An Automation Tipping Point? The Rise of 'Robotics as a Service' (venturebeat.com) 104
"Robotics-as-a-service (RaaS) is about to eat the world of work" argues Hooman Radfar, a partner at the startup studio Expa who's been "actively investing in and looking for new companies" catalyzing the change."
Companies buy massive robots and software solutions that are customized -- at great cost -- to their specific needs. The massive conglomerates that sell these robots have dominated the field for decades, but that is about to change. One major factor driving this change is how dramatically globalization has reduced hardware production costs and capabilities. At the same time, cheap and powerful computing and cloud infrastructure are now also readily available and easy to spin up. As a result, vertical-specific, robotic-powered, solutions can today be offered as variable cost services versus being sold at a fixed cost. Just as cable companies include the costs of set-top boxes in their monthly bill, robots and their associated software will be bundled together and sold in a subscription package. This change to the robotics business model will have profound implications, radically transforming markets and at the same time changing the future of work.
With a new variable cost model in place as a result of subscription packages, it's simple to calculate when a market is about to tip to favor RaaS. A market has hit its automation tipping point when an RaaS solution is introduced with a unit cost that is less than or equal to the unit cost for humans-in-the-loop to conduct the same task... One market that has already reached its automation tipping point is the enterprise building security market... Crop dusting ($70 billion), industrial cleaning ($78 billion), warehouse management ($21 billion), and many more service markets are tipping. When these sectors hit their automation tipping point, we will see the same level of industry disruption currently taking place in the building security market.
The changes taking place in the enterprise will also deeply impact consumer markets, and ultimately society, in profound and potentially challenging ways. We are at the start of a massive shift in how work gets done.
One study predicted the worldwide RaaSS market would be $34.7 within three years, according to the article, which also explores how the building security market is already being disrupted. "Instead of manning a building with three to four people, you can have one human managing a few remote robots" -- at a cost that's 30% cheaper.
"Moreover, all the data and insights collected via these robots is organized and made available for building and security optimization. It isn't just cheaper, it's better. There's no turning back -- this market has hit its automation tipping point."
With a new variable cost model in place as a result of subscription packages, it's simple to calculate when a market is about to tip to favor RaaS. A market has hit its automation tipping point when an RaaS solution is introduced with a unit cost that is less than or equal to the unit cost for humans-in-the-loop to conduct the same task... One market that has already reached its automation tipping point is the enterprise building security market... Crop dusting ($70 billion), industrial cleaning ($78 billion), warehouse management ($21 billion), and many more service markets are tipping. When these sectors hit their automation tipping point, we will see the same level of industry disruption currently taking place in the building security market.
The changes taking place in the enterprise will also deeply impact consumer markets, and ultimately society, in profound and potentially challenging ways. We are at the start of a massive shift in how work gets done.
One study predicted the worldwide RaaSS market would be $34.7 within three years, according to the article, which also explores how the building security market is already being disrupted. "Instead of manning a building with three to four people, you can have one human managing a few remote robots" -- at a cost that's 30% cheaper.
"Moreover, all the data and insights collected via these robots is organized and made available for building and security optimization. It isn't just cheaper, it's better. There's no turning back -- this market has hit its automation tipping point."
As before, as the future? No. (Score:2)
The future takes what we did in the past and builds on, not just copying it for the new form. Robots as a service sounds like a bad idea. The main reason Software as a service idea was created was the internet allowed instantly delivery, maintaince, and updating. Can't do that with hardware. Renting rarely makes sense unless you are getting an awful lot of service to the product or need to expand/shrink very quickly.
Also, I do not see security as something robots will take over. I could see them as
Re:As before, as the future? No. (Score:5, Insightful)
Robots as a service sounds like a bad idea.
And you're absolutely right for the reason you gave, but there's other reasons why some might get into Robots as a service. The initial cost of some of these service bots might be too high starting out and renting up-front would allow some to work on the cost at a volume side of the equation first and foremost and then offer their investment for others at a reduced cost. Much like, it doesn't take much investment to run an email server, but it takes quite a bit to run a large email domain properly. The cost of doing that, having the skilled labor, the iron to support it, the years of experience talent to run it correctly are all high to do "right". However, a company that's got folks doing nothing but offer that via an email provider, prevents your company from having to make that massive initial investment.
Can't do that with hardware
Absolutely, delivery will definitely be one of those things that make or brake a company getting into this. Faults and repairs are also going to be tricky, but that's backend if you have an army of bots at the ready. Delivery and getting bots into place on reasonable timescales for initial deployment and replacement of defective bots is going to be a real definer.
Also, I do not see security as something robots will take over.
