Guest_northshore_* Posted November 15, 2015 Share Posted November 15, 2015 I think the internet itself has created a monopolisation process that creates single winners in a range of fields both from 'apps' like Twitter, Whatsapp, Amazon etc. But also products bought via these channels much of it through price and customer experience information. This process both increases inequality...few massive winners and many losers not dissimilar to the career structure of footballers or artists. AND increases efficiency thru scale and destruction of the losers. This is probably right now a much more significant process than the much lauded robotisation/automation. Quite. As a network I'm not sure the internet and related tech progress could be much less practically co-operative or mutual. Quote Link to comment Share on other sites More sharing options...
Guest_growlers_* Posted November 16, 2015 Share Posted November 16, 2015 I think the internet itself has created a monopolisation process that creates single winners in a range of fields both from 'apps' like Twitter, Whatsapp, Amazon etc. But also products bought via these channels much of it through price and customer experience information. I would caveat this with the possibility that we are in a massive bubble - with tech companies at the epicenter - and the monpolisation trend could be due to the companies strategy rather than the internet. How many of these tech companies actually turn a profit? In Amazon's case the investor thesis seems to be that they will 'turn on the profits' at some point in the distant future i.e. the market is giving them huge leeway (that isn't afforded to other retailers) to pursue growth and investment at the expense of profits. How long will this last? Is this effect an artifact of a low interest rate environment? What happens if / when rates rise and/or investor sentiment changes? This article is specifically about Amazon: http://jessefelder.tumblr.com/post/133081731710/amazon-the-dumbest-competitor-in-americas-most Still, our loyalty to Amazon runs only as far as they offer the best price. As soon as someone offers a better price, consistently and with the same ease of purchase, our business will shift to them and away from Amazon just as quick as it shifted to Amazon and from its, “bricks and mortar,” competitors in the first place. To me, this suggests Amazon’s ability to, “turn on the profits,” is almost perfectly inversely correlated to their ability to generate sales growth and take market share. With respect to the whole automation debate (effect on net jobs / distribution of income etc), I am biased in that I don't want this to happen but I am struck by how unbalanced the whole debate seems (not on hpc but more broadly) i.e. everyone seems totally bought into the idea. It has the feeling of a social panic to me. Tim Halford had an interesting article on this recently: http://www.ft.com/cms/s/2/0b16754e-45c6-11e5-af2f-4d6e0e5eda22.html He queries why this IT revolution does not show up in productivity figures (perhaps the calculations do not capture new online goods?). He also mentions the hollowing out of the workforce - one of the individuals he references ascribes it to technology - and I recalled that this effect (Global loss to the Western working class) was also referenced in the Nangle's article on demographics: http://www.voxeu.org/article/labour-power-sets-neutral-real-rate Is it fair to say: that the impact of technology on jobs is highly uncertain? That (almost identical) predictions have be made time and time again and these have never been correct in the past? That there is little emperical evidence that recent technology leads to increased productivity? And that many of the outcomes (inequality, unemployment etc.) that have been ascribed to the impact of technology can be better explained by other factors? I also wondered why Haldane wrote a long rambling speech lamenting about Robots and jobs - something that might be an issue in 20 years - and didn't even mention the pending demographic cliff we have just passed (peaking of the ratio of working to non working population in 2012). I found that astonishing. Quote Link to comment Share on other sites More sharing options...
XswampyX Posted March 10, 2017 Share Posted March 10, 2017 Quote Link to comment Share on other sites More sharing options...
the_duke_of_hazzard Posted March 10, 2017 Share Posted March 10, 2017 My wife's a psychotherapist - I imagine she's pretty safe. Only thing safer I can think of is hooker Quote Link to comment Share on other sites More sharing options...
tyres Posted March 10, 2017 Share Posted March 10, 2017 (edited) 7 hours ago, the_duke_of_hazzard said: My wife's a psychotherapist - I imagine she's pretty safe. Only thing safer I can think of is hooker think again A Blowjob Cafe Staffed By Sex Robots Is Opening In London Edited March 10, 2017 by tyres Quote Link to comment Share on other sites More sharing options...
tyres Posted March 10, 2017 Share Posted March 10, 2017 An AI Completed 360,000 Hours of Finance Work in Just Seconds By 2030 students will be learning from robot teachers 10 times faster than today! Quote Link to comment Share on other sites More sharing options...
