I know this may be horrifying to some, but there are those of us who would rather make a killing machine. We'll settle for a robot that carries supplies for soldiers or watches the enemy.
Defense contractor employees are typically patriotic in a way that would be really alien to the normal tech crowd. Part of the motivation is to do something for the USA. Having Boston Dynamics get diverted to cutesy home stuff is a decades-long setback for the nation.
There is also the matter of legs. Legs are cool.
Household robots are fiddly. They have to deal with a complicated environment. They can't just look at a toddler obstructing the path and go "Fuck it, I'm blowing that away, EAT FLAMING DEATH!!!".
Household robots are cheap plastic, with the shapes and colors totally determined by marketing.
When you buy a company that pretty much makes weapons, what kind of employees do you think you will get? Do you really expect that they will be happy that you have "rescued" them from being forced to make killing machines? Ha, ha, ha.... NO. You took away the coolness. You took away the awesomeness.
I don't think this was a major problem -- after all, BD knew in advance that Google isn't working for the military. However, the article claims Google wanted a quiet wheeled robot for household tasks, which would ruin years of research into bipedal (or more) robots -- the foundation upon where BD was standing.
I don't know that it would ruin anything to also build a Roomba competitor, or booze wagon or whatever Google had in mind.
If BD has any kind of intention of selling to anyone other than the military or researchers like themselves, they'll need a brand, manufacturing and distribution. Might as well bootstrap that now with some less ground breaking tech.
I'm not so sure building a robot capable of killing toddlers 'does something for the USA' - at least not anything positive.
I'm not particularly horrified by military robots, but I suspect they'd have to deal with complications and subtleties in the environment even more effectively than domestic robots in order to be remotely usable.
> I'm not so sure building a robot capable of killing toddlers 'does something for the USA' - at least not anything positive.
I think that's the point. Rather than building a kick ass military bot, they'd be building a lame domestic servant. People who thought they'd be doing the former may not find satisfaction in the latter.
Making a household robot could be patriotic in your sense of the term... It could spy on people, perform home inspections for restricted possessions, restrain it's owner in case of emergency, etc... I'm sure that you could think of many other wonderful applications. In addition to being helpful to the owner, of course.
>Having Boston Dynamics get diverted to cutesy home stuff is a decades-long setback for the nation.
I don't think an American company being the first to enter or rather create a whole new market is less of a win than being the first to get some cool new weapon.
Wars are won with money remember.
(And that's without even talking about the tactical benefits of having a whole army of spies in the homes of elites all around the globe)
Before Google bought Boston Dynamics, BD was strictly a DoD contractor. The military wants big, strong, rugged robots, and BD was headed there. (Even though the USMC rejected the Legged Squad Support System, the militarized version of BigDog.) DARPA was providing much of the funding. It took about $120 million to get to the good version of BigDog.
Then Google bought six robotics companies, including BD, which dropped their existing customer base and went quiet. For several years, the robotics community was wondering what great things Google was developing. In the end, it turns out the answer was "not much".
The amusing thing is that Google/Alphabet worried about their image from owning a company that makes a big humanoid robot. Google survived paying $500,000,000 to DoJ to avoid felony charges after they were caught knowingly selling AdWords to an FBI undercover operation pretending to be a Mexican drug lord trying to take over the US steroids market.[1] It's not like Google has a cuddly image.
If Toyota wants a robotics company, they should probably buy Schaft, which is already in Japan, from Alphabet. There just don't seem to be civilian applications for BD's hydraulic machines. Also, it's about time for Marc Raibert, the BD CEO, to retire; he's about 67 now.
Utterly bizarre to think that Google is worried that videos of humanoid robots behaving like big dumb pets present the image of a company developing technology "ready to take humans' jobs" when they've invested so heavily in self driving cars with the stated aim of altogether eliminating a large employment sector.
That much was obvious when Google (Rubin) bought them. Expecting a commericial product from BD in short order is absurd. Expecting them to abandon 35 years of work on hydraulic legs and go back to the drawing board is worse.