I typically see them about the same as security cameras, but with wheels or flying in the air. So if you've got a drone that can hang out at 300 feet in the air and follow someone who just stole something, you'd have something that's definitely and augment of what we currently have. That said, it's really tricky to know ahead of time if this is even worth it. I totally concur with ya here. This is a hard sell and without a few more decades of advanment, I think we're a bit premature on security bots ruling the world.
Too many ways for a hacker to take control of a 'security' robot, turning it into the inside-man, or should I say, inside-machine?
Well, I don't think this is a fault with the people making the robots. Typically, bots come with audits of the systems to ensure there hasn't been a compromise. Of course, companies rarely do the usual maintenance of hardware and software, so........ Yeah, exactly.
Re: As before, as the future? No. (Score:5, Interesting)
You are correct a robot system is a lot more upfront even though it is cheaper in the long run than humans.
However two things currently apply to robots.
Robots are a capital expense, and in many cases tax deductible.
As well as
You can calculate the ROI, on how long until the savings pay for the robot.
With that you can easily get loans and lease agreements to cover the robots cost.
Speaking as someone who sells robotic systems for hazard enviroments
People misunderstand the security guards (Score:2)
Re:As before, as the future? No. (Score:4, Funny)
The future takes what we did in the past and builds on, not just copying it for the new form.
Actually I'm pretty sure this a case of PaaS (product as a service). You know like how you buy food at the grocery store? Well I have this brilliant idea where you have food that's already cooked for you, no need to deal with the prep, cooking, cleaning, dishes and so on. I will call it FaaS (food as a service). Now to patent it, before anyone else has the same idea...
Re: (Score:2)
Re: (Score:2)
Robots as a service sounds like a bad idea.
My Cherry 2000 [wikipedia.org] disagrees with you there, and is willing to debate the issue with you . . . after you have been "serviced".
Re: (Score:2)
You don't deliver the robot, you deliver the fruits of its work.
It's already quite common, e.g. 3D printing. You send it a design file, it makes it for you and ships you the resulting objects.
This will become more common. We are getting closer to fully automated electronics assembly, for example. You can already pay next to nothing for a factory in China to make your PCBs, populate them with components and do some basic tests, and a lot of that process is automated. Eventually all of it will be done by robo
Re: (Score:2)
Software as a service idea was created was the internet allowed instantly delivery, maintenance, and updating. Can't do that with hardware.
Pretty sure my TV is doing this, I know my computer is. As far as the hardware goes, two-day delivery is probably fine.
Too many ways for a hacker to take control of a 'security' robot
Well, no, there are limited number of ways for a hacker to do that and, depending on the setup and precautions, some won't work. Of the ones that will, the goal of security is not to prevent 100% of hacking, it's to raise the bar for what is required to do it so high that the only players that can are unlikely to see the time, cost, and risk as justifying the reward.
The real problem
Crop dusting? (Score:2)
Re: (Score:3)
I tried, but the more modern ones can actually regain their balance, no matter how much you try to tip them. Robot-tipping is a thing of the past, get back to cows.
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
If I ever set a company up that worked like that, I think I'd name it after a big river.
Re: (Score:2)
LOL I came to post that but was beaten to it. I'll have to revert to my secondary snark. Minimizing VC money loss to $34.70 allows for idiot investors to spread the wealth about in a much greater field of fiscal irresponsibility.
Invest it and they will lose it...
Re: (Score:2)
Re: $34.7?!?! (Score:2)
Somebody's gonna be a HUNDREDAIRE!!!1!
Hooman Radfar? (Score:2)
Remotely controlled Semi-Automated Robots (Score:2)
With cheap and fast international telecommunications the robots do not have to be fully automatic. They can be partially remotely controlled from a country with cheap labor.
I think we will see quite a bit of that.
Re: (Score:2)
Great, just with telephone service third world toilets scam us, now with robot minions they can actually break and enter and steal. Maybe even do arson, industrial espionage or even contract killing. What fun
Re: (Score:2)
Wrong. If they were remotely controlled by human laborers, they wouldn't be robots. They'd be waldoes.
Re: (Score:2)
> Noun. waldo (plural waldos or waldoes) A remote manipulation system in which a slave device mimics the motions of a master device manipulated directly by the operator.
These wouldn't fit the definition of a waldo either, because a waldo is a slave device that mimics the motions of the controller exactly. It's more likely these devices will have a lot of internal AI but get general direction from the human operator, thus not meeting the above definition of "mimics the motions".
This seems to be putting the cart before the horse (Score:2)
My experience, being in the mobile robot industry for the last 10 years, is that bringing a successful robot to market with a sustainable business model is incredibly difficult. Many have tried, but the vast majority have failed. The fundamental problem being there just isn't yet a very large overlap between what a robot can actually do well, and what people are willing to pay for a robot to do.