wonderpup Posted March 10, 2017 Author Share Posted March 10, 2017 (edited) Quote My wife's a psychotherapist - I imagine she's pretty safe. If some talking AI system with the ability to identify facial expressions, voice stress patterns ect and trained using real world therapy session data were to become available I could see the NHS being tempted to go down this road- not because it would be a good idea but because it would be cheaper and could be implemented on a mass scale. Every patient could have access to their own 'therapist' 24/7 at relatively low cost just by logging on to the system. The AI 'Therapist' need not be better or even as good as a human solution- as long as it worked to some degree and would be cheaper to implement then the temptation would be to offer it. The rationalisation would be the wider dissemination of the benefits of therapy- the real reason would be saving money on expensive professional services. One real barrier here would be patient confidentiality- any AI therapy system would need to be trained on huge amounts of actual case studies in order to be remotely plausible. But again if the argument could be made that using that data in a suitably anonymised form would benefit the thousands at present unable to access a human therapist then I could imagine some argument being made that sessions be recorded as grist for the AI psychotherapy mills of the future. Granted an unlikely scenario- but the same was said of self driving cars less than a decade ago. The problem with trying asses the limits and potential of AI today is that the latest deep learning techniques are scalable in a ways that previous methods were not- so some things that seem likey in the future may never be possible- while other things that seem unlikely might in fact prove perfectly possible. So your wife's job is probably quite safe- but it's at least possible to imagine scenarios that would reduce that margin of safety. The thin end of the wedge for the use of AI in the professions is the argument I laid out above- which is the idea that even an imperfect and inferior AI solution would be better than no solution at all- then- having got their metaphorical feet under the table via this rationale- AI's could slowly colonise the higher reaches of the professions by building up a reputational cache via their work among the less fortunate who could never afford a human practioner themselves but will settle for an erzatz AI driven substitute. Do I qualify for a free tin foil hat at this point? Probably- but there is an insidious plausibility to these scenarios that is hard to completely shake off. Edited March 10, 2017 by wonderpup Quote Link to comment Share on other sites More sharing options...
Economic Exile Posted March 10, 2017 Share Posted March 10, 2017 11 hours ago, the_duke_of_hazzard said: My wife's a psychotherapist - I imagine she's pretty safe. Only thing safer I can think of is hooker Only if there is enough nhs funding for mental health and/or enough people capable of paying her fees! Also a lot of people with mental health problems can now access more and more info/forums online and they find out that the bottom line is that they have to help themselves and build up a good support network with immediate access rather than wait for an appointment two weeks away! Quote Link to comment Share on other sites More sharing options...
Funn3r Posted March 10, 2017 Share Posted March 10, 2017 We have an r&d unit which comes up with all sorts of vaguely cool things which never seem to translate into anything saleable. They are trialling a diagnostic system at a hospital though where human doctors with 25 years experience can expect a 97% success rate. The magic box though can do 100% apparently. Quote Link to comment Share on other sites More sharing options...
ChewingGrass Posted March 10, 2017 Share Posted March 10, 2017 4 hours ago, tyres said: An AI Completed 360,000 Hours of Finance Work in Just Seconds By 2030 students will be learning from robot teachers 10 times faster than today! Seems inherently pointless if the AI can do all that, and if it is true AI then taking out a whopping loan for a university education is pointless as well. Don't think many people would fancy being fast charged with information (an education!) by a frikin robot. Quote Link to comment Share on other sites More sharing options...
wonderpup Posted March 10, 2017 Author Share Posted March 10, 2017 Quote We have an r&d unit which comes up with all sorts of vaguely cool things which never seem to translate into anything saleable. They are trialling a diagnostic system at a hospital though where human doctors with 25 years experience can expect a 97% success rate. The magic box though can do 100% apparently. Anyone claiming a 100% success rate for any technology has to be smoking something. But what if the choice were between a diagnostic system with a success rate of 85% or nothing? The way the NHS is going it's going take a month or more to get to see human GP anyway- so that magic box might start to look like a good alternative to some people, even with an error rate of 15%. It's like the argument that self driving cars would need to be 100% safe to ever be used and would need to be equipped with moral philosophy modules to navigate the trolly problem- (a scenario in which the car must choose between killing it's driver or some unfortunate third party(s)) Neither is really true- a self driving car that consistently outperformed humans on safety would be good enough,and the trolly problem is so vanishingly rare a situation that-for all real world applications- it could safely be ignored. It's an odd quirk of our attitude to AI technology that perfection seems to be the minimum standard deemed acceptable- in the case of all other technologies it's more or less accepted that the early implementations will be a little flaky- that they will improve over time. But we seem to have this species wide snobbery in the case of AI- unless it's perfect we won't give it the time of day, perhaps because the idea of artificial intelligence is threatening to us in ways that things like artificial muscle power or contact lenses are not. The former seems to touch a nerve in our self esteem that the latter do not. Quote Link to comment Share on other sites More sharing options...
wonderpup Posted March 10, 2017 Author Share Posted March 10, 2017 Quote Seems inherently pointless if the AI can do all that, and if it is true AI then taking out a whopping loan for a university education is pointless as well. It's a good point- a robot smart enough to train you to do something well ought to be smart enough to do that thing itself! But there's always; Those(robots) who can do, those(robots)who can't, teach. Quote Link to comment Share on other sites More sharing options...