I guess we'll never know what Rubin was thinking when he went on a shopping spree with daddy's credit card; or if he was thinking at all.
Hopefully Toyota will be more accomodating; and maybe somewhere in the 2030 timeframe it will be time to start thinking seriously about commercializing useful bipedal robots. A company with decades of experience mass producing machines with 1000s of separate components I'm sure will be much better equipped to handle that than Alphabet.
What's truly ironic is that the "Don't be evil" company bought a DoD contract robotics company in the first place. It was obvious all along that robots like the Chetah were designed to push the state-of-the-art in robots towards robot killers.
- feed and walk the dog (and tell him off if he does something bad)
- change bulbs, repaint some walls while you are on holidays
Would be like a new revolution similar to washing machines or dish washers, freeing up lots of people (mostly women) time and for instance giving them a better chance to focus on their career. People would be queueing to buy one.
Less sexy but a much bigger market than another Terminator in my opinion.
The problems you list (if they even are problems) are nothing to do with robotics. They're AI problems. And that's AI in the sci-fi novel sense, not the oh so hip "let's build a cluster of 10000 GPUs running a convolution network to recognise cat pictures" sense.
I am not sure batteries would be a problem for a household robot. There are power sockets everywhere. And the robot can take a break from most tasks to charge itself (and like a dog it would spend most of the day sleeping anyway!).
An advanced household robot would not even have to take charging breaks because it would be well up to the task of plugging itself in wherever it would work (except for the dog walking part, obviously)
In a similar line of thinking, I am actually kind of surprised that there are no wired quadrocopter swarms yet, that would form a bucket line of short cables to trade range for the power and infinite endurance of a grid connection. I am quite sure that there must be applications where this deal would be worthwhile.
To transmit power, its better to use high voltage and low current. Reduces transmission losses; uses much lighter cables. Then where the power is needed its stepped down with transformers or 'buck converters'.
Might still be cheaper (storage, transport, setup, investment) than a crane or scaffolding when you don't need to support heavy loads. And scaling would only be limited by how much power you can push through the chain, whereas sticks need to be quite massive just to support themselves once you go beyond a few meters of length.
It's not a swarm, but I think the one drone from 3D Robotics can be tethered for essentially unlimited flight time. So that's some progress towards what you have in mind.
Because the dog needs to be walked several times a day, whether you need to work late / go on holiday in a non pet friendly place / go out that night / wake up late, or not.
I love my dog, but I work and It would be nice to have someone come to my house in the middle of the day to interact with her. Take her for a walk and make sure she didn't accidentally flip over her water bowl, etc. It would be peace of mind for me and entertainment for her.
It's still more useful than what 90% of startups are doing, so I wouldn't complain about it. Being able to provide care and entertainment for pets in situations in which you can't be with them is not a bad thing.
I am actually surprised that we haven't seen assassinations by drones yet (I mean non USAF drones). Consumer drones can't carry a big payload, but I'd guess they can carry enough explosive to kill someone. And even someone super protected like a US president would be helpless facing 5 drones travelling at 50km/h coming his way. I guess most terrorists are dumb and all recent attacks in the west have shown to be low tech, but I'd rather not count on that to protect us.
But in the US it's fine, since the drones have to carry a license number. That will protect us!
Agreed, and I can't imagine why you were modded down.
It's certainly true that most terrorists are complete dumbasses, but not all of them are. In this case, the scenario includes technology that can carry out asymmetric warfare without placing the bad guy in any personal peril, allowing fanatics other than the suicidal variety to play the game.
I don't think there's much that can (or necessarily should) be done to mitigate the risk, but I'm sure that efforts will be made soon enough.
I think you are right; and what you're saying reminds me of a story from Malcolm Gladwell about the 1981 Traveller Trillion Credit Squadron tournament [1] that Doug Lenat won. He used some type of genetic algorithm (he never released his source) to find that swarming the challenger with many small boats would dominate as a tactic. Miniature swarms, which drones are cheap enough and more than capable for, are not a tactic that people are use to. I fear that a swarm could be terrible in its destruction if used to hurt people, individually targeted or for mass harm.
