July 27, 2017 at 12:34 pm
dld - Thursday, July 27, 2017 8:28 AMI think it's important to distinguish between robots and AI. Where I work we have a robot that stacks boxes of milk on pallets. It has displaced 8-10 people who used to do the task by hand, but it is far from an AI system. In fact it sits inside a cage because it is too dumb to shut itself off if a person walks into its space. There is a lockout system to gain entrance to the inside of the cage that automatically shuts off the robot and protects the human from being clobbered.
My concern is that as human beings we need work, not just to make money to meet our needs, but also to give our lives meaning. We feel good when we are productive and hopefully the world is better for our efforts. As work is redefined or eliminated due to machines (robots and AI) how do we adapt to the changes?
There certainly is a difference, but we also have robots that do things like scan systems for data, that do perform assembly, and more. Some of those are getting AI capabilities, not so that no humans are involved, but humans are involved less. Some manufactuers are using AI so that robots can do more than in the past, making adjustments to their programming based on the situation.
The concerns you raise are real, valid concerns. They worry me as well.
July 27, 2017 at 12:37 pm
Rodney Landrum - Thursday, July 27, 2017 9:27 AMAlexa, will there be a computer that can "think" and "reason" in the next 20 years?
"Sorry, I am not sure."
Alexa, can you think?
Sings Technology song.
Alexa, who let the dogs out?
Alexa, boots and cats.
Alexa, stop.I think we are all safe for now, Steve.
See you this weekend in BR.
See you there, but don't judge based on today. Could you have imagined the capabilities of Alex five years ago, ten? Not sure I could have.
July 27, 2017 at 12:43 pm
Rod at work - Thursday, July 27, 2017 8:31 AMI'm of two minds on this topic. I think there's a tendency to over romanticize AI. Like with Star Wars, where the robots are basically mechanical humans (C-3PO) or pets (R2-D2). Or they're more like the Terminator out to kill all humans because we've screwed up the world so badly. Then there's the robots who want to "be like humans", like Commander Data of Star Trek: The Next Generation. Just for grins, I just asked Cortana is she wanted to be human. Her response was, "No, but I have the deepest respect for humans. You invented calculus. And milkshakes." Cute. And most likely a canned response.But what does bother me is the disruption to employment. My son is on the autism spectrum. He's high functioning, but there's a lot he can't do that most of us take for granted. Because he's "on the spectrum" I'm more attuned to seeing it elsewhere. Locally there's an organization called Adelante, which seeks to help people with developmental disabilities. I've known some friendly young men who would do things like picking up trash, or ride in the trucks Adelante uses to pick up sensitive documents for shredding. What will happen to those people? To my son? In a world that the author in that HBR article wrote that have to improve the emotional IQ? Trust me, that ain't gonna happen for these people; for my son.
My sympathies, Rod, for your son and his future. What will become of him? I don't know. In some sense, I think the lower workers might survive well, because the robots and infrastructure to do very simple jobs, or those that need some basic human interaction, are too expensive. Pay a robot to change bulbs? Not unless we find a way to standardize lots of things. Even then, why? Find trash in the public world? Maybe, but I could see humans doing this.
Where I think this is disruptive is in more middle areas. Sell insurance? Why would we not use algorithms that might better analyze needs of humans in various ways. Humans can do this, but we can't necessarily keep all the variables in our heads all the times. AI/ML systems might do this much better. Those middle area jobs are places where I could see a few people with AI/ML systems handling monstrous loads of work.
July 27, 2017 at 1:36 pm
Steve Jones - SSC Editor - Thursday, July 27, 2017 12:37 PMRodney Landrum - Thursday, July 27, 2017 9:27 AMAlexa, will there be a computer that can "think" and "reason" in the next 20 years?
"Sorry, I am not sure."
Alexa, can you think?
Sings Technology song.
Alexa, who let the dogs out?
Alexa, boots and cats.
Alexa, stop.I think we are all safe for now, Steve.
