Meanwhile, many of the tasks that seem most basic to us humans—like running over rough terrain or interpreting body language—are all but impossible for the machines of today and the foreseeable future.

> Wow no.

...wow indeed. No research at all. Often I just read the comments here and skip the article, but that made me go and look to be sure.

For anyone somehow still unaware (BigDog creepiness factor was nearly viral back when I was in college around 2008), BigDog has been handling rough terrain for over a decade: https://www.youtube.com/watch?v=W1czBcnX1Ww

Here's BigDog getting abused to see what it can recover from (including being kicked in the side): https://www.youtube.com/watch?v=4PaTWufUqqU

Here's a running cheetah robot jumping over unexpected obstacles (2015): https://www.youtube.com/watch?v=_luhn7TLfWU

And here's a two-legged robot walking around outside (2017): https://www.youtube.com/watch?v=Is4JZqhAy-M

⬐ PeterisP

And for the "interpreting body language" part I recall seeing demos of transcribing sign language, and also of measuring speaker stress levels from video from all kinds of body factors. If it had some widespread commercial application there's no reason why we wouldn't have machines that read body language better than many humans.

⬐ sharemywin

Don't forget Handle.

https://www.youtube.com/watch?v=-7xvqQeoA8c&t=34s

⬐ throwaway080383

The latest video I saw was a two-legged robot doing a backflip.

⬐ Animats

My "running on rough terrain" video from 1995.[1] On non-flat surfaces, traction control dominates the problem.[2][3] This was before Boston Dynamics. (There's much more that could be done in this area, but there's no market. BD still doesn't do speed changes fast. Their machines start by walking, running or trotting in place and then extend the gait. Humans start by falling forward, for a faster start, and go far off vertical for fast direction changes.)

As for interpreting body language, here's the code on Github.[4]

[1] https://www.youtube.com/watch?v=kc5n0iTw-NU [2] http://animats.com/papers/leggedrun/leggedrun.html [3] http://animats.com/papers/articulated/articulated.html [4] https://github.com/shahqaan/kinect-body-language-analysis

Just because he ruled it out doesn't mean that it's not a factor.

People become attached to lots of inanimate items, and can feel emphathy for robots [1] [2], though perhaps not as strongly as with living creatures. US Military soldiers can develop an emotional connection with bomb disposal robots. [3]

I definitely sympathize with this dog-like robot slipping on ice:

https://youtu.be/W1czBcnX1Ww?t=1m23s

Even though it's clearly a robot with no attempt to make it look or act like a dog.

I can definitely believe that people can become attached, even fall in love with, their robot dogs.

[1] https://www.huffingtonpost.com/entry/human-empathy-for-robot...

[2] http://www.slate.com/blogs/future_tense/2013/05/02/robot_hum...

[3] https://spectrum.ieee.org/automaton/robotics/military-robots...

I'm now hoping to see the golf cart in a video like this one: https://youtu.be/W1czBcnX1Ww?t=1m26s

As easily as you and I:-

Snow: http://youtu.be/W1czBcnX1Ww?t=1m12s

Ice: http://youtu.be/W1czBcnX1Ww?t=1m24s

https://www.youtube.com/watch?v=XBsl3HlB8VE

Insight, creativity, and general intelligence are all different and active areas of research.

There is quite a lot of progress in creativity. Google for 'creative software'.

Insight can mean many things, but Watson can now provide insights into cancer diagnosis.

Consciousness is actually fairly well understood in terms of attention, focus, and other aspects.

Artificial general intelligence is a very active field seeing quite a bit of progress.

Here is a bird that flies with flapping wings: https://www.indiegogo.com/projects/bionic-bird-the-flying-ap... (By the way, that is completely unrelated to AI).

Walking over uneven terrain, (also completely unrelated to artificial general intelligence): https://www.youtube.com/watch?v=uVG4J29JZI0 https://www.youtube.com/watch?v=W1czBcnX1Ww

Deep learning is beyond statistical pattern matching. It does involve both supervised and unsupervised learning. Deep learning is currently the most successful technique, but not the most ambitious approach in AGI.

Google for 'AGI', 'deep learning', 'sparse autoencoder', 'hierarchical hidden Markov model', 'OpenCog', 'spiking neural network', 'Hierarchical Temporal Memory'

On a more flippant level, isn't it pretty obvious that robot ponies are going to be a thing that is possible to have some time in the next century or two? We've already got something that's kind of moving in that direction, and this was a few years ago: http://www.youtube.com/watch?v=W1czBcnX1Ww

In case you haven't seen the videos of some of the precursors to cheetah.

Big Dog: http://youtu.be/W1czBcnX1Ww Littledog: http://youtu.be/CEQlZtCi7IQ

This is slightly less spooky than BigDog:

http://www.youtube.com/watch?v=W1czBcnX1Ww

⬐ hippich

big dog uses RL while this bird robot do not have any learning abilities (at least nothing is said about it) and is just a remote controlled unit.