Same idea as Gen. Paul Van Riper's strategy in the Millennium Challenge war games ( https://en.wikipedia.org/wiki/Millennium_Challenge_2002 ). He spammed the enemy with cheap speedboats and motorcycles, forcing the organizers to rewrite the rules to make the game "fair."
I guess a cloud of drones can probably be shot down with some sort of EMP. It's unclear to me how to ensure the EMP is powerful enough to go through the shields of the drone without damaging the systems of the warship it protects.
Both of these articles are fascinating - underdogs with clearly superior strategies dominate, and rather than learning from that, the powers that be just change the rules to prove themselves right. How shortsighted
> Mr. Rubin is an entrepreneurial spirit who likes to run his own show and was facing constraints on his activities at Google, a person familiar with the executive and Google said. A Google spokesman declined to comment on why Mr. Rubin left.
> “It’s surprising and sounds pretty unplanned,” said Scott Strawn, an analyst at research firm IDC. “If it was voluntary on Mr. Rubin’s part, you would think he would see part of the robotics project through to completion to have something to show publicly before leaving.”
Wow. The Motorola patent portfolio pretty much saved Android.
One might argue that he should have been more aggressive in purchasing other IP portfolios - maybe Nortel (although that may have been before he was at Google) and Sun of course.
But Motorola gave Google exactly what it needed - a mature patent portfolio which other companies would have to cross license.
Google got out of owning Motorola at a $9.5 billion loss, modulo some real estate and the patents.
The Nortel patent portfolio sold for $4.5 billion, which was higher than Google was willing to pay.
Andy Rubin did not like what Samsung was doing with Android and wanted a captive OEM. This doubled Google's headcount, it changed Google's financial numbers materially, it cost a fuckton of money even at Google scale at a time when it was apparent that no phone OEMs were very profitable, and it was strategically incoherent.
Andy Rubin was exactly the right guy to make Android a success. But buying Motorola and intending to keep it as a captive OEM was a large mistake.
"As a startup of our size cannot spend 30-plus percent of our resources on things that take ten years," Rosenberg said, adding that "there’s some time frame that we need to be generating an amount of revenue that covers expenses and (that) needs to be a few years."
This professorial tone is baloney. "A startup of our size"? You decide what kind of giant company you want to be, Google. If you don't want to think 10 years out, why did you buy so many research labs? Own your decisions and figure out what you want.
This article is kinda of empty. Lots of implication, but it never actually says anything.
At best as I can tell Boston Dynamics wanted to just try things, and Google wanted a household robot? Yet it also says that Boston Dynamics was worried about generating revenues in a reasonable timeframe.
A military bot is a lot simpler to make then a household bot, ironically. Soldiers can be trained to use a tool a specific way its intended (provided it's ultimately useful). And the range of things a military bot needs to do is actually somewhat more limited to mostly the things we've already seen.
I'm not really sure what a BD household robot would even do.
Are you kidding? For every soldier in the field, how many households would pay for having a robot to do their washing, cleaning, cooking, taking parcels, caring for the dog, etc. It's a huge market. And no export restrictions!
That may be the case in certain part of the world (or even of the US) but if you think of London or New York, a typical household can certainly not afford to have someone full time taking care of these things. Western economies have been converging toward rather egalitarian societies where domesticity pretty much disappeared (and who would be the domestic to the domestics?).
Living in London I can hire cleaners and people to move stuff at very affordable hourly rates (the fact there's not an extremely popular Uber-type service for it says something about the demand...) and get food delivered whenever I want. And I'd expect sufficiently motivated humans to be considerably better at those tasks than a generalist robot.
I suspect that the purchase, maintenance and software licence costs of an extremely complex piece of machinery with incredible AI isn't going to compare that favourably with a minimum wage human anyway, even spread out over a few years.
Minimum Wage Human Capable of operating 24/7 in the $15/hr world is $131,400.