Economic Exile Posted March 10, 2017 Share Posted March 10, 2017 37 minutes ago, Funn3r said: We have an r&d unit which comes up with all sorts of vaguely cool things which never seem to translate into anything saleable. They are trialling a diagnostic system at a hospital though where human doctors with 25 years experience can expect a 97% success rate. The magic box though can do 100% apparently. Doctors have a 97% success rate regarding diagnosis? I do not believe that. Heard so many people relating their stories of weeks, months, years going to doctors/hospitals before getting a successful diagnosis. Several deceased people I knew endured this trekking to find out what was wrong....then finally got a diagnosis.....then in a few days, weeks or months they died. Quote Link to comment Share on other sites More sharing options...
Arpeggio Posted March 10, 2017 Share Posted March 10, 2017 5 hours ago, tyres said: By 2030 students will be learning from robot teachers 10 times faster than today! Teaching them what? A skill that will lead to a Job? That isn't done by AI...? lol seems slightly self contradictory. Quote Link to comment Share on other sites More sharing options...
One-percent Posted March 10, 2017 Share Posted March 10, 2017 30 minutes ago, Arpeggio said: Teaching them what? A skill that will lead to a Job? That isn't done by AI...? lol seems slightly self contradictory. It's also very silly to suggest that the robot will teach ten times faster. I would think that the students will be left behind. Reminiscent of those dusty old profs who dronebon regardless of whether anyone is listening. Quote Link to comment Share on other sites More sharing options...
spyguy Posted March 10, 2017 Share Posted March 10, 2017 56 minutes ago, Economic Exile said: Doctors have a 97% success rate regarding diagnosis? I do not believe that. Heard so many people relating their stories of weeks, months, years going to doctors/hospitals before getting a successful diagnosis. Several deceased people I knew endured this trekking to find out what was wrong....then finally got a diagnosis.....then in a few days, weeks or months they died. No where close. There was a us study where a ai took a written problem sympton and competed to diagnose. The ai was 30% of the dr. Now, consider the symptons were written text. Imagine a setup where the ai gets a vast ocean of data yo sift thru, all digitised. Then youll see ais vastly out diagnosing humans. Quote Link to comment Share on other sites More sharing options...
spyguy Posted March 10, 2017 Share Posted March 10, 2017 6 hours ago, tyres said: An AI Completed 360,000 Hours of Finance Work in Just Seconds By 2030 students will be learning from robot teachers 10 times faster than today! Dont need a computer. Ive slways found paper books more useful than teachers. 60% of my gcse teachers were making it up as they went along. Quote Link to comment Share on other sites More sharing options...
Arpeggio Posted March 10, 2017 Share Posted March 10, 2017 18 minutes ago, One-percent said: It's also very silly to suggest that the robot will teach ten times faster. I would think that the students will be left behind. Reminiscent of those dusty old profs who dronebon regardless of whether anyone is listening. I think the optimum speed to teach would be to be always knowing the level of understanding of the student in order to know at what stage, technique and speed to use. I suppose a robot may eventually be good at that. A robot may be able to tell when a human is pretending to understand when something is really over their head. Then again a human may not feel the embarrassed / need to pretend they understand something in front of a robot. Perhaps a robot cannot make a child think it's "cool" or earn respect and therefore attention. Also.... Robo: have you revised?..........student: no (and what the heck are you going to do about it) human teacher: have you revised?.............student: err no (feels awkward and comes up with excuses, there is more motivation). Quote Link to comment Share on other sites More sharing options...
