Showing posts with label AI. Show all posts
Showing posts with label AI. Show all posts

Universal Basic Income

Monday, April 24, 2017
It seems Ontario has now begun a Basic Income program in Hamilton, Lindsay & Thunder Bay.

http://www.cbc.ca/news/canada/hamilton/wynne-announcement-hamilton-1.4082476

The program will pay $17K for singles, $24K for families.  That's definitely a modest amount but probably reasonable for a pilot program.  Seems it will last 3 years of data gathering, then I suppose they will use that to either scrap it all together or I what I hope is to make it better.

Here's a good write up of this fascinating and ever more relevant topic...at the end of my "lecture" I point out some highlights from the link below, it's a bit long so you may prefer just to check that

https://en.wikipedia.org/wiki/Basic_income

One of the most common complaints critics have is that UBI will create a disincentive to work, to some degree this is true (and there is data on this -- check the Wiki) what most people who are not in High Tech or at least very plugged into the scene (such as your humble correspondent ;-) don't realize is that this question will be moot in as little as 10 years!  Sometime in the 2020's we will reach a true AI (as I've discussed in numerous posts!) at that point pretty much any job could be replaced by said AI.  Now, this is a complex subject which I don't want to go off on too big a tangent but these AI should have rights and hence they may not want to do all those jobs etc.  So how that will all shake out is hard to say.  But don't get too hopeful, Nanotech and 3D printing are also on the rise, so between these 3 amazing technologies the future of most jobs, not just manufacturing jobs but also most types of Lawyers, Doctors and Quants (such as myself)...etc will all be taken!

The point I'm trying to make is that  disincentive for work is irrelevant when there's no work!  So it's really with good foresight that many countries are seriously testing out what form of this program (likely will require a multi-variable approach in the sense that we change/adapt the form every 5 years or so to adapt to the changes technology is making to society) works best now!

Also, related to this are the people who don't believe in UBI, they usually advocate "jobs for everyone"...I find this to be a very antiquated mode of thinking!  It's like going back to the 1800's and outlawing textile looms because it put seamstresses out of a job (which incidentally what began the group called Luddites!).  For people who understand where technology is going they can see this is not the answer, there's been a long trend of automation, going back to the very first types of work, agrarian in nature...look how many people it now takes to run a farm, just a few!  As everything becomes more automated, we all have more free time (though parents to young children such as myself have trouble seeing that sometimes!) and will continue to do so as tech progresses.  Simple example would be a Nanotech based Replicator, like in Star Trek, this actually isn't that hard compared to many of the other future tech on that show.  No more shopping, no more cooking, cleaning up dishes etc.  Think how much more free-time we'd have!  My point is, if we force companies to keep human labor we will stifle this tech evolution, for instance we may not get to the stage where we have Replicators :-(  If you think about it, do you really want to work or is the more basic thing you want...money?  I know it's the latter for me so I certainly do not want to halt technology by handcuffing companies...UBI is the answer (what form is best?...that requires much research and experimentation!).  I do want to point out that companies will pay a significant portion of UBI, not sure how the optimal details will work out but take for instance a company akin to Microsoft that may gain a monopoly on AI labor, and another on Nanotech...these 2 companies would make Apple look like a Ma N Pa operation!  They would be ridiculously rich and so should pay their fair share to UBI!  If they don't...they lose anyway as nobody would have any money to buy their products!

Some people also question...what would you even do if you didn't work?  I think this is the ultimate 1st world problem! haha  I personally could list dozens of things!  But for those who lack imagination I'm sure we could run a variety of suggestions by you and at least a few of those would interest you!  Also the size of the UBI would limit what you could do but that's just a transitional phase, as Nanotech leads to buildings for example being created for a few $, money will go the way of the DoDo haha.  Even today with a modest UBI, you could study free online courses at great institutions like MIT on say M-Theory (potentially a Grand Unified Theory of Physics) or the collected works of Shakespeare or Mandarin...etc!  You could take up painting, carpentry...any number of activities!