>McDonald's former CEO Ed Rensi: "It’s cheaper to buy a $35,000 robotic arm than it is to hire an employee who’s inefficient making $15 an hour bagging french fries."
Since we're talking about whether there's a mass market for domestic robots, that quote supports my point more than yours. A robotic arm consisting of non-novel technology capable of doing nothing more than bagging chips according to a set program costs $35k.
As a householder, I don't need someone to bag chips for me 24/7. Having someone spend an hour cleaning my flat once a week would be quite useful. Unfortunately, the robot arm doesn't have enough moving parts or AI to do that, and even if it could, I could hire someone on a $15 minimum wage to do that on a weekly basis for 44 years for the same price as the robotic arm.
For the cost, you could say the same with all the technology packed in a smartphone or a laptop. The fact is it has become very cheap once it is mass produced.
Cars, on the other hand, remain expensive. A robot with sufficient versatility to cook, clean and walk the dog is significantly more mechanically complex than a car and likely exposes manufacturers to similar liability issues.
And if the hardware costs mean it's not a mass market product, the software licences aren't going to be cheap either.
I doubt the price of cars is driven in big part by actual costs. Also, at least in my area, you can buy good enough used cars for as little as $400, which isn't very much money. A lot of people drive in cars that cost less than a decent PC.
Cellphones have no moving parts and do not exert large forces on things. High quality high power motors are expensive and will always be expensive compared to the consumer electronics we are used to.
Roomba etc can be cheap because the forces are small and required accuracy is low. An arm that can e.g. load a dishwasher or plug itself into an electrical socket is another beast. Consider that to lift 1kg payload the arm also has to lift itself, and arms typically weigh several kilograms.
Mass producing robots will help costs, but they will still be expensive like a car, not like a TV.
Probably. But you wouldn't buy one of these robots every year. You would replace them at the same rythm as any other mature technology (fridge, dish washer, etc). Obviously not the early models, but democratisation is not about the early models.
i think the entire military spending in the U.S. was >650B. It is probably insanely hard to train a robot to walk around an unknown environment, pick clothes up, launder & dry them, and then have the dexterity to fold them. What percentage of people would buy v1 of this in america. A standard shitty top down analysis tries to capture 1% of a market. Here, your absolute max is 1%, likelely 0.01%. So, capturing 1% of the world market of people who could possibly even afford this is 80k people. assume this analysis is terrible, but you get the idea. many countries spend on military/security. This would branch into police dept ect. Obviously, this woud be terrible for humans, but a robonanny is less marketable than a military robot because the tech is there for a military bot but not precise enough to do chores, which aren't super valuable relative to en masse robot manufacturing costs.
I'd assume the problems will be fairly similar. The terminator will have to evolve in an unknown environment, be careful of not hurting civilians (or the wrong army) and will face a deliberately deceptive enemy. Either way there will be a v1, and a v2, and a v3, and it will take a while before it becomes a cheap mainstream product. But the civil use I think has a multiple of the potential.
Now the reality is that it's probably not going to be one or the other, both will advance in parallel and share technologies.
i am totally in agreement that they will largely evolve in paralell. i just think the market op is much higher for mil tech now.
Most technologies are funded and developed in research stage for military applications, internet being one example. your example was (and i assume you mean far future) a super hard problem. terminator was a fully intelligent super being.
i am talking about the development of the class that boston dynamics is/has built. they can run around in open environments and could do recon, supply and very basic defensive or offensive maneuvering.
i think (and could be wrong) the granular dexterity to fold a shirt is super hard. also, training would be non trivial.
i think robots and augmentation will be mil -> commercial -> consumet. and this is happening. soon all 3 will evolve in parallel and converge on apecific applications
> I'm not really sure what a BD household robot would even do.
Start with basic tasks - carry grocery bags in from the car, walk the dog, carry heavy things from room to room. Vacuuming, mowing the lawn, etc could be next, then picking up clothes and toys from the floor. Maybe even chasing the kids through the woods, so you can keep an eye on them?