I find amazing how difficult it is to successfully perform apparently trivial tasks such as walking.

There's more information about the 'LittleDog' on the Boston Dynamics website: http://www.bostondynamics.com/

I was really impressed 2 years ago by their video of the 'BigDog': http://www.youtube.com/watch?v=W1czBcnX1Ww (check out t=40s)

⬐ njharman

I find it amazing HN readers believe walking is a trivial task.

⬐ jrp

The key word being "apparently".

⬐ seldo

I don't even think it's that apparent. It takes us 2 years to learn to walk, even with millions upon millions of years of evolution designing us perfectly to do it.

⬐ cma

>even with millions upon millions of years of evolution designing us perfectly to do it.

Birds can walk on two legs in very little time.

⬐ demallien

Yes, but many four-legged animals can walk mere minutes after having been born, and this is a more valid comparison for the task facing a four-legged robot.

⬐ None

None

⬐ pohl

Yes, the way it recovered from that kick is incredible.

And, if you'll forgive me, here's one additional link for anyone who hadn't yet heard of Big Dog. (Unlikely, I know...) http://www.youtube.com/watch?v=VXJZVZFRFJc

⬐ po

If you saw the original Boston Dynamics BigDog Youtube video, should check out the version that researchers here in tokyo have come up with:

http://www.youtube.com/watch?v=X5BPrlOVX2o

This is still my favorite robot: http://www.youtube.com/watch?v=W1czBcnX1Ww

This video is old - but I though it was worth watching again after reading the paper. http://www.youtube.com/watch?v=W1czBcnX1Ww

I thought Toyota was pretty good until I followed the link in the article to Big Dog. http://www.youtube.com/watch?v=W1czBcnX1Ww Big Dog is truly awesome. It runs, jumps, recovers from slips, climbs over rubble, ...

⬐ scott_s

Both are impressive - I can imagine that a bipedal robot is a harder problem than a quadruped.

⬐ peregrine

Amazing what an Industrial Military Complex can do with half the US Budget. I'd like to see them put some money into space myself but robots are good too.

⬐ jerf

Arguably, putting money into robots is putting money into space. Along with a lot of other things, too. Robots have already long since passed through the hype cycle and now we're firmly sitting at the end of the cycle in "ho-hum", but it's worth taking a moment to realize just what this all means, and that the old sci-fi predictions may not have happened on schedule but are about to start popping up.

While everyone's busy moaning about the economy and how terrible everything is, I've been looking down the road five or ten years and I see amazing things coming at us, fast.

⬐ gnaritas

OK, I'll bite, what's a few of the most amazing things we'll see in the next five years?

⬐ fauigerzigerk

One thing that might happen within the next 5 years is that robots start to learn more from what they see in their environment and from the feedback they get. I'm not sure how much that is already part of current robots, but I haven't heard about it so I think it's not an important part.

One way we humans learn is by observing things, trying to immitate them and learning from the feedback (either physical or via communication). If that principle takes hold in robot development, the "evolution" of robots could get much speedier I believe.

The ability to act in the physical world and actually cause things to happen, cause reactions, is what differentiates robots from other types of computers. So that's what I would try to exploit thinking about the principles of programming robots.

⬐ jerf

I don't see a lot of household robotics in that time frame, but I think we're going to see a new generation of industrial robots. Some of them have already started to pop up, and robots like this story will continue that trend.

Right now, robots only work in spaces explicitly designed for them, which has really limited their effectiveness. They're going to be able to move beyond that into more human spaces, which has two major advantages: A lot of human spaces exist, and those spaces are set up to do more different things, which robots will be able to participate in. An assembly line is a fine thing, but not terribly flexible. (Though increasingly, this is a fuzzy line, and the line will get fuzzier; indeed, this is one of the effects I am anticipating.)

Computer vision seems to have taken a new step forward lately, and while they're still miles away from "full human vision" (which is AI-strong, ultimately), they're moving from total toy problems into really useful problems.

Brain interaction is also moving forward lately.

Now, you might say I'm not being specific. Well, here's why. In the five-to-ten year timescale, what I see is that where today these things are still the cutting edge of research, in the five-to-ten year timeframe this is going to be available to small startups and dedicated hobbyists. I don't know exactly what will be made of advanced computer vision and smart robots and direct brain interaction and another five-to-ten years of hardware development and two guys in a garage (oh, and we'll still have the internet, of course, and pervasive cellular connections), but I find it quite likely that it'll be very significant... in exactly the same way the Internet is significant. It may take a genius to push the frontier of computer vision or robotics, but thanks to the miracle of programming, it won't take geniuses to apply them in new ways.

