Boston Dynamics will start selling its dog-like SpotMini robot in 2019 | Hacker News Reader

Boston Dynamics will start selling its dog-like SpotMini robot in 2019

techcrunch.com

155 pointsandyjohnson0162 comments

How this isn't exclusively military tech is beyond me. If this sort of thing is available to anyone with a chequebook then what does the military have?

It is begging for a weapon to be mounted to it and sent to patrol the streets of dystopia.

The Police will for sure be buying stuff like this with the argument that it will save cop's lives in hostile situations. From there we'll have to get used to seeing automated enforcement. Oh poop, I am getting into the obvious slippery slope arguments, it's just so hard to not be cynical about tech like this. The world is rapidly changing again </luddite>.

Those things remind me of the Rat Things from Snow Crash.

I don't want to derail the conversation either, but they're owned by softbank which is a glorified holding company of Alibaba (slight exaggeration, granted), so there is a valid tinfoil hat national security argument to be made if there's a million of them running around the street in 10 years.

SoftBank is Japanese and has significant stakes in the US (eg Uber) as well, fwiw.

I'm sympathetic to the underlying sentiment, however. China is starting to look scary and our IP might be worth watching out for.

Softbank has $150 billion in Alibaba - drew them back from the brink way back iirc. Any foreign ownership in Alibaba relies on the goodwill of the Chinese government. Who knows what relationship they can or will have - don't actually think it's too much of a stretch to question.

Completely agree on the IP

AFAIK the DoD lost interest in Boston Dynamics BigDog, because their petrol engine was too noisy on the field and the electric version was too weak.

Did anyone serious buy that? I mean, come on. They literally have sci-fi terminator in their hands, except it's better than the movie.

If you haven't seen the black mirror episode "Metal Head", it will be right up your alley.

This was the first thing I thought of when I saw the demo.

Fun fact, that episode was actually based off of their previous videos.

It's really hard to see how a wepaonised version could be of practical use in an active situation. Granted, it has that nice Hollywood sci-fi look about it, I feel it wouldn't be any more accurate than throwing a grenade into a room.

I would imagine one group opposed to such a device would be the police themselves. They are as likely to get shot as anyone else.

I think a more practical application would be camera mounted infantry for SWAT teams. And, based on some of the incidents we've seen recently, assessing situations is where many interactions go south.

Mounting a facial recognition face-tracking camera-gun combo is somehow going to be inaccurate? Walk the dog round the corner of a bank-heist hostage situation, have it stand on guard and shoot whenever it gets a human verified lock?

I know saying use 'technology' is like saying use 'magic', but all of this seems totally within the boundaries of what is available today or near-tomorrow at the very least.

As for cops being shot up by one of these, I am not sure many criminals would see the value of investing significant amount of money and know-how into one. Especially as it would easily be traced back to a purchaser.

SWAT teams with just the recon implementation is worrisome enough as that's clearly just a stone-throw to a weaponized version once they gain confidence in it.

> bank-heist hostage situation, have it stand on guard and shoot whenever it gets a human verified lock

It's about time someone disrupted the field of killing hostages.

Why can't we do the same thing with a swarm of cheap drones with a single shot zip gun? Navigate to within a couple feet and shot them in the face?

The future is bleak: https://www.youtube.com/watch?v=TlO2gcs1YvM

Because then you grab yourself another swarm of the same cheap drones and program them to shoot their zipguns at the enemy drones.

This discussion reminded me of a scene from the movie The Jackal in which an assassin basically mounts a rifle on a three-axis motorized mount, connects a calibrated camera to a mounted scope, and then mounts the thing in a truck/van and parks it near the target and uses it over a laptop w/ wireless connection. I think this was in the 90s and I remember thinking, at the time, video bandwidth and latency would be the biggest obsticle but now days it seems like this setup would be near trivial.

When I just tried to google the scene I saw an ISIS setup of virtually the same thing :/

>It's really hard to see how a wepaonised version could be of practical use in an active situation.

Dallas Police used a bomb defusing robot to blow a guy up... so it's not too hard to imagine.

https://www.theverge.com/2016/7/8/12128230/dallas-police-bom...

This is super edge case, and the robot didn't shoot.. it placed a bomb. This is apples and oranges to putting a gun on a mobile robot and have it effectively target a human.

That seems like a distinction without a difference. If the bomb disposal robot had been remotely controlled to detonate the bomb it was carrying while standing in front of the suspect, would that be different in any ethical or legal way from a Boston Dynamics robot being remotely controlled to pull the trigger of a gun while standing in front of a suspect?