See you this weekend in BR.See you there, but don't judge based on today. Could you have imagined the capabilities of Alex five years ago, ten? Not sure I could have.
I was hopeful in 1986 with ELIZA on a Commodore 64. Then in college teaching myself LISP. I agree that we have come a very long way absolutely but we will probably have a MagLev hover board and Land Speeder long before computers are able to reason. I would actually prefer a consciousness upload to the Cloud over AI so we can all be ghosts in the machine. Only then will we have the true AI...when the Soul Meets the Silicon (TM) (R). Just kidding on the trademarks. I just thought that sounded good.
July 27, 2017 at 1:46 pm
What we call AI is really just a form of information power tool. Huge amounts of information can be algorithmically sorted and searched, but not comprehended. The only true comprehension that occurs in that chain is when a human looks at the final output and integrates that into their knowledge and decides to take an action. Faster and bigger is good, but it does not fundamentally change what is going on. It's still impossible to have even a slightly intelligent conversation with Siri or Google Home. Any seeming intelligence is in the imagination of the human (as was the case with Eliza decades ago).
In the US colonial era, over 90% of human activity was dedicated to raising and producing food (now it's less than 5%). If you were to tell someone at that time 'we're going to build machines so powerful that a 100 acre (or 1000 acre) farm could be run by only a handful of people (which is the case now) the immediate reaction could well be 'what will we do for a job'. But the same forces that reduced the labor of farming opened up human ingenuity and endeavor for countless other jobs and wealth, now only a very small part of the population works in farming.
Of course, too rapid a change can be harmfully disruptive, but real change in the world almost always happens a lot slower than the advocates expect.
...
-- FORTRAN manual for Xerox Computers --
July 27, 2017 at 1:47 pm
It seems to me that for a company to simply replace employees with AI or robotics is taking a short view. Automated vending machines that sell hot sandwiches have their place, but very few people would consider that a replacement for a good cafe. Many industries like healthcare, financial services, or restaurants automate work that was previously manual, but yet they also continue to leverage a lot of human staff too in order to improve the overall service level of their business and remain competitive.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
July 27, 2017 at 1:55 pm
Eric M Russell - Thursday, July 27, 2017 1:47 PMIt seems to me that for a company to simply replace employees with AI or robotics is taking a short view. Automated vending machines that sell hot sandwiches have their place, but very few people would consider that a replacement for a good cafe. Many industries like healthcare, financial services, or restaurants automate work that was previously manual, but yet they also continue to leverage a lot of human staff too in order to improve the overall service level of their business and remain competitive.
The other day I was looking at one of those fancy drink vending machines where the robotic arm reaches up and grabs the desired product. Except, something had happened and several bottles had fallen down into the mechanism. The robot shut itself down with an out of order. It struck me that the dumbest of human stock clerks would have been able to evaluate the situation and place the bottles back on the shelves. Yet the fancy machine was helpless.
...
-- FORTRAN manual for Xerox Computers --
July 27, 2017 at 2:35 pm
Steve Jones - SSC Editor - Thursday, July 27, 2017 12:43 PMRod at work - Thursday, July 27, 2017 8:31 AMI'm of two minds on this topic. I think there's a tendency to over romanticize AI. Like with Star Wars, where the robots are basically mechanical humans (C-3PO) or pets (R2-D2). Or they're more like the Terminator out to kill all humans because we've screwed up the world so badly. Then there's the robots who want to "be like humans", like Commander Data of Star Trek: The Next Generation. Just for grins, I just asked Cortana is she wanted to be human. Her response was, "No, but I have the deepest respect for humans. You invented calculus. And milkshakes." Cute. And most likely a canned response.But what does bother me is the disruption to employment. My son is on the autism spectrum. He's high functioning, but there's a lot he can't do that most of us take for granted. Because he's "on the spectrum" I'm more attuned to seeing it elsewhere. Locally there's an organization called Adelante, which seeks to help people with developmental disabilities. I've known some friendly young men who would do things like picking up trash, or ride in the trucks Adelante uses to pick up sensitive documents for shredding. What will happen to those people? To my son? In a world that the author in that HBR article wrote that have to improve the emotional IQ? Trust me, that ain't gonna happen for these people; for my son.
