This is surprising to me. It seems that many prominent Googlers are very optimistic about robotics.
For example, in the FAQ of Jeff Dean's recent talk in Seoul, he mentioned how Deep Learning has a lot of potential to reinvent the field of robotics. Also, Demis Hassabis recently tweeted about progress in learning 3-D environments. I'd be surprised if Google wasn't looking into general purpose robotics...
Perhaps Google is disappointed in their robotics acquisitions and wants to start from scratch? It seems that they are farther on the software front than anyone at the point. I wonder what they'll do in their hardware/power divisions...
(Also, it kind of seems like Tesla and Google are on a crash course here. Tesla is ahead in power/hardware and is developing a top-tier AI team for self-driving cars. Elon also seems very interested in Robotics + AI. Google seems to be working from the opposite end.)
It's a big company so it should be no surprise that some sections of Google are very excited about robotics. For example you may have seen https://www.youtube.com/watch?v=iaF43Ze1oeI which is research done by Levine et al at Google. One thing to notice in the video though is that those are all custom arms. Someone was able to get $1M+ to build out that lab and the work is on going.
I think the thing that Google has recognized is that cutting edge robotics research that takes advantage of the work they have already done in deep learning doesn't require the kind of robots that Boston Dynamics (or many of the other companies they bought) build. Cutting edge robotics (from the perspective of learning systems) is back at the level of getting a single arm and a camera to learn basic tasks. In fact, there is a lot of research that can be done in simulation with these new deep methods that doesn't even require hardware.
The software and learning systems they can develop will very likely make their way into super advanced complex hardware like Atlas in the future, but for now it's probably not the best use of their time.
When I was in school in the late 90s it seemed like all of the robotics and AI research were in create algorithms to emulate behavior. Boston Dynamic's work seems to be extended from that. They purpose built a robot with purpose built balancing and walking algorithms. The being pushed over and getting back up demonstration seemed to confirm that to me.
Everything I'm seeing these days with AI research seems to be about developing systems that learn for themselves. Rather than telling the robot/AI exactly how to behave in a given situation, it's about allowing the robot/AI to experience as many situations as possible and learn what the appropriate response should be so that in the future it can independently identify and react accordingly.
I was also in AI school in the 90s and I agree with you to some extent. But when you think about nature, much of the behaviour of animals is not cognitive, but instinctive and physically dependent on their structure. In other words if you can build a robot that knows how to walk because you built it that way, why not do so and build more interesting learning at the decision making level?
Well the problem with that approach is that you have to account for every possible situation that might occur and program them into the robot. When a situation arises that you didn't anticipate then bad things happen. When a learning machine, it can teach itself, adapt and understand.
The former approach works fine for a factory floor robot that is in a controlled environment but doesn't lend itself well to other situations. An example might be a biped robot walking up a mountain and it falls. Most of the examples I've seen, the Boston Dynamics one included, have the robot detect it's falling and put itself into a crash position where it remains until it comes to a rest before attempting to recover.
If your robot is rolling down a mountain it might be destroyed before it comes to a rest. Having a feedback loop, reflex reactions and the ability to access the situation and recover dynamically would be much more useful.
Having a feedback loop, reflex reactions and the ability to access the situation and recover dynamically would be much more useful.
Agreed. But none of that implies learning. So why are you talking about learning in the beginning of your comment?
This happens all too often in AI conversations. Learning gives you a special and powerful kind of flexibility, of course. But not being able to learn doesn't imply it can't cope with an infinite range of situations. A robot that's unable to learn could be programmed with enough flexibility to walk on any surface it could possibly encounter.
I meant to say that a learning robot could adapt to a situation and try out possible solutions, measure success and adapt where as a preprogrammed robot would only ever try what it's been infused with. In my analogy of falling down a hill, the learning robot might not be able to stop itself on the first try but hopefully it might adopt a strategy that could allow it to regain control of the situation.
I think it does imply learning. Human babies "learn to walk", by using the feedback from lots of little experiments to improve their ability to navigate uncertain and varying terrain.
