#robotics #artificialintelligence #economy #technologicalunemployment
#robotics #artificialintelligence #economy #technologicalunemployment
Originally shared by David Fuchs
I read the UK report on robotics and automation and job loss again this morning and realized something. The conclusion that 45% of jobs can be automated into nonexistence is off, they do not seem to be taking Moore's law into account in any significant way. The math breaks down this way, 20yrs times 12 months per year div 18 months per doubling (Moore's law). In 20 years computers will have undergone 13 doublings in processing power and will have become 8,192 times as powerful as they are now.
To put that into perspective, the new NVIDIA Tegra chip does 192 billion instructions per second, in 20 years, for the same cost, you will be able to get a chip that does 1,572,864 billion instructions per second, or a straight up 1,572,864,000,000,000 instructions per second. Or to really put it in perspective, your cellphone will be as smart as you are, and in some cases smarter than you. While that is a bit of an exaggeration, and computing will take several more years to reach that point, it does drive the point home.
Combine the processing power we will see in twenty years, with software capable of working without human supervision, and the UK reports employment numbers are way off, and look downright optimistic.
http://www.futuretech.ox.ac.uk/sites/futuretech.ox.ac.uk/files/The_Future_of_Employment_OMS_Working_Paper_1.pdf
http://www.fiercecio.com/story/another-study-paints-bleak-jobs-future/2013-11-26
#sciencesunday #robotics #automation #robotoverlords #employment
David Fuchs, you make a very good point with this analysis. Assuming Moore's Law doesn't break down, the next 10 years are going to see huge change, far less what will happen in 20.
ReplyDeleteMichelle Cameron Huge change doesn't even begin to describe what is going to happen over the next 10-12 years.
ReplyDelete3D printing becomes a mature technology, cheap energy and storage start becoming seriously disruptive, digital currencies become a hedge against inflation, robotics and automation cause massive job loss, immortality should occur around 2025, nanotechnology will exist in primitive form, and about 12 other technologies come of age. All converging on the year 2025, then in 2033 nanotechnology converges from about 20 different directions.
May we live in interesting times.....
David Fuchs you're getting ahead of yourself :P
ReplyDeleteKurzweill is predicting the Singularity around 2045, but I sense you're thinking this will happen sometime after 2033?
Michelle Cameron I think that Kurzweil is right about computing power, if there is a singularity it will probably not be a singular event. More than likely a whole many small ones, going on over the period of years to centuries. There will be a ton of people wanting to opt out for whatever reasons and so it will come in waves.
ReplyDeleteWhat I am talking about is the convergence of disruptive technologies on society. 2025 and 2033 are the big ones before Ray's singularity.
Each of the technologies I mentioned above (previous comment), have the ability to bankrupt pretty much any nation, and we are going to have these high technologies converge on one point in time, twice over a 10 year period.
David, I don't disagree with you regarding disruption, though I'm not as bullish as you on some technologies.
ReplyDeleteI do think we will experience a single point in time that we will look back on as the beginning of the singularity, and that moment won't last a day or several hours as is predicted by some Singularitarians. I suspect it might be a period of several days or weeks, simply due to propagation.
For quite some time I've felt that Kurzweil has mentioned 2045 to avoid the fate of most futurists who got it wrong. 2045 is sufficiently far enough out as to be more certain, and sufficiently close to say after the fact that nobody could really predict the Singularity.
I do however disagree that we will have immortality by 2025. Age reversal is likely, and the curing of many common diseases and disorders, but too many illnesses are simply not getting enough research money, personnel, and resources thrown at them yet to consider we'll have cures by 2025. In that respect, I'd think 2045 is more realistic.
Michelle Cameron Myself, Kurzweil , Aubrey De Grey, and a bunch of others have 2025 pegged (plus or minus a little) as the date when aging is figured out and possibly reversed.
ReplyDeleteGoogle seems to be following the singularity university philosophy of, affect a billion people in a short time span. From past investments google has a 5-8 year cycle on projects. Calico is google's venture into the anti-aging and immortality arena. They wouldn't have started the company if they didn't think was a solution could be found in under a decade. Personally, I think the 25 generations of cloned mice will be the key to cracking this. But hey, thats just me.... :-)
Kurzweil chose 2045 because that is where the exponential computing curve leads.
If the singularity happens, computers figure out everything, then do not just shut themselves down out of sheer boredom (Larry Niven twist to the singularity). I think we might see a place and a time as the first singularity with others that follow.
Consider it this way. The people that are part of this ever accelerating mass of information and technology are like an accelerating train with no speed limit. Once they get to a certain point, there is no way to just hop on, or even catch up to hop on. This is one of the reasons why I believe there will be several singularities over time. One of the others is some people will just not be interested, and their kids might be. There might end up being a future high tech version of the Amish Rum springa, where instead of being excommunicated you become god like. ;-)