And some people do enjoy mowing their lawn themselves or cooking themselves. But I think most do not. And it would mow the lawn in the middle of the week when it is the least disrupting to neighbors. And it's one thing for Mum to enjoy cooking. But if the family has a choice between Mum's specialty (a bunch of potatoes floating in grease) and an electronic Gordon Ramsay with skills and recipes for every kind of cooking in the world, they might vote for the latter.
Boston dynamics worked on legged robots, signed contracts with DARPA, made cool videos showing how effective they were, people got spooked by the said videos, google PR decided image might damage company's face, so they decided to not associate Google name with the videos and robots. That and the fact that Google wants robots that work indoors while BD are making pretty much outdoor robots as of now.
Google is seeing the writing on the wall about web ad revenue and is thrashing around looking for a replacement cash cow. Reigning in their moonshot attempts is reasonable but I think they're throwing the baby out with the bath water on a few of these.
The entire article is very confusing re: Google/Alphabet - I thought that Boston Dynamics now rolled up through Alphabet, and was totally separate from Google. Or is the author just pretending/ignoring the Google/Alphabet reorg. Or just doesn't know?
Regardless, seems like the sort of thing that should get a mention at least.
I think it's just that even though the company formally is called Alphabet now, people still call it Google and that's fine. Not like Alphabet/Google is working hard on rebranding or something.
The article claims BD didn't like Google's plan for a household robot on wheels because BD makes legs. This seems oddly specialized.
The bigger topic (probably not a revelation?) is how loosely integrated the robotic purchases have been, and will this continue?
I wonder how much more tightly those teams will ultimately be integrated. Acquisitions should ultimately be integrated and used to form the best teams for internal projects, you don't get the benefit of scale if they operate like an island...
Sounds more like, 'Google Takes Hands-Off Approach Before Hand-Off'.
The entire BD tech portfolio is about making really good legs, including thing like balance and recovery. So I can understand how they might be a bit miffed when Google wanted them to build something with wheels.
Of course they were. And I was frustrated (and creeped out) by their "wacky robot dog prancing through the forest getting kicked by the handler" videos.
Evolution seems to tell us certain things about legs. I assume they have reasons for clinging to this approach. But it creates an obvious perception problem because, man, those videos are weird.
Google's got rainbows popping out of their logo meanwhile these guys are punting pets. A culture clash surprises whom exactly?
I'm not correcting the guy's grammar for god's sake. I'm pointing out a subtle but serious prejudice and discrimination that is pervasive in our industry. But go on and cover your ears and pretend it's not happening
Indeed, but the goal of HN is to be interesting, and every thread turning into anger about the same four or five things is not interesting. I'm sure that wasn't your intent, but it's the default trend regardless of intent, so there needs to be a countervailing factor. Such discussions are predictable and achieve little except to get participants' juices flowing.
Both your comment to the GP and your comment here make uncharitable assumptions that convert immediately into anger. Please don't do that when posting to HN. It corrodes the civil discourse that we're trying for.
Yeah, I think your criticism is appropriate and hardly "off-topic" when it's a direct reply to a comment. The replies to your post did degenerate into a bit of a shit show, but that's not your fault.
Wait... if we select a 25 y.o. engineer over a 85 y.o. engineer it means ageism? We must pretend that cognitive functions doesn't decline with age? (Which has been completely proven as any biologist can tell you)
Yeah and white people are worse than asians at math and blacks are more likely to rob your store
The fact that there are still people that don't understand discrimination in this day and age is infuriating. If you think your interview candidate is potentially senile then you should screen for senility and not lump people based on statistical averages for whatever pigeon holes your constructed for them. People need a chance to surpass whatever stereotypes you hold even if they're based on statistically facts.
Not only is it down right illegal, it what perpetuates discrimination and causes so many of our social problems
PS: I'm sorry this is off topic, but this is a very important issue - and simplistic narrow minded thinking like this needs to be pointed out for the common good
No. Please no. It would be like screening 70 year old women for stripper roles. Yeah, maybe one of them is attractive to mass public, but still it doesn't make any sense, because the costs of screening all candidates is much higher than preselection. And if you have a problem with that then capitalism is your enemy, not "ageism".
but still it doesn't make any sense, because the costs of screening all candidates is much higher than preselection
The cost of screening 70 year old stripper candidates depends on how many such candidates their are to begin with. I doubt there are many, because the candidates select themselves out.