Basically, flexibility and the ability to deal with novel situations is close to synonymous with the ability to learn.
A self writing behavior tree that uses behavior to write new behavior based on past behaviors. Sounds human enough to become an addict of some sorts. :-).
Cutting edge robotics (from the perspective of learning systems) is back at the level of getting a single arm and a camera to learn basic tasks.
Which is a nice way to say that robotics really hasn't progressed that much in a sense an outsider would see as progress (where progress on outsider-terms would be Boston-Dynamics style robots that can, say, learn task and repeat them), though I'm sure researchers can point to a lot of progress on their terms.
To have a useful anthropomorphic robot you need better:
batteries
actuators
software
computers
Only in software and computers are we seeing fast progress. (Big improvements in ConvNet algorithms, and the fast video cards you need to run them) But actuators need to be more powerful and much lighter, and batteries need to store at least ten times as much power. Progress in these fields has been slow, since energy storage is a mature field, and you don't see routine doubling of performance like you do with CPUs.
And, of course, an economically useful anthropomorphic robot has to be dirt cheap, as well.
And, of course, an economically useful anthropomorphic robot has to be dirt cheap, as well.
Indeed, people talking about this subject often don't realize that humans are really cheap in many if not most circumstances (Boston Dynamics is working on poison-gas-protection-suit testing robot. Finally figured out a job a person wouldn't do).
And it's an evil equation where once a given task is mastered by robots, it makes humans cheaper in many other tasks - because it increases the competition and because it decreases the cost of maintaining the human.
So we've seen incremental automation and steadily declining living standards. Not a world that screams out the benefits of technology.
Yet another reason basic income would be useful. By leveraging existing capital to establish a more stable floor under that progression.
E.g. "Why would I do that for that much? I don't have any driving physical needs pushing me to poison/injure myself performing a dangerous menial tasks for minimum wage." Which puts a floor on human desire to do basic jobs. Which helps continue to support investment in improving automation/robotics. Which makes the world a better place.
Which actually sounds a lot like a carbon tax and the struggles alternative energy sources have gone through. Call it a self-aware employment tax.
Jonathan Hurst, the inventor of the ATRIAS robot, has argued that we have the actuators necessary and that efficiency solves some of the problems with batteries[0]. A brushless motor hooked up to a big reduction gearbox can have a pretty high torque density. If we design our robots to be light and design them such that they don't throw away energy with every step, then they can go further with current batteries.
Is the hardware really that much of a problem ? It seems like the Boston Dynamics robots are already very far along on the mechanical aspects. I've also seen videos of industrial robots which appear to work with greater speed and agility than humans. My impression is that solving the control problem robustly and rapidly is a bigger obstacle than mechanical limitations.
I do see the power source as being an issue. Many of the most impressive robots are tethered, but even a tethered humanoid robot could be extremely useful.
BigDog uses hydraulic actuators, and is powered by a two stroke gasoline engine. Not so good indoors.
Stationary industrial robots, the only real success story of robotics, have great speed and power, at the cost of incredible weight and power consumption.
15Kg payload, pretty okay, (Try holding a 15Kg weight at arm-length) but it masses 540Kg and has a rated power consumption of 5KW. (Three-phase power, of course) And that's just the arm! The NX100-FM controller it's specced with masses another 120Kg.
Mobile robots hate weight. Cutting weight forces a lot of other compromises, in speed, power, and cost.
EDIT: Also for the Motoman, is it possible it needs so much power because of how fast it can move that 15 kg mass around ? There's no beating conservation of energy.
If you studied robotics, you'd have learned that some walking/running problems are easier to solve when you have an engine with "unlimited" short-term torque. Low-powered electric actuators are very bad, hence BD uses gasoline engines to provide that short spike necessary for some differential equations to have a nice solution (e.g. one without power-pumping or without needing a few cycles around to reach your desired state).
Batteries have changed a whole lot since the 90's, though. Lithium-ion has a much higher energy density than lead-acid or nickel-cadmium. And without these improvements, many modern robots (roombas, quadcopter drones, etc) would simply not be possible.