Also, I could probably go on for another three or four times longer, but this is an HN comment, not an essay. (Though maybe I should make it a blog post.)

You wake up today and it's much the same as yesterday, but you know, look back at what the Internet has done to so many industries in the past ten years, and what it has yet to do, and I think you start to get a reference frame for what we may be looking at with the convergence of some of this stuff. I know I'm not giving you product names and the bullet-point feature lists, but that would be almost exactly like me trying to predict Facebook back in 1990. Even those who were on the right track in the broad sense were wrong in many ways... but they were right that big things were coming.

Big things are coming.

⬐ modeless

Here's a specific technology that's going to be a huge boon to robotics in the next 5-10 years: Flash LIDAR cameras. This is the technology behind the XBox 360's "Project Natal", but it's a lot more than a game controller. A Flash LIDAR camera directly senses the distance to every object seen by the camera by measuring the time-of-flight of a reflected laser pulse. With accurate distance data for every pixel a lot of really hard computer vision problems suddenly become much, much eaiser.

⬐ None

None

⬐ lutorm

That's very cool. Does it use vision or just tactile feeling of where it puts its feed down?

⬐ joeyo

I wouldn't be surprised if it uses neither. My guess would be it has encoders to know the state of the joint angles (so, a type of proprioception) but probably not for foot contact.

If someone has a reference to some technical details, though, I'd be very interested.

⬐ JabavuAdams

Posted: http://news.ycombinator.com/item?id=731300

It uses an IMU, optical vision, lidar, encoders, force sensors, and ground contact sensors.

Why do you think joint angles alone would be enough?

⬐ joeyo

Thanks for the link! I just figured it would use the simplest solution that could get the job done. I don't know how important measuring foot slip/ground contact is for quadrupedal locomotion or how well it can be approximated by measuring joint angles and computing forward kinematics. I guess load cells are cheap so why not use them?

Again, thanks for digging that up. It's heaps better than idle speculation.

⬐ DougBTX

Getting the job done is important! A simple solution is no good if it doesn't work. Presumably they need to know where the ground is / how firm the ground is, before putting all it's weight down. There is a good bit in the video of it climbing over rubble.

⬐ trop

I ran into an MIT engineering grad who was talking trash about Big Dog. He said "When you see a video of a new robot, remember that what you see is all it can do." In the case of Big Dog, this apparently means that it can run but not walk. And we only get to see the runs where it doesn't collapse. OTOH, it is truly impressive and sinister.

I think big dog is still cooler:

http://www.youtube.com/watch?v=W1czBcnX1Ww

Asimo too:

http://www.youtube.com/watch?v=cfaAiujrX_Y

⬐ pg

Imagine something that balanced dynamically like Big Dog, but was a biped like the Asimo, and you'd have:

http://www.youtube.com/watch?v=_CQ5AKaEi3U

http://www.youtube.com/watch?v=ZnTy_smY3sw

⬐ jderick

This robot makes small jerky movements. I think the others are much more natural in appearance. Perhaps it is more of a hardware issue than software, as it does seem to have pretty good balance.

⬐ pg

"more natural in appearance"

Truer than you know. The reason the Asimo's gait looks so smooth is that it's preprogrammed. It doesn't balance dynamically. It looks smooth for the same reason animation does.

⬐ ivankirigin

That's not technically true. Asimo has preprogrammed paths, but not motor movements. Even in a semi controlled environment, a preprogrammed biped would fall.

The difference is between electric motors and pneumatics: today's pneumatics are just jerkier.

And static balancing doesn't mean that it is preprogrammed, just that at any point (unless it is running), all movement could stop and the robot would stay standing.

If humans stop moving, we collapse.

I guess the best reason to say Asimo is not more natural is that no animal in nature is statically stable. All animals with legs have actuators with the characteristics of pneumatics and springs (in humans, tendons are springs).

We can also turn to efficiency. The electric motors on Asimo will _never_ be practical. It takes too much energy for the bot to ever be sold as a product. No animal in nature has a 15 minute battery life :)

⬐ ivankirigin

On second thought, a flamingo standing on one leg might actually be statically stable, and there might be other examples. But flamingos are weird, amirite?

⬐ pg

static balancing doesn't mean that it is preprogrammed

Really? That's what I was referring to when I said the gait was preprogrammed. I assumed that static balancing meant that the robot was only willing to traverse certain paths through the n-dimensional space of possible limb configurations, and that since these could be calculated in advance, they would be. But since I know you know about robotics, I'm willing to be convinced otherwise.

⬐ ivankirigin

There are constant adjustments made by the dozens of motors in the legs and torso to keep balance, based upon sensors like accelerometers and gyroscopes. That's not preprogrammed.

My point about a reasonably controlled environment is that in the real world, even on a "flat" stage for a show, there are enough irregularities to make complete preprogramming hopeless.