Rather than concentrating at this time on pie-in-the-sky infrastructures modeled on SAAS, it seems to me the eff
$34.7 within three years (Score:2)
"One study predicted the worldwide RaaSS market would be $34.7 within three years"
not sure on which currency, but I've got a $50 in my pocket and will happily buy this entire industry now.
Tesla tried it too, then chose humans. (Score:5, Insightful)
The training costs for robots can be immense. When manufacturing theory says a 3-axis robot can do a certain job, only to find that it can't be trained, and the training constraints mandate a 5-axis robot is necessary, with an increasing bump in training costs.
At some point, the flexibility, adaptability, trainability and observational and deductive capabilities of having a human "in the loop" become far more cost-effective than robots, especially in those tasks that most affect revenues and profitability.
I've worked on many robotic systems, and always encourage their use in places where humans would be exposed to needless risk. Nuclear power. Munitions factories. Mining. The list goes on.
When involving humans, the key is to prevent boredom and encourage active interest. Most of the time, that's the fastest path to both the market and profitability.
Toyota learned this when setting up their US production plants back in the '90's. As did GM with Saturn (RIP). Tesla relearned it more recently. Industrialists keep reaching to robots as a controllable cost that won't strike, only to find the hidden costs and delays can be far more expensive.
DIFFERENT (Score:2)
Those are the Model-T of the robots; to use a car analogy.
New robots have evolved beyond that. They still take time to adapt and train but that is a big part of the evolution. An old assembly line robot is custom made; now they are far more generic and mass produced - but need custom software. The other thing going on now is they need LESS custom software because they have some learning and adaptability ability. Applied Intelligence still takes work (real AI) but it's more capable than the techniques use
Re: (Score:2)
Re: (Score:2)
Same reason slavery went out of fashion. You have to buy a robot, maintain it, program it... with workers, you can just rent one and if it goes bad, throw it away and rent another one.
Re: (Score:2)
It's probably more complex than "it could have been done already".
The food line at McDs has to service a dozen different sandwiches. A robot would probably mess up kitchen layouts which can't easily or inexpensively be enlarged to accommodate it, and there's so many existing locations that would need to extensively remodel.
More surprising to me is why some VC hasn't gotten behind a new hamburger chain that IS as nearly fully automated as possible. This gives them the ability to provide a limited menu whi
Re: (Score:2)
McDonald's already has an image problem
Not with the hundreds of millions who eat there.
Every burger stamped out exactly the same over and over. "Mc - whatever" implies cheap, mass-produced items,
Which is why hundreds of millions eat there - across the world.
I imagine putting robots on their grills would destroy whatever remaining restaurant credibility they still retain.
Ask me how I know you don't have kids,
No-one - and I mean absolutely no-one - would care if all McDonalds workers were replaced tomorrow b
Re: (Score:2)
I've really wondered why McDonald's hasn't automated hamburger making yet.
Because teenagers making minimum wage(with no benefits) are cheap,
while robots are unreliable and fucking expensive.
We've been replacing employees with robots (Score:2)
This is the trouble with our current industrial revolution. It's not as immediate as the last one (where steam & waterfall powered looms replaced workers overnight and in mass). 5 million jobs lost to robots since 2000, with nothing but crap service jobs and Uber to replace them. It's bit by bit. And the people who lose their jobs blame themselves.
Test comment, Please ignore. (Score:4, Insightful)
Test comment, Please ignore.
, .
, .
Re: (Score:2)
Odd, I can't see it.
You may want to try again.
Re: (Score:2)
Was trying various unicode characters to see what /. can handle.
None of them.
Which is easier: kill a human or kill a machine? (Score:3)
which also explores how the building security market is already being disrupted. "Instead of manning a building with three to four people, you can have one human managing a few remote robots" -- at a cost that's 30% cheaper.
It takes a special kind of fucked-up person to kill a human security guard who happens to be in the wrong place at the wrong time. But how many burglars/thieves/industrial spies would hesitate to take out a robot? Rhetorical question the answer is obvious. Of course it won't even be necessary to blow them up or completely destroy them, they'll just get hacked and put out of commission that way. Then the human security guard supervising them will have to go find them, to discover that they weren't effective at all. It'll at best become an arms race between the criminals with robot hacking skills and the robot company engineers.
Re: (Score:2)
Of course it won't even be necessary to blow them up or completely destroy them, they'll just get hacked and put out of commission
The burglar will just press the big red button.
Re: (Score:2)
I'll just steal the robot as well. If I can't sell it en bloc, I can certainly take it apart and sell the parts, if everything fails at least the copper and other metals will have some value.
Ever tried that with humans? Unless you really know a few people on the shady organ market, it ain't easy to monetize a corpse.
Re: (Score:2)
Re: (Score:2)
While the hidden internal GPS tracker is reporting your position to the base station of the cellular network. By the time you have it torn down, the police will be at your door.