The PR2 I discussed in the blog post does indeed have a lithium ion battery pack. It has a 1.3kWh capacity, (188 times bigger than an iPhone 6's battery! Probably part of the reason the robot massed 220 kilos) which gave it a rated... 2 hours of runtime.
Actuators, yes. At last. The power to weight ratio of motors has improved considerably in recent years. The current record is 5KW/kg [1], for a Siemens motor for aircraft. Tesla's motors are around 3.5KW/kg, although that may be peak, not continuous. Water cooling, as used by Tesla and Schaft, helps a lot. Schaft's innovation was to apply water cooling to small motors.
You can run electric motors far above their continuous rated values for short periods. Also, electric motors specifically designed for brief overloads (high-temperature insulation, temperature sensors) are quite possible. Every automobile starter motor is such a motor. With synchronous brushless motors ("brushless DC" and "variable frequency synchronous" motors are the same thing; motors above a few KW tend to be called the latter) and big power IGBTs, you can have huge torques briefly without much difficulty. If you have the electric power available.
Batteries, maybe. Running time between charges is going to be a problem for a long time to come. There's a huge battery industry trying to get energy density up, with modest success. For many applications, trailing a power cord most of the time is an option. Especially if the robot can plug itself in, which the Hopkins Beast was doing in the 1960s.
The power density of electric motors is pretty amazing. Take for instance quadcopters, a couple of brushless motors can lift their own weight, batteries, and still have plenty of thrust left over to accelerate[0]. Yes those power densities are probably for continuous power.
The problem is not power density, but torque density. Brushless motors spin really, really fast with low torque, which is the exact opposite of what we need for robots.
Electric motors have maximum torque at zero speed. Maximum power is at half of no-load speed.[1] You can design motors for higher torques at lower speeds; it's a standard design parameter.
The maximum torque an electric motor can produce is proportional to the magnetic field in the windings which is proportional to number of turns and current. Increasing the number of turns means more mass, increasing the current means more heat. More heat is particularly insidious, because as temperature goes up resistance increases, which makes more heat, which increases temperature, and so forth and so on.
SCHAFT found a solution to the more current problem with their ultracapacitor driven water cooled motors. Except one cannot drive said motors continuously and alone they still don't have that much torque, so if one wants more torque more windings must be added.
You can design motors for higher torques at lower speeds but the torque density suffers. Luckily we have compact high reduction gearing to transform high speed low torque to low speed high torque.
You've made some good points in this thread about actuation. Can you point me to more info on SCHAFT's actuation strategy? I've been trying to turn up info, but haven't found anything great.
>> For many applications, trailing a power cord most of the time is an option.
If the charging time was close to zero, would this problem be solved, for many applications ?
Or the other alternative, phinergy's aluminum air battery, which has 2000 wh/kg, but cannot be recharged, just replaced and "recycled", but probably in a cost effective way ?
Primary batteries have higher energy densities than rechargeable batteries. The military uses high-energy-density primary battery technologies for one-shot items like torpedoes. This might be worthwhile if you were building robots to help decontaminate Fukushima. Commercial applications, no.
Most of Google's robotics acquisitions were driven by Andy Rubin, who was trying to build a robotics division within Google. Once he left the company, it probably left their future uncertain.
It's always a bad sign to me when a company's strategy lurches based on who's around rather than on changing external circumstances. It suggests to me that the strategy isn't a coherent, collectively understood plan, but a political balance.
I think it's a stronger phenomenon in cutting edge areas, as there's less group-think to fall back on or use as a crutch.
Everyone knew what to do with a typewriter, even if the visionary leader left. If the person who was saying that microcomputers were going to be the next big thing quit? Nobody else has that idea in their heads.
Is this from a real-world example? Because I'd expect the opposite to be true.
Cutting-edge areas tend to attract people there for the vision. And the harsh commercial realities of innovative markets mean companies get in trouble if they get complacent. Whereas people in larger, older companies in stable markets can let their vision die and just go on doing whatever worked before. At least until it doesn't work, and then they're screwed.