I don't know enough control theory, but it could be stated that an open loop controller will be doomed to fail for a biped.

http://en.wikipedia.org/wiki/Open-loop_controller

http://en.wikipedia.org/wiki/Closed-loop_controller

Higher level, there are sequences of commands that yield a step forward or to the side. That is almost certainly preprogrammed. When you own an asimo, the api they expose is actually really limited. For example, to lift the arm from the side to the air will only have a few states in between, and you choose which one you want. This is not low level control at all. They do this because they know Asimo is a marketing machine, not a real robot, and they don't want grad students making it fall over on youtube.

Here's some specific suggestions. I was most recently fascinated by this video on head tracking "virtual reality".

http://www.youtube.com/watch?v=Jd3-eiid-Uw

Another cool thing if you're into DIY robots (not really educational, but I found it inspiring) is the Yellow Drum Machine.

http://www.youtube.com/watch?v=_RyodnisVvU

Not sure why I got on the robot tangent but here's another good one, a 4-legged robot that's pretty good at navigating hills and obstacles. Especially the recovery from a fall sequence around 1:25.

http://www.youtube.com/watch?v=W1czBcnX1Ww

And here is hacking the Wii Fit controller and a Roomba to do a sort of vacuum-by-surfing... thing?

http://www.youtube.com/watch?v=tLbprdjTX0w&feature=relat...

⬐ mattjones

This excites me. If they become common, people are going to spend a lot of time exploring the world -- an antidote to sitting behind your laptop all day.

These things would make great pack animals. When you go hiking, you'll be able to toss your gear into one of these things and have it trot along behind you. You'll be able to take more risks because robots will be able to rescue you from tight places. If you get tired or sick, you'll be able to ride them.

It's funny to think that the streets of San Francisco in 2100 might superficially look like a town in a developing country. ("Oh, those aren't mules in the street, they're robots.")

⬐ jonah

My question is why not a real mule? More flexible refueling options, has already been through many iterations, and not nearly as expensive. Seriously.

⬐ waleedka

Because this is only the beginning, and soon you'll have versions that carry ten times more than a mule, travel faster, and can be controlled remotely.

⬐ mattjones

Interesting question. Perhaps we will witness a renewed proliferation of domesticated animals in this millennium, a sort of "domestication 2.0". Robotics may actually help inspire it, along with our knowledge of so many more kinds of animals. As robot designers study lots of different critters to improve their products, people will start thinking, "Hm, some of these animals have skills." The range of animals we've made use of so far is probably tiny compared to the range that could be useful. And our ability to breed or engineer them in desirable directions has improved a lot. So yes, maybe there will be real mules too, or some other kind animal we haven't thought of yet. (Have you ever seen guide horses? http://www.guidehorse.org )

⬐ brent

The bigdog is better at recovery when slipping on an ice patch. I need that.

⬐ dejb

Ever heard the expression - 'Stubborn as a mule'

⬐ mechanical_fish

The price tag is not a bug, it's a feature. What, you think DARPA would pay you millions of dollars to invent a real mule? Buck Rodgers movies didn't have mules!

I will say, though, that compared to the typical military-industrial spendthrift extravaganza, this project looks really worthwhile. Think: powered prosthetics for quadriplegics. Think: remote exploration of Martian mountains. Think: lots of generally applicable algorithms for real-time physics.

⬐ Prrometheus

>Think: remote exploration of Martian mountains.

Think: Donkey in spacesuit

⬐ Hexstream

"You'll be able to take more risks because robots will be able to rescue you from tight places."

"In other news, the number of accidents this year has boomed because of the carelessness induced by the presence of friendly robots."

⬐ jrsims

I admire your optimism. It is indeed exciting technology, but I think this will have military applications long before it is available for public use.

For now, Roomba will have to do.

⬐ kajecounterhack

eh, I kicked my roomba... and it broke.

gotta get one of these pack mule buggers...

⬐ LogicHoleFlaw

Just watch out. Pack mules kick back.

⬐ nazgulnarsil

It's getting scary. It's not that we're going to have robots going haywire and killing people on its own. It's that the elite crust of humanity is going to use it againstt us. Think blackwater is bad now? How about when they're not even human beings?

The thing keeping the elites in check for all of human history has been that they can't get too crazy or their own soldiers will mutiny. Hitler was a study in charismatic leadership. He was able to commit atrocities only because the german populace liked him so much. WHat attrocities will be possible when it is impossible for your soldiers to disobey?

⬐ thorax

That's when the hackers here step in, right?

⬐ sabat

As long as hacking without a license remains legal.

⬐ None

None

⬐ SwellJoe

Sense. Your comment makes none.

Hackers (or crackers, if you prefer) don't wait for permission.

⬐ sabat

Direct object. Your sentences use it unconventionally.

Anyway, I was being wry. You may not be waiting for permission, but it does matter if it's illegal.