Re: (Score:2)
overhyped (Score:4, Interesting)
Re:overhyped (Score:5, Interesting)
Once upon a time, there were accountants and book keeping clerks who said the same thing - most large companies had a huge accounting staff, with dozens of book keeping clerks looking after their own column of numbers or area of data entry and consolidation. Fill out a number there, add it up, fill out the total there, pass the figure on.
They were destroyed by the advent of the spreadsheet and accounting software. Change one figure in one box, and all the calculations updated, no need for dozens of workers to do that job. Put a lot of people out of work.
The typing pool, the floors and floors of architects working on small parts individually, phone operators manually switching extensions etc etc etc. The computer did away with then all.
So yeah, this isn't new, this is just this iteration of the automation cycle - its already made its way through the white collar workers, and now its hitting the blue collar workers.
Re: (Score:2)
This isn't overhyped at all. In many places, it already exists. My son does maintenance and repair for a German company called Witron ( https://www.witron.de/en/autom... [witron.de] ) that specializes in warehouse automation. In Aurora, CO, they supply and run all the warehouse automation for King Soopers/Kroger, which supplies dry goods to over 230 stores in a several state region.
The only thing not automated in the warehouse is moving pallets on and off trucks. All orders in and out are received, packed/unpacked and
Stupidest shit I've ever read (Score:2)
Robot programs are highly customized to the customer's demands. This shit just makes no sense at all. At we haven't even talked about the latency problems with this whole cloud setup.
For anyone wondering what automated security (Score:2)
I'm actually a bit surprised Crop Dusting hasn't already been replaced by drones. Amazon's well on the way to warehouse management, but not sure what they mean by industrial cleaning.
Re: (Score:2)
I'm actually a bit surprised Crop Dusting hasn't already been replaced by drones.
Probably too hard to get a drone to fly like it's a 1920's barnstormer on coke. Some of those guys are crazy.
Why not? (Score:2)
Re: (Score:2)
They mostly make hamburgers; but on any given day you'll need to restock the straw dispenser, clean up the ketchup on table 5, open and close; do some cleaning in various areas, unload supplies from the truck; and so on. You could, if you wanted a tech demo, rebuild the entire operation around robots, with maybe one or two cleanup/customer-facing employees plus techs dispatched for
Re: (Score:2)
Plus add in the limited expansion space in most existing McDs kitchens and a ton of franchisees unwilling to gut their kitchens and start over.
The trick is probably to design an automated burger place from the ground up, starting with a limited menu that makes the automation process easier. By restricting what you serve, you simplify the automation.
It probably also makes sense to make the automation modular enough that any given store has a minimum of 2 automated burger lines. If one goes offline, the oth
Re: (Score:2)
With out current tech, having robots imitate humans and slot into their roles is just about doable as a tech demo; but ASIMO and friends are wildly expensive for something that lags behind a minimum wage teenager in terms of physical prowess and not requiring an engineering team for support.
All the settings where robots are delivering almost
Re: (Score:2)
Humans are cheaper in the long run.
With robots, you first have to buy them, you have to maintain them, you have to service them and every time you need to replace a part, it costs you money. Humans can easily be rented and if some part of them fail, you simply throw them away and rent another one.
Re: (Score:2)
McDonald's automated ordering kiosks are very much like pick-a-path books where you never seem to get the ending you wanted.
They are terrible.
Regarding automation, start with fries. A machine that can track time, temp, use weight to know when to make more, and a big freezer hopper with an automated arm or something to operate the fryers. Better quality/consistent fries, less danger to workers (ever had a grease burn?).
Not that kind of robots (Score:2)
They're probably talking about a different kind of robots.
Apparently there's a class of software that can manipulate the user interfaces of other (standard business) software to perform standard tasks like adding users or creating and sending invoices that take to long to do manually.
Aparently organizations prefer to add another layer of software instead of adapting the underlying software to better automate such tasks.
You could see it as a next generation of cross-program macro's, except that it's apparent
Design for robotic manufacture (Score:2)
Currently there is no good software for designing products for 100% robotic manufacture. The Tesla would be hughly profitable if those cars were designed for total rono
Hooman Radfar (Score:2)
OK, who else thinks that maybe Mr. Radfar is a thinly-disguised robot sent back from the future to bring about the Robopocalypse?
Anyone? Anyone?
Already Tppied Over (Score:1)
Competition abounds (Score:2)
We already have it (Score:2)
it's called software. People write software to automate tasks. In fact, tats allowed a company who would have 200 accounts in 1940, have 20 people today, that do more work, with more data, and can cut that data into reports no would would have thought was possible 80 years ago.
The idea robots means physical machines must die. It's just automation.
What about McDonalds? (Score:1)
I'm surprised no-one has asked why Maccas hasn't been automated yet.
well... (Score:2)