In this instance, Google is a larger, older company.
Or to extend the analogy, if someone goes to IBM as director for New Technology X, and then leaves five years later, what's the likelihood people in New Technology X Division are going to be able take his or her responsibilities over?
Big companies are big companies and they usually don't encourage or reward employees overly much for striking out in a brave new direction. Which is hilarious given that they'll continously try and hire exactly those people externally.
Though I suppose you honestly can't encourage too much rebellion when you make your money from a crank being turned (and happen to need 1,000 bodies to just shut up, turn the crank, and get paid).
I think it's the common mode in American business culture the last few decades, but I think that's a result of the managerialist paradigm that has come to dominate.
Most startups avoid it to begin with. There's a really strong incentive for having a clear mission and high customer focus; many things get easier. Wikipedia has done a good job with "Imagine a world in which every single human being can freely share in the sum of all knowledge." Toyota has done a fantastic job organizing around the Toyota Production System. The best restaurants, bakeries, and the like generally have strong shared understandings. The same is often true of multi-generational family firms.
I think even Google did a good job for a long time rallying around "organize the world’s information and make it universally accessible and useful". Which makes this especially sad to me. Around the time of Google Plus and "more wood behind fewer arrows" I think there was a culture shift that is probably irreversible.
>> I think it's the common mode in American business culture the last few decades
Agreed. I find it doesn't hold for most contemporary Japanese corporations though. Their leaders tend to be switchable placeholders whose main purpose is to efficiently represent a consensus view.
Interestingly, though they do have momentum, they seem to lack strategic direction. Perhaps only charismatic leaders can provide the latter. Japan can produce such people (e.g. Morita at Sony) but the current environment values continuity over vision.
Very interesting. I suspect that Toyota is a similar result of a visionary leader. I want to believe there's some way to get the benefits of both approaches at scale, but I have yet to see an example.
Do you have anything you'd suggest I read to get a better understanding of the current Japanese situation? Most of my knowledge is about historic Toyota, which I'm sure gives me a distorted view.
Sorry, my opinion is formed only by observation and discussion with related parties. I don't have any direct experience of Toyota, except with one of their trading company's subsidiaries. Uniquely, this company does have a visionary leader at the helm, yet I believe there is no correlation with the parent's leadership style because Toyota Tsucho is run at arms length.
My rather uninformed opinion of Toyota Motors is that they are succeeding exactly as other Japanese companies used to succeed. If this is right, and I hope not, then they may be fated to see the same stagnation in time. A more optimistic view is that Toyota have something unique. If so, I don't believe it to be charismatic leadership. It would be baked into their culture.
Of note is Toyota's recent decision to invest $1B in AI research in Silicon Valley and Boston. They are trying to get ahead of the coming tech for autonomous driving and factory automation.
http://www.nytimes.com/2015/11/06/technology/toyota-silicon-...
Thanks! Yes, I'll be interested to see how Toyota goes. Although many people have taken the TPS lessons and applied them in software, Toyota itself is not one of those companies. (Indeed, their software appears to be terrible. [1]) That makes me think that the visionary leadership that created TPS is long gone, and that they are coasting. But yes, what they're coasting on has a continuous improvement component, so it could be that an advantage is a permanent part of the culture.
It is often true, but not always true. I have worked at or worked with companies that could express their strategy. But usually they are not big multi-division companies like Alphabet.
Yeah, but we're talking about shaping the future here: If you're relying on external circumstances for your hints as to what to do next, you're just reacting to the present....
It makes more sense once you take on a senior level role in those big companies, you start to see how a single, charismatic and convincing person can mentally mobilize (over a period of time) a large group of people to align with their hopes and dreams.
I was baffled by this as well, until I got closer to it and got to see "how the sausage is made".
It is amazing what a confident, strong personality can do in a group of people in _any_ setting.
You probably don't have much experience with larger companies. I envy you for that.