(Clarification: I use hacking in the same spirit as this site, meaning to code with a hobbyist's or experimenter's ethic. Not meaning the media obfuscation of the word.)

⬐ nazgulnarsil

many issues in the next few decades will involve technology outrunning our ability to make ethical choices about their uses. what will the ultimate effects be of crops with terminator genes be? what happens when biotech created in a lab winds up mutating out in the wild? human cloning and genetically tailored babies are great examples that are emerging right now.

of these robots being used as somewhat intelligent disposable soldiers scares me the most because it is the easiest to abuse and the effects would be felt immediately. when it becomes a matter of pressing a button to suppress an uprising, who will decide when it is appropriate to press that button? remember that winners get called freedom fighters and losers get called terrorists. The USA wouldn't exist today if our "freedom fighters" hadn't resisted british rule.

⬐ 1gor

What's scary is how much this comment reminds me of Reddit...

⬐ nazgulnarsil

I understand, i'm usually against conspiracy theory bullshit myself. using robots on the battlefield is the only subject I get a little spooked over.

⬐ astine

Robots on the battlefield doesn't spook me at all. It seems to me to be an improvement over sending people to risk their lives.

⬐ motley_fool

This is an overly simplistic and dangerous view of warfare. Increasingly, warfare is not about soldiers being sent to fight each other in some empty battlefield.

The pattern of U.S. wars since WWII has been expeditionary warfare where the U.S. sends its troops to fight with irregular forces, among civilians of other countries.

No war is clean, but this is a particularly evil setup, because the worst that American non-combatants experience is the return of their dead or wounded troops. On the other hand, the host country feels the full horror of war.

So, frankly, I don't care nearly as much about the volunteers who are sent to risk their lives as the civilian non-volunteers who are necessarily and inevitably harmed.

This is not a U.S. specific thing. It's just that as citizens of the top military power, Americans are spared the full horrors of war, yet have the ability to visit those horrors on others.

There is a reason why Europeans are not nearly as accepting of warfare.

⬐ astine

The reason Europeans are not as accepting of warfare is because they screwed themselves over twice with it. War is always the last resort of negotiation. When nations become too belligerent or stubborn conflict will always follow. Europe forgot this lesson twice in as many generations and suffered the consequences.

There are only two ways to prevent conflict: 1. be reasonable in all of your negotiations and 2. make sure you are always ready for a conflict.

If you fail to do the first, others may see no choice but to fight you. If you fail to do the second, they may not see any downside to it. This boils down to one principle, you must always back every one of your demands, even the reasonable one which you have a right to, with sufficient force.

In short, there is nothing wrong with using machines to do our fighting for us. There is something wrong with killing civilians and unnecessarily invading foreign countries; but there is nothing wrong with using better weapons to do the same job.

⬐ motley_fool

> There is something wrong with killing civilians and unnecessarily invading foreign countries; but there is nothing wrong with using better weapons to do the same job.

You're assuming that these two issues are independent. I'm proposing that handing our rulers robot warriors will result in harder-to-expose killing of civilians and more unnecessary invasions.

Of course, if all the killer robots were working for me, I would deploy them wisely and reduce civilians deaths.

⬐ SwellJoe

"It seems to me to be an improvement over sending people to risk their lives."

So, who are the robots fighting then if not people? I'm just thinking what a frightening thing it would be for an empire builder like GWB to have a conscience-less army of mechanical killers. At least our armed forces generally want to avoid killing innocents. Big Dogs with M50 machine guns don't care about such fine distinctions. So, yes, it could be scary.

Now, of course, as soon as these are available to the general public, I fully intend to exercise my second amendment right to own robotic death machines. That's in the constitution, right?

⬐ gills

Because maybe, just MAYBE, when we realize our robots are fighting their robots and it's really just an expensive video game, we'll quit fighting and actually sit down and talk.

That is, if the robots don't join forces and turn on us first.

oops..spelling edit.

⬐ motley_fool

Nonsense. People don't sit down and talk when they have less to lose. They negotiate when they have more to lose.

So what will actually happen is that these ultra-tech tools will just further shield voters from the nastiness of war, and allow their leaders to visit horror on those less technologically advanced.

What I'd like to see happen is the creation of some kind of independent news agency that uses UAV and robot technologies to bring back uncensored, uninvited news back from warzones. Of course if this were truly effective, that news agency itself would become a target.

⬐ angstrom

On a long enough timeline every discussion on the Internet converges on Hitler and Nazis.

⬐ euccastro

To be honest, I'm quite tired of all this "ugh, this is so reddit!" I wish people would discuss the (de)merits of comments in their own terms.

⬐ 1gor

OK. I'll explain why it reminds me of reddit.

The author of the comment in question uses tired populist cliché (elites) and tries to impose their own political views on the reader (you think blackwater is bad).