I really want to see a 85 year old trying to sue for being "discriminated" for a stripper role. The amusing part would be the tumblr rage over "supporting" the stripper.
Can you show any links to reliable research ahowing cognitive decline is inevitable in 70 year olds? Or is that just something you know and think is obvious?
I can't believe I'm being down voted for this. Is like when someone says males have naturally more upper body strength and then some feminists get mad, even when is written in a biology study about musculature.
This obsession for political correctness is corrupting scientific knowledge and research.
You're being downvoted because you're hyperfocusing on one dimension of people.
That 25 year old will expend many neurons and lots of that fresh cognitive capability on waste due to inexperience. Have you ever met a practicing 70-year old engineer? That perspective and experience has an outsized impact.
The male vs female upper body strength thing is totally different. For one, it's measurable. Second, it's not a job attribute that is offset by expertise -- a firefighter has to be able to carry an average sized person or a roll of hose. 30 years of hose carrying doesn't make it lighter. Usually, fire departments recognize this by allowing fire fighters to retire starting at the 20 year mark.
I know what you are trying to say, but I have known engineers in their 70s who have a really hard time grasping new concepts, we should be good with each other and nice with each other, but sometimes that is not in line with how the market/capitalism works. I don't think I was completely right but I don't think neither are you.
Brain plasticity is a real thing, and it goes away really quickly, is depressing but is a reality, I'm 27 but I already feel my reflexes are a bit worse than when I was 24, here is a nice video that gives some perspective about the issue, you may have seen it before (reverse bicycle): https://www.youtube.com/watch?v=MFzDaBzBlL0
It's easy to see why you are being downvoted. You precisely expressed the essence of discrimination and then defended it exactly the same way people always justify not hiring _____ race/class/gender/etc.
Diversity requires you change how you think about talent and work. Change from 80lb containers to 30lb containers and then you can hire people with less upper body strength while reducing injuries.
I know what you are trying to say, but I have known engineers in their 70s who have a really hard time grasping new concepts, and your splitting weight analogy doesn't cover such things, you can't "divide a concept".
Yeah, we should be good with each other and nice with each other, but sometimes that is not in line with how the market/capitalism works. I don't think I was completely right but I don't think neither are you.
Brain plasticity is a real thing, and it goes away really quickly, is depressing but is a reality, I'm 27 but I already feel my reflexes are a bit worse than when I was 24, here is a nice video that gives some perspective about the issue, you may have seen it before (reverse bicycle): https://www.youtube.com/watch?v=MFzDaBzBlL0
Yeah but why would I want to hire some inexperienced 25 year old that's just going to come in hung over from being drunk all the time and then raise my insurance costs due to venerial diseases, until I ultimately have to fire him for either incompetence or HR violations?
They didn't kill it, they sold it. They might have harmed it by causing internal conflict and altering the company vision, but it's still doing fairly well.
I know this may be horrifying to some, but there are those of us who would rather make a killing machine. We'll settle for a robot that carries supplies for soldiers or watches the enemy.
Defense contractor employees are typically patriotic in a way that would be really alien to the normal tech crowd. Part of the motivation is to do something for the USA. Having Boston Dynamics get diverted to cutesy home stuff is a decades-long setback for the nation.
There is also the matter of legs. Legs are cool.
Household robots are fiddly. They have to deal with a complicated environment. They can't just look at a toddler obstructing the path and go "Fuck it, I'm blowing that away, EAT FLAMING DEATH!!!".
Household robots are cheap plastic, with the shapes and colors totally determined by marketing.
When you buy a company that pretty much makes weapons, what kind of employees do you think you will get? Do you really expect that they will be happy that you have "rescued" them from being forced to make killing machines? Ha, ha, ha.... NO. You took away the coolness. You took away the awesomeness.