Economic Exile Posted March 10, 2017 Share Posted March 10, 2017 1 minute ago, spyguy said: No where close. There was a us study where a ai took a written problem sympton and competed to diagnose. The ai was 30% of the dr. Now, consider the symptons were written text. Imagine a setup where the ai gets a vast ocean of data yo sift thru, all digitised. Then youll see ais vastly out diagnosing humans. I confess that I can't fully decipher spyguy text (sorry) but I think the gist of what you're saying is that gps, even consultants might be ousted by AI. I agree. FFS, I can do better myself in comparison to the medical "profession" on the internet regarding an ailment and possible solution to a health problem. If the medics don't buck up in diagnosing and treating people appropriately, IMO, they'll be downgraded and pay lowered due to a combination of rational self diagnosis/treatment or AI systems that outperform them. Quote Link to comment Share on other sites More sharing options...
One-percent Posted March 10, 2017 Share Posted March 10, 2017 6 minutes ago, Arpeggio said: I think the optimum speed to teach would be to be always knowing the level of understanding of the student in order to know at what stage, technique and speed to use. I suppose a robot may eventually be good at that. A robot may be able to tell when a human is pretending to understand when something is really over their head. Then again a human may not feel the embarrassed / need to pretend they understand something in front of a robot. Perhaps a robot cannot make a child think it's "cool" or earn respect and therefore attention. Also.... Robo: have you revised?..........student: no (and what the heck are you going to do about it) human teacher: have you revised?.............student: err no (feels awkward and comes up with excuses, there is more motivation). A good teacher knows exactly where her students are. I'm not convinced at all that a robot can better a human at this. We are social animals. Teaching and learning is a social process. Yes, something simple and practical can be learnt from the Internet, instructions or video. I do this a lot myself. An example, today, my windows 10 computer failed yet again to communicate with my printer. Easiest way to solve it was a google search and follow the instructions. However, if you want to really understand something complex, it needs someone else to help you. This person will prod, probe and ask questions. You will be able to ask questions back. You will laugh, smile and chuckle together when you get it right, or even wrong. Quote Link to comment Share on other sites More sharing options...
spyguy Posted March 10, 2017 Share Posted March 10, 2017 32 minutes ago, Economic Exile said: I confess that I can't fully decipher spyguy text (sorry) but I think the gist of what you're saying is that gps, even consultants might be ousted by AI. I agree. FFS, I can do better myself in comparison to the medical "profession" on the internet regarding an ailment and possible solution to a health problem. If the medics don't buck up in diagnosing and treating people appropriately, IMO, they'll be downgraded and pay lowered due to a combination of rational self diagnosis/treatment or AI systems that outperform them. Yeah on tablet. http://m.slashdot.org/story/317409 The comparison was a bit of a scam. Getting a computer to analyse prose. Bit like challenging a dolphin to the 100m race. Imagine if the diagnosis was crunching gigabytes of numeric data and sensors? Quote Link to comment Share on other sites More sharing options...
spyguy Posted March 10, 2017 Share Posted March 10, 2017 28 minutes ago, One-percent said: A good teacher knows exactly where her students are. I'm not convinced at all that a robot can better a human at this. We are social animals. Teaching and learning is a social process. Yes, something simple and practical can be learnt from the Internet, instructions or video. I do this a lot myself. An example, today, my windows 10 computer failed yet again to communicate with my printer. Easiest way to solve it was a google search and follow the instructions. However, if you want to really understand something complex, it needs someone else to help you. This person will prod, probe and ask questions. You will be able to ask questions back. You will laugh, smile and chuckle together when you get it right, or even wrong. I had 2-3 competent teachers at school. The rest were winging it. There are not thatmany good teachers. Quote Link to comment Share on other sites More sharing options...
spyguy Posted March 10, 2017 Share Posted March 10, 2017 31 minutes ago, One-percent said: A good teacher knows exactly where her students are. I'm not convinced at all that a robot can better a human at this. We are social animals. Teaching and learning is a social process. Yes, something simple and practical can be learnt from the Internet, instructions or video. I do this a lot myself. An example, today, my windows 10 computer failed yet again to communicate with my printer. Easiest way to solve it was a google search and follow the instructions. However, if you want to really understand something complex, it needs someone else to help you. This person will prod, probe and ask questions. You will be able to ask questions back. You will laugh, smile and chuckle together when you get it right, or even wrong. Thats coaching. its mainly done by parents. Quote Link to comment Share on other sites More sharing options...
One-percent Posted March 10, 2017 Share Posted March 10, 2017 6 minutes ago, spyguy said: I had 2-3 competent teachers at school. The rest were winging it. There are not thatmany good teachers. There are slackers in every profession and many I've come across in my professional role are very good and dedicated Quote Link to comment Share on other sites More sharing options...
One-percent Posted March 10, 2017 Share Posted March 10, 2017 6 minutes ago, spyguy said: Thats coaching. its mainly done by parents. Nope, it is what a good teacher does Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.