Which is likely to generate a long, heated and shallow political debate rather than a discussion about the technology (or about the social implications of the technology).

⬐ rms

Blackwater being bad is not a political view; it is close to moral truth. I have talked to multiple US soldiers back from Iraq who have described Blackwater committing what would normally be war crimes. However, Blackwater is not subject to laws, so they're not war crimes.

⬐ rms

Details of one of the incidents:

Soldier was in the Air Force, in northern Iraq near the Iranian border. They had a warrant to raid a house looking for a weapons cache. They did the raid with some guys from Blackwater. When they got inside, there was a guy laying on the couch watching TV. Normally, anyone inside a house during a raid would be put in flexcuffs. A guy from Blackwater just shot him, while he was laying on the couch.

⬐ yummyfajitas

>Think blackwater is bad now? How about when they're not even human beings?

Then they run at a lower threat level.

A humans's decision process: "Is that guy a suicide bomber? OMFG SHOOT!" A few minutes later: "Fuck sarge, I got scared. I screwed up big time."

A robot's decision process:

    if uncertain_of_hostility(target):
        speaker.warning() #This robot only costs us $50,000. Hearts and minds, right?
    else:
        fire_warning_shot(target, bystanders)

    ## if fu_bar(target): #Don't delete or uncomment this line, or else bad things will happen.

⬐ cheponis

I worked on the original hopping/jumping/running robots at CMU and later MIT with Marc Raibert. I can assure you that everything done was above-board, published, and open. Dr. Raibert is one of those rare scientist/engineer/entrepreneur types who is interested in progress for people. He was asked by DARPA to develop this vehicle, so he wanted to help them out.

The Sci-Fi stuff in these comments are, well, crazy.

⬐ mechanical_fish

I've got to say, this just looked like yet another in the line of CMU robot videos -- impressive, but only to someone who actually knows how hard this is -- until the kick. And then the ice-patch sequence, which is just awesome.

Kudos to you and your coworkers. This obviously represents one hell of a lot of patient work.

⬐ jrsims

Well, thank goodness for Dr. Raibert. And while we're at it, thank goodness for Albert Einstein.

It's the people who would exploit the technology developed by these benign characters that frightens me.

So it is perfectly reasonable to be paranoid about emerging technologies that have the capacity for great good or great evil.

⬐ andreyf

The Boston Dynamics site is here:

http://www.bostondynamics.com/content/sec.php?section=BigDog

It seems to me that there is a lot of promising work that will be done in this area of robotics. There is a recently published book on this...

http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&...

...which suggests we should evolve brains and bodies together in order to get the most natural/most efficient gaits from them.

⬐ boredguy8

I was watching thinking, "Yawn". Then it got hit -hard- in the side and use its 'elbows' to stabilize itself. I'm just glad it doesn't hit back (yet).

⬐ WilliamLP

Yes, wait until it figures out that the best way to stay upright is to subdue the human before the kick can happen!

⬐ TrevorJ

It seems like a very robust system, and the way it displays the ability to fail gracefully really demonstrates some amazing software. I am very impressed that they are able to un-tether it as well. The fact that they apparently were able to get the power plant and the processing hardware onboard is really very impressive.

⬐ ivankirigin

I love this robot. It is the right way to do legged locomotion.

Leave it to "a few guys in a garage" to crush Honda's 15 year effort with Asimo. Granted, those few guys were key in the MIT leg lab...

Note that BDI is all about the mechanics. They haven't done much work on the perception side of things. If you work in that area, contact them.

⬐ alex_c

It actually makes me think of a robot horse more than a dog.

Does that mean in 20 years we'll all be riding robot horses around?

Once again, the Japanese are way ahead of us... http://www.absoluteanime.com/saber_rider/index.gif

⬐ motley_fool

I want my Chevaline, dangit!

⬐ Goronmon

You know, I knew there was something weirdly familiar about watching that video.

Then someone remarked that it reminded them of a Hunter from Half-Life 2. Now I'm scared.

⬐ Readmore

That was exactly what I thought too. It looks just like the new Hunters. Good thing it's being built for the U.S Military and not the Combine....

⬐ alex_c

Here's the headcrab:

http://www.bostondynamics.com/content/sec.php?section=Little...

⬐ gojomo

And the standalone sentry machine-gun came out from Samsung a while back:

http://www.newlaunches.com/archives/samsung_develops_machine...

Forget the headcrabs -- what about those flying chainsaws in HL2?

Also: will the high school curriculum ever catch up? One of the wackiest 'long bets' offered:

"By the year 2150, over 50% of schools in the USA or Europe will require classes in defending against robot attacks." http://discuss.longbets.org/postlist.php?Cat=&Board=86

⬐ s3graham

Wow, really impressive. The ice was amazing (though as someone with bad knees, transferred anthropomorphization makes me cringe).