Yes, companies seem such opaque entities from the outside, but they are mostly driven by the wills of the topmost two layers. Remove key players from those layers and you see billions of dollars moving in strange new directions.
Uh .. that was his job. If he didn't do that he would have no job function in the company. If he couldn't change the faith of the company that employed him, then he was a bad employee at that particular job level.
That's what SVP/ director level / C-level people do for a living. They don't barely do any coding or any technical work - they create the vision and hire the people (or buy other companies) to make it real.
I am baffled by the cognitive dissonance here. This kind of single individual driving direction is LITERALLY the intended purpose of these high level roles.
Well, if you think of e.g. Marissa Meyer at Yahoo, big multinationals look like ships with so much momentum even the captain has trouble changing their direction.
Of course, reforming one arm of a business is a lot easier than an entire company, especially if nobody really depends on that arm.
Bill Gates famously "turned the battleship" at Microsoft in 1995 to focus the entire organization on the Internet. They actually did it quickly and succeeded (although more recently have lost their lead). So the right captain can make it happen.
I hear that analogy a lot, but on actual big ships, all the captain has to do is give the word and the ship turns pretty quickly. Maybe slower than a jet ski, but there's only one or two orders of magnitude difference there.
Likewise, sure, a big company has a lot of momentum but it shouldn't take more than a year to realign it.
You actually make a good point. The analogy gets used a lot because big ships do have big turning radiuses. But the better analogy is probably something more along the lines of the captain gives an order and various groups within the ship don't really agree with the order, so they say they're doing their part to make the ship turn but they're really not. Maybe the rudder have been moved a bit but the propellers are acting against the change. Furthermore, some percentage of the crew has decided that the ship should really be a plane and that's what they're working on.
Not really. On a ship (i.e. naval vessel), if the subordinates don't obey the commands, they get sent to prison for mutiny.
In a big organization, if the CEO says to do something (i.e. new vision... PKI's!), most people can ignore it with little chance of reprisal. It's much more of a political/social hierarchy than a military one.
A large company, like any large organization, is not a single thing. It's made up of individual parts, and parts within parts; all of them have different motivations. Frequently the most important competitors for some parts are other parts.
Think of it more like a feudal society. At the best of times, the headman at the top has the loyalty of most of the upper levels, and each of those has the loyalty of most of whoever is beneath them. But each of them has their own desires and plans: what they think will make the organization better or make their position in it better. At worse times, you have the War of the Roses or Game of Thrones sort of stuff.
The problem with bringing in outside organizations, like buying Boston Dynamics, is that they have their own, mostly fixed feudal structures, and integrating them is difficult. (Consider, at the very least, someone is going to go from the king of their own personal world to having to ask before they head to the executive washroom.)
A single person with a vision can have a great effect, because they only have to inspire (or convince) a small group of people around them to go along and that small group will bring along their own people. On the other hand, if the person with vision leaves the organization, whatever they were holding together falls apart pretty quickly.
(My favorite example is the Westinghouse(!)-CBS-Viacom thing.)
A single person can provide vision. A really good leader, gets his followers to want to follow them. If they leave, and nobody steps up to really lead...or if that leaders vision really isn't accepted then the organization is pretty much hosed. You can have the best group of employees in the world, and get by. Perhaps keep innovating for a few years. With an awesome leader at the helm, people show up extra early, they stay late, and they "believe." I've seen it too many times, in the military and civilian tech world. The "reality distortion field" that they speak of Steve Jobs possessing is real, and it's something only some people develop. Find those people and latch on, they'll take you for a fun ride.
I understand everyone comments but still find it not healthy that one man vision cannot be spread so other people around him to continue it even if he is not there anymore. I guess that is because that vision wasn't shared with a lot of people.
This is one of the reasons why BDFLs while being very effective, are also very dangerous. As when there is no equivalent replacement, the entire structure collapses.
There is a limited supply of geniuses. You lose one, all the B-players (that might consider themselves geniuses as well) will make sure you are going to take the wrong turns.