I suppose one small difference would be whether the bomb disposal robot in that situation was destroyed by the explosion it caused, since it presumably wouldn't be destroyed by shooting a gun (unless the suspect returned fire or had a dead man's switch on his own bomb). I highlight this difference not because of the ethics of sending robots to their "deaths", but the minor practical consideration of whether a given robot is considered reusable, that is, able to amass a potentially unlimited number of kills.

From my perspective the difference is the control design loop.

To place a bomb is a linear bang-bang controller (pun intended) And the decision is already made wrt collateral damage

But to target a moving object from a moving object holding a gun on a 3-dof mount is an order of magnitude harder to control and the question of collateral damage is an unknown.

>the question of collateral damage is an unknown.

Sure... but collateral damage is already a problem, as humans are very poor at actually hitting targets. People who are trained to shoot in high pressure situations are only in the area of 18-30% accurate.

I suspect if your goal was 30%< accuracy during an active shooting, it's absolutely within the realm of possibility today.

>According to a 2008 RAND Corporation study evaluating the New York Police Department’s firearm training, between 1998 and 2006, the average hit rate during gunfights was just 18 percent. When suspects did not return fire, police officers hit their targets 30 percent of the time.

I can't speak for police, but for military, i could see this being an actual life saver for an infantry rifle platoon or medical emergency. It could run in with extra ammunition or medical supplies enough to stabilize any injured. This same thing could be used for civilian emergency situations or disaster relief for hard to reach areas that a normal vehicle might have trouble reaching.

Why spend money on training and maintaining a human infantry at all??

If one of these things, maybe the more advanced biped version could be bought for $1M it would have to be cheaper and more reliable than a human. No salary, no pension, no health insurance, no time off, no training hours, just pure measurable, reliable killing. I guess there would be a huge crew of maintenance engineers though, but maybe automate the servicing too.

For one, machines haven't (yet) proven as capable as them in dynamic situations. The tech both hardware and software simply isn't there yet. The fact that these Boston Dynamics robots are both loud, and have a relatively short battery life cements this current fact. This problem will likely be overcome in the next 5-10 years or less. That kind of makes it a nonstarter in and of itself. As it stands today, it is the same reason you don't (yet) have machines replacing fast food workers. The economics for most places or the actual tech simply isn't there 100%. Not yet at least.

Some potential non-military applications that come to mind:

1. Scientific expeditions. For example, it may be able to get closer to flowing magma or hot geysers for samples than humans safely could.

2. Extraterrestrial expeditions.

3. Construction industries.

4. Rescue operations (post-earthquake; collapsed buildings; etc.)

For sure there are non-mil applications, no denying that. It doesn't negate the concern that the military must be craving something like this though.

Why would it be a bad thing if the military had access to more tech?

My point was that this is crazy high tech that I am surprised isn't limited to the military. I am wondering if this is consumer grade, what does the military have.

That said, more tech for the military means more investment into the idea that killing each other is a splendid idea. To me that seems counter to what technology is ideally about. I am an idealist though, so dismiss that comment as you wish.

I'm not sure I understand your perspective on that. Ostensibly, the purpose of your defensive forces is to protect you, and don't you want them to be well equipped for that purpose? To me, your statement looks equivalent to saying that equipping the police to do their job is like an investment into the idea that putting people in prison is a splendid idea. I'm not sure that's what you were saying, although you could say that if you are hyper cynical. But even if you were, I know where you will call when someone is breaking into your house at night.

It can be used to inspect off shore oil rigs. They are full of stairs and something like this could be ideal. Oil companies are also less price sensitive. I wouldn't be surprised if the Spot Mini costs as much as a luxury sedan, unless they sell them at a loss.

I thought stairs were on of the hardest aspects for robots to master. I know the four leg bots are better but I still think having these heavy dense robots climbing 4+ stories of stairs would present more of a danger then it would help mitigate. Some kind of special built inspection system that runs along rails or something seems like a better use case.

They are, but if you pull up the videos of Big Dog and its smaller chassis-brethren, you find that they're quite a ways towards solving that problem.

That having been said: special-built inspection systems would make more sense if it's cheaper to retrofit one in than to buy a robodog.

>Some kind of special built inspection system that runs along rails or something seems like a better use case.

Drones are getting popular ,with laser scanners mounted on them, for older platforms.

But what really makes the most sense to me is just locally networked sensors stuck right on the equipment. Not IoT, L(an)oT.

The logical response is that of course it _could_ be weaponized, but, there are a whole host of cheaper drones that are much better candidates than this for that purpose. I’d bet on inspection/rescue in mines, demolition sites, and anywhere indoors with stairs or rough surfaces where people should not go. Maybe you could use them for mine sweeping.

Real question is whether they’ll make a version that makes no sparks so it can work in flammable atmosphere environments.