My sympathies, Rod, for your son and his future. What will become of him? I don't know. In some sense, I think the lower workers might survive well, because the robots and infrastructure to do very simple jobs, or those that need some basic human interaction, are too expensive. Pay a robot to change bulbs? Not unless we find a way to standardize lots of things. Even then, why? Find trash in the public world? Maybe, but I could see humans doing this.
Where I think this is disruptive is in more middle areas. Sell insurance? Why would we not use algorithms that might better analyze needs of humans in various ways. Humans can do this, but we can't necessarily keep all the variables in our heads all the times. AI/ML systems might do this much better. Those middle area jobs are places where I could see a few people with AI/ML systems handling monstrous loads of work.
WOW, I hadn't even thought about insurance. Yes, I can see a disruption in that as well. And insurance is something which, at least to my point of view, is a specialized task. Speaking only for myself, I know that I don't understand insurance all that well.
In a similar vein this week I listened to Richard Campbell's RunAs Radio podcast episode, "Self-Driving Vehicles with Mark Minasi". It was a very interesting episode, where they discussed issues similar to what you brought up today. Both Uber and I think Lyft are working on producing self driving semi-trucks. In the podcast one of them mentioned that over 20% (I think it might be as much as 30%, but I'm not positive of that) of the USA's economy is tied to transportation, especially trucking goods. Replacing much of that with self-driving semis will have a really huge impact upon a lot of people's lives. That's going to hurt a lot of people. My brother-in-law makes his living driving semis in the eastern US. But I know that we, as consumers, will probably love it because it will likely mean lower costs of goods; after all you're not paying anyone to get it from point A to point B. I don't expect to see this happen in the next couple of years, but I do expect it will be reality in the next 10. Perhaps even the next 5.
Kindest Regards, Rod Connect with me on LinkedIn.
July 27, 2017 at 2:54 pm
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
July 27, 2017 at 3:32 pm
Rodney Landrum - Thursday, July 27, 2017 1:36 PM... but we will probably have a MagLev hover board and Land Speeder long before computers are able to reason.
reasoning is closer to sentience. "Smarts" are more like being able to do a better job based on the situation, without specific programming.
AI is closer to the latter than the former.
July 27, 2017 at 3:34 pm
jay-h - Thursday, July 27, 2017 1:55 PMThe other day I was looking at one of those fancy drink vending machines where the robotic arm reaches up and grabs the desired product. Except, something had happened and several bottles had fallen down into the mechanism. The robot shut itself down with an out of order. It struck me that the dumbest of human stock clerks would have been able to evaluate the situation and place the bottles back on the shelves. Yet the fancy machine was helpless.
That's an example of an automated machine that has no machine learning or AI. Just does tasks by rote. that's not an example of "fancy" AI or ML.
July 27, 2017 at 3:36 pm
Eric M Russell - Thursday, July 27, 2017 2:54 PMPlaying Devil's Advocate, I'm not sure who I'd trust [least] behind the wheel of a truck haling 40 tons of cargo on an 8 lane highway during a rain storm, an AI or a human driver who has worked three straight shifts without any sleep.
Depends on how many vehicles are also on the road in each config. I'd lean towards AI, but only in limited environments. Known routes, maybe only between cities or on the highway. Not necessarily in city traffic (highway inside the city) , which is a much larger problem domain
July 28, 2017 at 1:16 am
jay-h - Thursday, July 27, 2017 1:46 PMIn the US colonial era, over 90% of human activity was dedicated to raising and producing food (now it's less than 5%). If you were to tell someone at that time 'we're going to build machines so powerful that a 100 acre (or 1000 acre) farm could be run by only a handful of people (which is the case now) the immediate reaction could well be 'what will we do for a job'. But the same forces that reduced the labor of farming opened up human ingenuity and endeavor for countless other jobs and wealth, now only a very small part of the population works in farming.