Does anyone know why 4 legs was chosen? 4 seems better than 2 for stabilization, but why not 5, or 6, or 8? Maybe looking like a familiar friendly animal is an end goal?

⬐ astine

I'm willing to bet that it was literally and deliberately modeled after a large dog. Engineers borrow design from nature all the time because its much simpler than starting from scratch. Why reinvent walking when dogs already do it so well?

⬐ huherto

Probably the 8 leg design was already tried by nature and it was not successful for large animals.

⬐ pchristensen

Whether in nature or manufacture, those legs aren't free. Is four more legs worth the energy required to grow/build them, and could the mental processing power necessary to run them be used better somewhere else?

⬐ kingnothing

Probably better for it to look like a friendly animal than a giant spider, although that could be nice from an intimidating militaristic point of view...

⬐ wallflower

T3 is inevitable..

2-legged animals are no match for 4-legged robots that can't even be knocked over - just ask the target of a K-9 pursuit dog

⬐ icey

This really has enormous implications over a lot of fields, not just military. I can see this technology being used in prosthetics, in home-care (walking bots), and all over the service industry.

The first thing I thought when I saw this video was "what's the stock symbol for this company?".

⬐ tokipin

what's interesting to me is it's not even that difficult to get life-like, self-balancing, responsive motion, etc with basic things like feedback loops (eg segway.) i don't know what to blame most robots' relative rudimentaryness on except their designers' lack of knowledge about things like that. to me the difficulty is much more in the physical mechanics than in the software. though after you get the mechanics down, the software would become the challenge

⬐ None

None

⬐ paulsb

I can't decide if the bit where he tries to kick it over or the bit on the ice is best.

⬐ chengmi

I think the ice part is better--I doubt any animal would have recovered from that without falling on its side. What I really want to see is a bloopers reel from early in the project.

⬐ thorax

Like those ED-209 bloopers...

⬐ tim2

It seems that it is much easier to stay up with three or four legs than two. Just lock them into a slightly outward direction and you cannot fall. If you're on one or two though, you've really got to balance the rest of the system much more carefully.

⬐ iamwil

Your comment got me thinking. From a technical point of view of traversing over unknown and uneven terrain, there really is no reason to go with bipedal, especially if it's easier to do control systems for tri or quadraped robots. Of the handful that can be bipedal (primates, bears, birds), it's not their preferred mode of movement. However, it's probably a psychological disadvantage for the robots to be 3 or 4 legged and work along side humans in a job that requires it to be treated as a human, like a waiter, house companion, etc. That alone is reason enough to continue bipedal robotic research.

⬐ tim2

For comparison, hikers on rough terrain usually bring a hiking staff and aim to keep at least two points of contact at all times.

I can see two legs being more efficient than four, but wheels are usually vastly more efficient than either..

⬐ carpal

There are only two real advantages to bipeds:

1) Height 2) Use of spare limbs for tools

Height doesn't seem to be much of an advantage for robots at this stage, and I think that robots could have 6 limbs with no problem. Therefore, biped robots are really only a novelty.

⬐ SwellJoe

3) Use of human devices that are biped specific

Ladders, automobiles and heavy machinery, household equipment, etc. all rule out quadrupeds.

While it would be possible to custom-build all of this stuff to be usable by a quadruped, it means you then have to have two of everything if you plan to cohabitate with a robot assistant. From the perspective of assistance robots, it's not going to be possible to accomplish everything without bipedal bots.

I think wheeled bots (like Trevor's Monty) will also be a great option--probably a tenth the cost of a walking mechanism, an order of magnitude more speed and endurance, etc. But they'll still be unsuitable for many assistance robot tasks. Stairs, for example. And unlike the Daleks, leveling the building will not suffice as a solution to the problem of climbing stairs.

⬐ carpal

Well, not necessarily. Those devices aren't biped specific- they're specific to thing with grasping mechanisms.

Who's to say a quadruped robot can't have 4 feet + 1 or 2 grapsing mechanisms? Or have two front feet that are also capable of grasping?

⬐ iamwil

While you can certainly have a 4 legged robot, 2 armed helping you around the house or a hospital, only using grasping mechanisms to climb the ladder to adjust the satellite dish on the roof, help you with the dishes, or get you out of bed, I think it's important to consider the reaction people have to a robot's form factor given its job.

You and I might think it's really cool to have 4 feet and two arms (like a centaur!), but if a robot reminds someone of a gigantic insect, and it's suppose to be a household companion/housekeeper, people might reject having it around the house. That's a good reason to have bipedal robots.

But on the other hand, if you get close, but not exactly with bipedal robots, people might get even more freaked out with the uncanny valley.

⬐ motley_fool

Think Centaur.

⬐ carpal

Good point.

Personally, I'd love to have a robotic centaur helping me around the house.