> Real question is whether they’ll make a version that makes no sparks so it can work in flammable atmosphere environments.

Can't you "dress up" the robot?

I don't think you can for environments with continuous exposure - http://www.rstahl.com/fileadmin/Dateien/tgus/Documents/ExPro...

usually there is a requirement not to use any relays that have physical contactors and other similar parts.

Carrying heavy things is (by my humble estimation) 4/5 of war. So if there’s something that can carry heavy things, yes it will be used for that purpose.

Big Dog was floated for that purpose and failed to meet the requirements.

(Specifically, it was intended to be a marines platform---I guess the main infantry doesn't need it because they already have trucks?---but it didn't meet the goals for power constraint or noise constraint).

Yah first time I saw the prototype I thought, “that’s a good way to lose the initiative on a patrol”. Perhaps a battery powered version with quiet servos will approach. It will also need to understand hand and arm signals to go on patrol.

The idea is incredible. You could greatly extend the range, duration, and lethality of the average patrol (9 men) by taking a few hundred pounds off their backs (or just adding that much in water and ammo to the robot).

Drone with a pipe bomb strapped to it makes the most sense to me.

Sticking an m-16 on the top of a dogbot that can wiggle around at 3mph... not so much lol

Today 3mph, next year it will be running at 30mph+.

I don't really see the threat considering how much cheaper and more effective it is to give that weapon to a young man rather than mounting it on that expensive robot.

Its a lot easier to find robots willing to massacre innocent people.

That's demonstrably false

ICBMs?

I believe that's BrainInAJar's point: the number of people killed by regular ol' humans with weapons of war is massively dwarfed by the number killed by drones. ICBMs (to my knowledge?) haven't actually been launched in a war ever, so their kill-count is "0".

This solves the General problem, from Bertolt Brecht's poem :)

The threat is to anybody who those in power label the out-group. For too many people, risking lives of the the "good guys" is the only cost weighed for violence.

Tell that to the guy balancing the military pension and medical compensation books.

You don't even have to arm it. Scouting and as a flanking unit in breaching situations is enough to show the value of the robot for tough police situations.

What's the counter? Weapons that disable robots..

Also they already killed someone with a robot,https://www.theguardian.com/technology/2016/jul/08/police-bo...

Here’s an alternate proposal: don’t add any kind of weapon to it, and use it to replace the police in everything but the most extreme situations. This way people don’t need to feel their life is in danger during regular interactions with police.

They tried marketing a similar system for military applications. The military refused it because it was too noisy. [1] The Police may have similar concerns.

[1] https://www.theguardian.com/technology/2015/dec/30/us-marine...

The Big Dog has an internal combustion engine (what I take to be an electrical generator) that is about as loud as a chainsaw[1]. SpotMini makes servo noises, but it's all-electric, and quite a bit less noisy.

1: https://youtu.be/cNZPRsrwumQ

Nah. If they're chasing a bad guy out of a building, this is great stuff: just like a barking K-9 officer, bad guy will either run or surrender. With a little armor, he won't be able to shoot it. And with some simple kind of IFF (infrared coded beacon maybe?) you could keep cops out of its sights.

Of course, the police will find a way to weaponise them. In the end, the Police will justify its weaponisation by saying things like “it is for Officer’s safety” or “It neutralises a ‘threat’ without putting a officer’s life at risk.” Similar avenues have been taken already to justify small police stations owning fucking tanks in USA. A more realistic route, I’m willing to wager on, is that Military will do it first and it will handed down to a Gestapo Force near you! Er, I mean Police Department!

Yeah, you’re spot on. The world is becoming an increasingly scary place.

> It is begging for a weapon to be mounted to it and sent to patrol the streets of dystopia.

Normally people use the humble Toyota Hilux for this; should that be restricted to the military?

The Toyota Hilulx is not autonomous...yet. There's still a human being holding the weapon.

Perhaps if the Hilux had been invented in WW1 it would have been a prime candidate for military application. But these days the military has Humvees and tanks and so on. A Hilux is no match for cutting edge equipment else the US Army would not have developed more capable analogs.

Reference to http://www.newsweek.com/why-rebel-groups-love-toyota-hilux-7...

"Conventional" military is the least interesting cutting edge now; much the actual fighting of the present day is extremely cost-sensitive re-purposed civilian equipment.

Another example: https://www.cnbc.com/2018/01/11/swarm-of-armed-diy-drones-at...

There's a term for when people gravitate to worst case scenarios not because the worst case scenario is likely, but because it is easy to imagine. I forget the term.

I think the term you're looking for is ethical engineering; you probably saw something like it at an ABET accredited university.

no, that would not be the term. Here are some candidates

"awfulizing" from Rational Emotive Behavior Therapy https://qz.com/989112/how-to-stop-jumping-to-the-worst-case-...