The difference is that in order to operate these machines, you didn't need much brain power. If you have an IQ of 70, you can be a Farm Worker, truck driver, working - supervised - on building site but you are unlikely going to be a mechanical engineer. Mechanization has reduced the amount of work, it has pushed some people on the lower end of the IQ scale either to go on benefit (disability benefit included) or to work in the black economy (including thief, basic drug dealing, etc).
Now it is true that it is not worth creating A.I. for the unprofitable tasks such as the one relating to caring for old people, etc. But then, there are so many people that are been pushed out of the work force that salaries in these fields are going to drop even lower than they are now.
A.I. and software engineering are creating jobs but too few to compensate the jobs they are destroying. We, as DBA/Software Developer/Data Scientist are well positioned to profit from what is coming. But not so much anyone else.
In the legal industry, wages have been diving since the introduction of automated workflow (for example for immigration, conveyancing or for personal injury). Accountant are soon going to disappear as well. In exchange a few Data Scientists jobs are going to appears, as well as jobs that consist in data validation for A.I. (menial jobs paid under the UK minimum wage). But would a qualified solicitor can easily retrain as a Data Scientist? What about Tesco's cashier that lost his job for an automated till?
July 28, 2017 at 2:34 am
Hopefully the oil and coal will all run out before the AI figures out how to turn people into electricity?
Personally I find it very strange, and I say this as someone who makes a living from automating tasks which previously required human work, that we are so intent on replacing such an abundant resource (people who want work) with such an increasingly scarce one (technology and energy products which are heavily reliant on fossil fuels both for production and operation). Especially given that the side-effect of using the abundant resource is that we create more wage-earning customers to fuel our economies, and the side-effect of using the increasingly scarce resource is that we degrade and destabilize our ecosystem to a dangerous extent. From the point of view of any given company's shareholders I'm sure it makes perfect sense for the time being, but I don't think it really makes much sense for anyone else involved. A few top tier technical experts, maybe, but not even as many of those as we currently have. Where I work we've made maybe half a dozen technical support staff redundant this year (to be replaced by e-learning, self-serve information products, and increased automation in our helpdesk and data transfer processes) and I'm sure there'll be more of that sort of thing to come.
Maybe once all we clever technology types realize that it's also our own jobs we're making redundant, and not just those of the lower pay grades, we'll alter our opinions of the benefits of this kind of technological development. Even then, I think altering our habits to match those altered opinions will be the real challenge.
July 28, 2017 at 6:38 am
Steve Jones - SSC Editor - Thursday, July 27, 2017 3:36 PMEric M Russell - Thursday, July 27, 2017 2:54 PMPlaying Devil's Advocate, I'm not sure who I'd trust [least] behind the wheel of a truck haling 40 tons of cargo on an 8 lane highway during a rain storm, an AI or a human driver who has worked three straight shifts without any sleep.Depends on how many vehicles are also on the road in each config. I'd lean towards AI, but only in limited environments. Known routes, maybe only between cities or on the highway. Not necessarily in city traffic (highway inside the city) , which is a much larger problem domain
Maybe this is what they're already doing, I'll admit I'm not too familiar with how exactly they do their testing, their but one way of testing self driving AI technology in the field would be to install the AI and sensors in a vehicle but just not hook it up to the steering and break system. Just let an experienced human with a good driving record go about their daily routine through different variations of traffic patterns. If the AI makes identical decisions as the human driver over a period of several months, then they know the technology is proving itself. For those situations where the AI made a significantly different decision from the human driver, software and DOT engineers can playback the dash-cam video, sensor data, and decision tree logs to determine if it made the right call.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
Viewing 15 posts - 16 through 30 (of 65 total)
You must be logged in to reply to this topic. Login to reply