> It seems that many prominent Googlers are very optimistic about robotics.
BD was heavily geared towards defence contracts. That stream is gone. They will now have to do some painful restructuring, I guess Google don't want the bad publicity.
There are probably conversations going on that made the acquisition worth at the time (favour with public figures etc, Google has long been cozy with spooks and military types), and losing those contracts probably changed the equation significantly enough.
It seems like there is significant distance between being able to move forward in a stable direction and be a generically useful and sellable product.
There are a lot of problems to be solved before such a thing could work in even an Amazon warehouse along side humans in a cost-effective fashion - and the use of this thing is it's potential working-where-humans-work ability, otherwise it's easier to have automated bookshelves. And I'd choose an Amazon warehouse because it's an ultra-structured, streamlined environment, anything else would be harder.
Google's clearly interested in robotics, DoD, etc.; my guess is that this exit is for cultural reasons. Honestly, surprised they just didn't shutdown the company if things aren't going well.
Google is specifically not interested in DOD work. I don't understand why Google decided to buy a company so tied to DOD contracts if it did not want to continue that work.
This was what I was thinking when they first bought Boston Dynamic. Boston Dynamic was a huge DoD contractor. When Google bought Boston Dynamic, they explicitly said that they do not intend to be a military contractor and was expecting the existing contracts to expire [1].
The contracts that Boston Dynamic had with DoD expired in March 2015 [2]. My guess (BIG UNEDUCATED GUESS) is that the DoD doesn't take too kindly to its contractors not renewing its contracts and these military contracts contradict the corporate culture of Google. What's interesting is that they sold Boston Dynamic exactly a year after the DoD contract expired (donning conspiracy hat).
> My guess (BIG UNEDUCATED GUESS) is that the DoD doesn't take too kindly to its contractors not renewing its contracts
You may burn some bridges in your example but think of DoD like any other organization. If it's not getting the spotlight, there will be no hard feelings when things turn.
> I don't understand why Google decided to buy a company so tied to DOD contracts if it did not want to continue that work.
My experience in this area is quite limited, but it's surprisingly difficult to find companies doing this type of research that are not tied to the DoD in some way. DARPA is a major source of funding for these projects, and many of them wouldn't get funded at all without the government.
Case in point, at least until now: iRobot, maker of everyone's favorite robotic vacuums ... and military support robots. (Now spun off, but a core part of the company's strategy for two decades.) Its first years were DARPA-funded for the military applications.
Some of Tesla's autonomous vehicle work is in conjunction with an Israeli company called Mobileye which has autonomous highway driving demos of its own, though Elon has been quoted in the press as saying that's not all of it...
Google mostly bought Boston Dynamics to annoy the Pentagon, right after the Snowden revelations broke. DARPA sees Boston Dynamics' research as key for future ground-drone combat. Guessing somebody made a deal to get them back.
I find this implausible. I don't see Google spending that kind of money with out something sort-of resembling a plan[1], and didn't that purchase have to go through SEC approval? The DoD does have some pull with the SEC.
[1] Project Azorian not withstanding. Plus, it was funded by the government.
Don't underestimate the pettiness of billionaires, and the lengths they will travel to embarrass one another.
In fact, I can't think of a plausible reason for purchasing them and then putting them up for sale so quickly that doesn't involve political variables.
For example, in the FAQ of Jeff Dean's recent talk in Seoul, he mentioned how Deep Learning has a lot of potential to reinvent the field of robotics. Also, Demis Hassabis recently tweeted about progress in learning 3-D environments. I'd be surprised if Google wasn't looking into general purpose robotics...
Perhaps Google is disappointed in their robotics acquisitions and wants to start from scratch? It seems that they are farther on the software front than anyone at the point. I wonder what they'll do in their hardware/power divisions...
(Also, it kind of seems like Tesla and Google are on a crash course here. Tesla is ahead in power/hardware and is developing a top-tier AI team for self-driving cars. Elon also seems very interested in Robotics + AI. Google seems to be working from the opposite end.)