"Catastrophizing" https://www.harleytherapy.co.uk/counselling/catastrophizing-...

"Catastrophic Thinking" https://psychcentral.com/blog/catastrophic-thinking-when-you...

I can't think of any other term than worst-case thinking, and this post by Schneier: http://www.cnn.com/2010/OPINION/05/12/schneier.worst.case.th...

There's more going on than just that though. It's like we've collectively forgotten that technology is double-edged. Maybe things like Black Mirror that people are referencing a lot (I haven't watched any of it so someone can correct me) only focus on one edge and that leads to things like availability bias. If popular media drives a lot of the heuristic when thinking about technology it's perhaps no wonder when not much of it is uplifting.

Availability heuristic?

Is it because we continue to believe that police departments in a town with two cops and a few hundred citizens dont buy $1MM or so of military equipment?

This doesn't seem to have any hallmarks of being useful to a military. It's just a mobile device, but there is no recovery from falling over. It also appears very fragile. Drones are cheaper and don't suffer from the same issues. Maybe if this were self-righting or was like a cheap RC that could maneuver upside-down it would have more value.

This actually seems to have less utility than many other items, other than pretending to be a dog.

Evocative of the opening scene of William Gibson's _Count Zero_[1]

"They set a Slamhound on Turner's trail in New Delhi, slotted it to his pheromones and the color of his hair. It caught up with him on a street called Chandni Chauk and came scrambling for his rented BMW through a forest of bare brown legs and pedicab tires. Its core was a kilogram of recrystallized hexogene and flaked TNT. "

1: http://www.williamgibsonbooks.com/books/zero.asp

There is also a "mechanical hound" that chases Montag in Fahrenheit 451 and a similar killer cyborg guard dog creature in Snow Crash.

Let's get some good fiction with optimistic views of robot animals, mmkay writers? ;)

That'd actually be the "rat thing" guard dog in Snow Crash. ;)

I can't remember the exact quote, but when someone comments on how horrible it is to cyberize a dog, its creator says something to the effect of "People keep saying that, but nobody asks the dog." Jump-cut to a couple pages from the point of view of the dog, which (relative to its past life on the street) is incredibly happy with its new function.

I can just as easily mount a weapon to my car and drive around killing people. How is this any different?

None, until you equip the dog with an AI that decides on its own when to kill people. I am pretty sure this is what the OP meant.

Hackers like George Hotz can and have decided to equip their cars with AI.[1] Again, how is this any different?

[1] https://www.theverge.com/2017/7/7/15933554/george-hotz-hacki...

People are hacking on everything from micro-drones to full sized cars. Why single out this platform? Explosive on drones or automated weapons on cars could all be misused.

In this context "People are hacking everything" is not entirely reassuring.

You really have no problem with a George Hotz purchasing a bot, soldering guns to it, installing some form of AI that controls the trigger, and then letting it loose into the world? Seems a bit irresponsible.

But to answer you questions about the legality all this - I don't know. Tesla and Uber are putting autopilots in their cars. Is that legal? Apparently, it is. Yet, those autopilots are already killing people. So expect a regulatory backlash, once enough people die.

I hope it won't take a massacre perpetrated by a buggy AI with a gun to get us there.

Something like this was weaponized - back in World War II - with mixed results:

https://en.wikipedia.org/wiki/Anti-tank_dog

Possibly ironically, one of the reasons BD is looking for consumer-sale alternatives is that it specifically didn't meet the bar for military tech. The Big Dog project started as an idea for a field heavy-pack machine to offload some of the weight marines deep-inserted into a theater need to carry. It failed to meet either the noise limitation or power demand needs for that use-case.

Turns out for military applications, "Just send some trained dudes with guns" is a lot easier, still, than ground-based drones. Air drones are, of course, a whole different kettle of missiles<delete>fish.

This seems like the most plausible theory to me. Even _if_ one day these robots can handle a lot of manual tasks, they won't have big enough batteries to last for more than a couple of hours unplugged

A slightly larger version of Sand Flea seems like a better fit. Disposable, can be deployed relatively cheaply in large quantities, with good camo nearly invisible on the ground, can scale obstacles by jumping. Could be made to carry a fragmentation payload and/or spy gear.

https://m.youtube.com/watch?v=6b4ZZQkcNEo

this is the top comment?! the most active discussion thread? Somehow dystopian handwringing and cheap, shallow, commentary is so much more interesting than progress, technology, and business to the audience here (of all places). get a grip. gossip elsewhere. </smh>

I don't think technology for the sake of technology is a good reason to build it.