⬐ drm237

He's alright but dexter can do it while wearing some sweet shoes: http://www.anybots.com/videos.html#walking_080302_title

⬐ nose

DARPA was the original ycombinator!

⬐ anupamkapoor

for a long time it seemed like two guys with their head+torso stuck in a funny box. but truly, wow !

edit: it seems very close to AT-ST from star-wars though.

⬐ sabat

Skynet became self-aware at 2:14am EDT March 28, 2008.

⬐ sabat

Aw, c'mon. I can't believe people voted me down for this one. No one else thought the 'bot seemed eerily self-aware?

⬐ pchristensen

I thought the way its legs moved and the way it stumbled and rebalanced were extremely lifelike. I was wondering how they control it - is it as simple as a remote joystick? Just point in the direction and it figures out how to get there?

⬐ ctingom

Wow! Amazing.

⬐ tim2

Hah, it slipping on that ice is hilarious.

⬐ noelchurchill

The jump is the best are you kidding me?! It's like Jordan haha

⬐ jbrun

That is crazy. It is almost like a real dog - which are free.

The real future is with genetically engineered dogs, horses... to run faster, cary heavy loads etc. We have already done it through selective breeding and soon we will be able to program DNA (see Craig Venter, Dawkins...).

These robots are so mechanically complex that they will ultimately fail. Nature (evolution) is the greatest designer. What complex robots have hit the mass market? None, because they are just too many moving parts.

Animals can even reproduce!

⬐ coffeeaddicted

You just did see a video of one of the most advanced prototypes(!) of a robot and you think it will fail because such things are not yet seen in mass market? Somehow that made me smile :)

⬐ jbrun

How many planes have you seen in people's driveways? The mass market is not a sign of success, all products have to start somewhere.

These robots will remain the realm of military and rescue missions because they have so many damn moving parts. Quality control is what drives up prices and quality control is exponentially linked to moving parts. So my argument is more along the lines of:

"This robot is technologically amazing and promising, but is built upon a foundation of complexity which inexorably leads to limitations".

⬐ coffeeaddicted

I don't think the reason you don't see planes in people's driveways is because of their complexity. Nearly all consumer products are increasing their complexity each year and there's no reason to think that this trends stops. Just take a look at any modern car and compare it with an old one. If people want such robots enough those robots will be produced for the mass market.

And the first application which did spring to my mind immediately was transport. There are so many situations where an automated transport, which does not depend on wheels or rails, would be useful that I can see a huge market for this.

The first mass market might be in the industry. But once prices fall there are also lots of possibilities for this to reach the consumer market. Certainly normal people also can have uses for such transports. Be it as intelligent luggage, as a tool for the garden (replacing the wheelbarrow) or simply as a very cool toy for the kids (just think of riding those things!).

This is a completely new way of automating transports and so I expect it to have an extremely large impact in the following years.

⬐ rokhayakebe

i think we are in deep shit. there is no way that such robots paired with artificial intelligence do not take over the world. we are a few years away from it. humans will self destruct not by killing each other, but by creating robots that will be out of control.

⬐ ArcticCelt

My impression has always being that if we create AI that can tweak itself to become even more intelligent, they will simply look a us with pity, smile, pat us on the head and fly away to explore the galaxies in their brand new interstellar ship.

⬐ microcentury

I'm just glad we get to see in advance what will be chasing us through the woods in a few years, the thum-thum thum-the-thum music playing the background, firing M15 bursts over our shoulders before the camera pans to the blackened sky.

It's going to be sweet.

⬐ alex_c

Well, only if we give them the ability to be self-powered and reproduce without any help. The first is feasible, but the second is so difficult that I can't see it happening by mistake, or being worth the added complexity compared to just making non-self-reproducing robots on an assembly line (on Earth, at least).

⬐ comatose_kid

Don't forget the sharks that shoot lasers from their eyes.

Or the dogs with the bees in their mouths, and when they bark, they shoot bees at you.

⬐ WilliamLP

That's nothing compared to dogs that can be trained to sick specific parts of the human male body!

⬐ ericb

One thing that struck me was when it was kicked or almost fell, I felt a little bad for it. I can see where the general public might someday award robots rights because we instinctually anthropomorphize them. I felt silly--I "know" it's just software, but what about the rest of the world with the same instincts and no understanding that's it's mindless code?

⬐ lastone

machines are our slaves, get over it

⬐ henning

I felt kind of bad in Trevor Blackwell's latest video where he was stopping Dexter from moving towards him. :(

I started constructing dialogs in my head along the lines of "Father, why don't you love me?"

⬐ ivankirigin

dexter: "one day, old man, you'll pay for this"

⬐ henning

the arrival of LOLbots is inevitable.

⬐ alex_c

Too late:

http://lolbots.com/

(went there to check if the domain's still available)