Advancing technology is interesting to work on, but deploying it in useful ways to solve problems is the part nobody wants or knows how to do. There's more to be done with what we have now.

That's a pretty pessimistic world-view you have there. How about science for the sake of science? or art for art's sake? Develop new capabilities and see what people can do with it - it's hubris to think that you can develop the future in a planned direction and be right. And as to not moving on until you've done everything with what you already have - that's silly. I mean, people are still doing novel stuff with pottery - I suppose we have to wait for your approval to move on to, like, bronze?

Have you seen the Slaughterbots video? They seem potentially more dangerous and the hardware is much simpler: https://www.youtube.com/watch?v=HipTO_7mUOw

If you're scared by this robotic dog, don't start thinking about quad-copter style drones too hard!

These things just look terrifying. The reversed knee joints, the square bodies, the vulture beak head, the color, all send a message of hostility. You want to put one of these things in my office or my home? Forget it.

Or hellz yeah - they’re amazing! Bring it on

I have absolutely no idea what you mean:

http://ew.com/tv/2017/12/29/black-mirror-metalhead-interview...

That's art-imitating-life though. The look of Metalhead was inspired by the original Boston Dynamics videos of the balance tests.

These traits would make it a pretty good guard dog though.

I agree that they look terrifying. I still would buy one. Maybe that is their angle to look terrifying?

Or do you mean the uncanny valley?

Seems like a solvable problem, stick some googly eyes and bunny ears on it. Maybe add speakers to make squeaking noises when it walks.

"DO NOT BE ALARMED. I AM A HARMLESS ROBOTIC ASSISTANT. I AM JUST HERE TO OBSERVE AND MAKE SURE EVERYTHING IS OKAY IN THIS SECTOR."

Said in the most synthetic-sweet robot voice they can find, pouring out of the speakers every two minutes. ;)

Holy shit, you weren't kidding. This thing reminds me of a creature from a fantasy book I read a long time ago - the Wyrsa: http://valdemar.wikia.com/wiki/Wyrsa

I always thought it looked very puppy like, not at all terrifying.

But that might be because i've seen videos of it like [0] where it likes to hop and bounce about (and it doesn't help that the yellow is a bit less "military robot sent to kill")

[0] https://www.youtube.com/watch?v=kgaO45SyaO4

> These things just look terrifying. The reversed knee joints, the square bodies, the vulture beak head, the color, all send a message of hostility. You want to put one of these things in my office or my home? Forget it.

I agree, although I don't know exactly why. Maybe it's our reptilian brain which sends a message of fear. All I want when I see this robot thing is to hit it with an axe until it does not move anymore...

Imagine in XYZ years, when the technology advances, and these things come equipped with fur, articulated joints, and lifelike facial expressions. And an AI that mimics a real dog's behavior. Woofwoof. Not sure if I am excited or creeped out.

I enjoy rereading http://www-formal.stanford.edu/jmc/robotandbaby/robotandbaby... from time to time, I still wonder how plausible it will be that we'll end up with laws that require robots to be somewhat menacing-looking to avoid people (especially children) getting too attached to them. (In the story, household robots have "the shape of a giant metallic spider with 8 limbs: 4 with joints and 4 tentacular. ... They spoke as little as was consistent with their functions and in a slightly repellent metallic voice not associated with either sex. ... Also robots were made somewhat fragile on the outside; if you kicked one, some parts would fall off. This sometimes relieved some people's feelings.")

Some of us will always want these robots to look even more horrifying though. \m/ And for myself I've always found each of the BD robots somewhat cute and always sympathetic whenever hockey stick man bullies them.

Human perception is a funny thing.

I used to find the look off-putting, but having seen more and more videos of the thing, I sort of find it adorable.

I guess that makes sense though; some people think insects are really neat. There's bound to be some variance in human reactions to appearance.

This is most likely also due to competition that is showing up like ANYmal [1].

[1] https://www.anybotics.com/anymal/ [2] https://www.youtube.com/watch?v=2RQDp0Q2vSo

I like how it uses its leg to press the elevator button.

While yeah, those things are super had to make I just could not help to think of how pathetic it looked compared to Boston Dynamic's Spot Mini.

Now. They'll get there, eventually. I give them 3-5 years max.

> I give them 3-5 years max.

why?

Hate to downplay anyone's efforts but these look like Boston dynamic's retirement home; they look painfully slow.

Damn, seems like this isn't intended for home use. I had my wallet out. No idea why, but I'd love a little robotic "pet" that was able to navigate the world.

Bonus points if it had an API to easily "train" it. Endless fun

Give it time. Shortly after someone is able to afford one and walk around town with it, there will be laws "regulating their use".

the Nao robot was over $20,000 -- how much do you think this will cost? Pepper is $7,000 and it has no legs. One arm and no navigation skills is $8000, two arms is $25000 from rethink, right?

I'm guessing it'll be over $100K

My guess as well, the R&D that has gone into this would certainly allow it to be that much. But once it's out in the open I'm sure some people will take it apart and hopefully we will see easier (cheaper) ways of recreating it DIY style.

The hardware on these things is pretty beefy too tbh. Like it's not just motors and servos from the electronics hobby shop.

it's like legitimate hydraulic suspension and batteries and power supplies and obviously, LIDAR + it's accompanying software

I would love to DIY this thing but it's a lot more advanced than DIYing even a Powerwall or a Model S

I forgot about LIDAR, in that case, my estimate was too low. LIDAR stuff is expensive, even the lowest end stuff is thousands.

An agricultural grade LIDAR sensor (for drones) costs $250-$500. I wouldn't trust it in my Tesla, but for a DIY Spot Mini clone it'll do.

Still, that's just the sensor...

(based on my very small amount of knowledge) -- the LIDAR in the recent SpotMini video looks high res/high quality -- would a cheap one mean mis-stepping on stairs? I imagine problems of quality abound for less than several thousand

Can you buy one of these with an SDK?

You'd have to roll your own... there are libraries for Raspberry Pi and Arduino though.

I think the parts are probably at least 20K, and the software is the proprietary effort of many, over a decade or more, so the homebrew version will be incomparably inferior, to the point where it's not really "the same thing", like how a Power Wheels car isn't a Tesla.

I spent some time with a Nao unit during a project in college - I don't understand how they've gotten away with this pricing.

By selling them to institutions, with support, rather than people, but also because it's decent. Cheaper robots are all jerky and jittery and slow and have no other sensors/cameras and no decent software interface (that I have seen, but I'm not even an amateur roboticist, I just read about it sometimes).

Ours overheated after a few minutes of extending its hands forward, so it tried to sit, but collapsed, because apparently it needed functioning arms to do that. Fortunately my lab buddy was there to catch it.

For my application I needed microphones which were plentiful, but placed on the head that also contained the CPU with a cooling fan - needless to say I got much better readings from my smartphone.

That being said our Nao was and old model so perhaps they fixed those issues.

Lol no idea - it was just a reaction to the title. When I clicked on the link, I lost hope. If they're selling what's in the image, the parts alone cost more than I want to pay :P

So, in 5 years it'll be a $1000 piece of hobbyist equipment like a Parrot drone and in 10 years a smaller version will be available for $100 for kids?

Unlikely, the NAO Robots have been around since 2008 and they still cost quite a lot of money. The hardware on these things is pretty impressive, and since it's more of a platform for R&D then a general product, the demand for them will be constrained to University research labs and R&D groups at companies able to afford the like $100k+ price tag, so they'll never be mass produced. The same situation exists with the NAO robots.

Wouldn't an Aibo be more appropriate?

I enjoyed its Black Mirror episode

Yeah, it was a nice break from the really terrifying episodes. Just a bit of casual murder.

<Black Mirror Spoiler> If you consider using a knife as a weaponized appendage and launching exploding balls with mesh-networked long-distance tracking-capable shrapnel pieces which embed easily in skin "casual" then I'm almost afraid to ask what you consider less mundane.

watching the video of the presentation I was saddened by the fact that the CEO said that one of their most requested features is to create some sort of wheelchair replacement based on this tech, but it seems they aren't really working on it (around 16:30 in the video).

Given how hard it is for wheelchair users to navigate the world, it would have been amazing if this had been one of the first applications. I know friends of mine on wheelchairs would give a lot to be able to just go for a "walk" on a normal park trail or on the beach without having to worry about it being wheelchair accessible.

It is understandable that from several points of view the surveillance industry is a much easier target for this platform, not to mention that of course carrying a person-sized weight might be difficult also from a battery standpoint, however I wish that was more of a priority for the product.

An alternative maybe of interest to your friends? https://en.wikipedia.org/wiki/IBOT "The iBOT has a number of features distinguishing it from most powered wheelchairs: By rotating its two sets of powered wheels about each other, the iBOT can "walk" up and down stairs, much like a cog railway or a rack and pinion with the two wheels as the "teeth" of the gear. The wheels can roll slightly at each step to compensate for a wide range of stair dimensions. When stair-climbing without assistance, the user requires a sturdy handrail and a strong grip. With an assistant, neither a handrail nor a strong grip is required. The iBOT is capable of tethered remote control operation, useful for loading the wheelchair up ramps into vehicles, or "parking" out of the way when not occupied. Custom software receives data via various sensors and gyroscopes, allowing the iBOT to maintain balance during certain maneuvers. For example, during curb climbing the seat remains level while parts of the chassis tilt to climb the curb. It allows the user to rise from a sitting level to approximately 6' tall, measured from the ground to the top of the head, and depending on the size of the occupant. It does this by raising one pair of wheels above the other to elevate the chassis, while a separate actuator raises the seat slightly more than usual. In this configuration the device is on two wheels, and the 'iBALANCE' software and gyroscope signals control the iBOT to maintain equilibrium, balancing much like the Segway scooter (which was a spin-off from the iBOT development). The user may also travel in this "standing" configuration. It can climb and descend curbs ranging from 0.1 to 5.0 inches, according to the manufacturer's specifications. The limits are determined by the rider's technique and risk tolerance. It is capable of traveling through many types of terrain, including sand, gravel, and water up to 3" deep."

>“The SpotMini robot is one that was motivated by thinking about what could go in an office — in a space more accessible for business applications — and then, the home eventually,”

Both in the home and in the office settings, wheel-leg robot (like Handle) would be more energy efficient, silent, faster and require much less maintenance. 99% of the time it would just wheel around.

I'm sure they can sell several thousands of these based on coolness factor alone, but I don't think it's the future.

If it had wheels, it wouldn't be able to navigate things like stairs. It also won't be cool and would just be a Roomba

I said wheel-legs, not wheels. https://www.youtube.com/watch?v=-7xvqQeoA8c

The article doesn’t mention what it’s purpose is? Anyone know?

It's supposed to be a platform users can mount their own hardware to and write software for. Its back is flat and contains mount points for custom hardware.

> Its back is flat and contains mount points for custom hardware.

Well considering the specs...

>The device weighs around 66 pounds and can operate for about 90 minutes on a charge.

Per physics and current battery technology that run time probably falls off really quickly if you make it lug around any appreciable amount of weight.

A general laborer costs the employer maybe $20/hr (all inclusive) in most settings.

Except maybe in situations where the cost of an employee is artificially high (e.g a role that requires a bunch of certification/training in a union workplace working on a government contract where the actual job duties are fairly menial) I can't see a use case where this makes sense.

I honestly have no idea what the use case for this is. Where can a $20k-ish robot that can carry around 30lb replace/supplement a person well enough to break even over its lifetime?

A couple of thoughts:

- sell it to large businesses that have "innovation labs" based on "coolness", see their use cases and needs, and then double down on developing custom solutions for that. The whole issue with BD in a civilian sense has been as a solution looking for a problem. It might help find their problem with some in the wild.

- battery life might not be a massive issue if they walk themselves back to the charging station and can auto-rotate as a pack.

- I can't see the use case either, but perhaps attach a monitoring camera to the arm and have it perform remote inspections of aircraft? Guessing some kind of in-person work that could be done by a central operator, but needs a bit more autonomy/environment sensing than a glorified remote control car?

Given some suitable attachments and add-ons, a device like this could be used to perform tasks in conditions too dangerous for a human, like inspecting a mining site full of toxic gas or radiation.

Or could be used to retrieve/neutralize bombs. Or weaponize it and send it into a hostage/standoff situation.

It might be adaptable for off-world robotic use cases such as exploration/mining/construction projects on the Moon and Mars.

Robotics has endless possibilities. Unfortunately, some of them are evil. There's been talk about the potential to program a mobile robot to crash through your door and steal your valuables. Or snuff you. Scary stuff.

The most obvious use case is short security patrols around warehouses etc. A single night shift guard at 20$ / hour * 8 hours = 58k / year so having even 5 of these things could be a net savings over 2 years assuming a ~20k price tag.

Flying drones seem like a better option for a many things but they have short battery life and significant weight limits. If these things can do 1 hour with 50 lbs over rough terrain that's probably very useful.

Most off the wall use case? Renting them from Uhaul to help you move heavy boxes up and down stairs. Something of a mid way point between a moving company and doing it by yourself. 50$/day * 100 days a year puts break-even around 4-5 years and 50lbs up or down stairs for an hour before charging say ~3 hours a day could easily be worth 50$.

Mount a light machine gun and maybe a rocket or two and make a few tens of thousands of them and it doesn’t really matter if the battery runs out in 90 minutes. Just air drop them into an insurgency and give then orders to shoot anything that moves.

Oh, that's great. Does anyone know how much .223 ammunition it can hold? Asking for a friend by the way.

I'd buy one to carry my groceries.

It's a shame the charge is only for 90 minutes. I could see this being somewhat of a novel packhorse for going camping or something. Throw some saddlebags on it and you're set. Not sure how to get it to follow you automatically though. Probably would have to write custom software and make it follow a Bluetooth signal on the person to follow.

Will it jump into my bed and sleep on top of me though

No price given. That's the big problem with this technology. It's expensive. It took Boston Dynamics about $125 million to get to Atlas, from looking at their DARPA contract data.

If they could sell 10,000 units a year, the price might come down to a few thousand. The next big improvement in price/performance will probably come when LIDAR gets cheap. This is about to happen as the various automotive LIDAR companies get their act together. Once that's a standard auto part, it will be cheap.

They got rid of the hydraulics. That's progress.

DARPA is already featuring spot mini like robot in one of their new YT videos https://www.youtube.com/watch?v=qUCHN2Fyt_0 at 1:56

And then one day SpotMini gets a firmware update to attack me in my sleep..

If by "dog" you mean killer alien AI, sure

What's the price, tho?

No idea, but I'd expect they'll initially price it for companies until they get the price per unit down... maybe make the business case that it can replace a night watchman type position so replaces one or more salaries.

I'd guess $40k based on a 2-year lifespan (I reckon it'll get longer but the initial versions will have glitches)

Any more guesses? Or... actual information?

This is already interesting. I thought for production readiness the price would need to be 5-10% of that.

Looks like what I’d imagine the mechanical hounds in Fahrenheit 451.

To the comments in this thread that are panicking about SpotMini: https://xkcd.com/1955/

/me starts googling diy emp

Remember that wolfbot that acted as a companion to Raiden in Metal Gear Rising: Revengeance?

What is the price point?

From Cordwainer Smith's classic short story "Mark Elf."

Carlotta stared at the machine. It had legs like a grasshopper, a body like a ten-foot turtle, and three heads which moved restlessly in the moonlight.

From the forward edge of the top shell a hidden arm leapt forth, seeming to strike at her, deadlier than a cobra, quicker than a jaguar, more silent than a bat flitting across the face of the moon.

"Don't!" Carlotta screamed in German. The arm slopped suddenly in the moonlight. The stop was so sudden that the metal twanged like the string of a bow. The heads of the machine all turned toward her.

Something like surprise seemed to overtake the machine. The whistling dropped down to a soothing purr. The electronic chatter burst up to a crescendo and then stopped. The machine dropped to its knees. Carlotta crawled over to it.

Said she in German, "What are you?"

"I am the death of all men who oppose the Sixth German Reich," said the machine in fluted singsong German. "If the Reichsangehoriger wishes to identify me, my model and number are written on my carapace."

The machine knelt at a height so low that Carlotta could seize one of the heads and look in the moonlight at the edge of the top shell. The head and neck, though made of metal, felt much more weak and brittle than she expected. There was about the machine an air of immense age.

"I can't see," wailed Carlotta. "I need a light."

There was the ache and grind of long-unused machinery. Another mechanical arm appeared, dropping flakes of near-crystallized dirt as it moved. The tip of the arm exuded light, blue, penetrating, and strange.

Brook, forest, small valley, machine, even herself, were all lit up by the soft penetrating blue light which did not hurt her eyes. The light even gave her a sense of well-being. With the light she could read. Traced on the carapace just above the three heads was this inscription:

        WAFFENAMT DES SECHSTEN DEUTSCHEN REICHES
        BURG EISENHOWER, A.D. 2495
        And then below it in much larger Latin letters:
        MENSCHENJAGER MARK ELF

For the benefit of non- German speakers,

    MENSCHENJAGER MARK ELF

    HUMAN-HUNTER MARK ELEVEN

like it's a pun on the word "mark" indicating a version of something. Feels weird that it's not rendered as MARK XI, but that's not as cute a name I suppose.

What’s the pun?

Mark Elf sounds like a name (in fact there is a famous guitarist with this name). I assumed that the purpose was to misdirect us so that there was an "a-ha" when you discovered the real meaning of the words, a-la a garden-path joke.

I might be reading too far into it though, apparently the story _was_ originally published as MARK XI[0].

[0] http://mporcius.blogspot.ca/2014/05/three-stories-by-cordwai...

I dont say this lightly.

But: I hate this company.

Spoiler[1]

  Law I / A robot may not injure a human or, through inaction, allow a human being to come to harm.
  
  Law II / A robot must obey orders given it by human beings except where such orders would conflict with the First Law.
  
  Law III / A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

[0] https://en.wikipedia.org/wiki/I,_Robot_%28film%29

Law 0 / A robot may not harm humanity, or through inaction allow humanity to come to harm.

This precedes the 3 laws.

Source: Isaac Asimov in Robots and Empire

Don't forget, when Asimov introduced this new law, those (few) robots that were governed by it also had a modified First Law: "...except when such action or inaction would conflict with the Zeroth Law" (paraphrasing). That got interesting, quick!

Reply on news.ycombinator.com