Finally, there's many more interesting tangents to explore in this topic, but I have to do more research myself so let's save it for a future post!  I'll finish with a few highlights from the Wiki on UBI:
  • Administrative Efficiency - The lack of means test or similar administration would allow for some saving on social welfare, which could be put towards the grant. The Basic Income Earth Network (BIEN) describes one of the benefits of a basic income as having a lower overall cost than that of the current means-tested social welfare benefits,[9] and they have put forth proposals for implementation that they claim to be financially viable.[10]
  • First, although most basic income supporters tend to be politically left, right-leaning supporters, at least since the 1970s, have argued that policies like basic income free welfare recipients from the paternalistic oversight of conditional welfare-state policies. Second, Philippe Van Parijs has argued that basic income at the highest sustainable level is needed to support real freedom, or the freedom to do whatever one "might want to do."
  • Fox Piven holds the view that an income guarantee would benefit all workers by liberating them from the anxiety that results from the "tyranny of wage slavery" and provide opportunities for people to pursue different occupations and develop untapped potentials for creativity

Update:  A story about San Fran mulling a proposal to tax robots that take people's jobs...the money can be used to retrain those workers etc.

https://hardware.slashdot.org/story/17/05/02/1958221/san-francisco-politician-jane-kim-is-exploring-a-tax-on-robots

Science News: Event Horizon Telescope & a big step towards true AI

Thursday, April 13, 2017
The cool idea I introduced a few blog posts ago, a virtual telescope (from Hawaii to Antartica to Spain) with a combined diameter of the Earth (10,000K!) built to observe the Event Horizon of Black Holes has been turned on and taken it's first series of pictures.  The story below provides details, alas the data goes to supercomputers and it will take months to create a "picture" from the data.

https://phys.org/news/2017-04-astronomers-piece-image-black-hole.html

Just saw this, new TED Talk about the EVT
https://www.ted.com/talks/katie_bouman_what_does_a_black_hole_look_like

In other news, a great piece (from MIT's Technology Review zine) on a more mysterious but almost surely more powerful form of AI which has begun to come to fruition!  The article isn't short but well worth the read, touches on many tangentially interesting topics as well.  This form of AI is more biological in nature, it's an extreme in a way, you give the program the power to observe and experience then it develops it's own algorithms (genetic algorithm's I presume) to try and achieve a task.  NVIDIA used this approach with an autonomous car and got some promising but somewhat scary results, the car certainly was driving but by the very nature of the approach, a black box is essentially created that neither the programmers or engineers can tell you much about...it's akin to the AI created a brain and now we need to study it to figure out how it works!

https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/

I highly recommend this story to even my readers who aren't super interested in science, AI has become almost a regular dinner table topic of conversation these days.  A far cry from the small corner of science it held back when I began to follow it in High School (long time ago!).  It will affect more and more of your life as it becomes intertwined in all aspects of society, so probably not a bad topic to at least casually keep up on!

Carnegie Mellon AI beats top poker pros — a first

Wednesday, February 1, 2017
Yet another step towards the Singularity :-) AI claimed Chess back in 1997, Jeopardy in 2011, GO in 2016 and now, Poker in 2017!  Notice the progression is not linear at all!  Looks exponential to me, as Kurzweil has been saying!

http://www.kurzweilai.net/carnegie-mellon-ai-beats-top-poker-pros-a-first

The Supercomp running the code (learning based algos) is still on the order of 1000x times "weaker" than the human brain so there's some room for improvement ;-)  Also, let's not forget that when we build a computer with a comparable number of interconnections (Current estimates say 10^15 [Quadrillion] -- Synapses connecting Neurons in humans) then we will be no match for that computer for 1 main reasons, processing speed!  Our synapses fire at ~0.1 - 2Hz  (compare to a typical i7 firing at 3 Billion Hz [GHz]), furthermore, the speed at which these signals travel is very slow in Humans (286 Mph) compared to nearly at the speed of light in computers (186,000 Mph -- or the actual speed of light in Optical computers).

If we think about the firing rate as a kind of time keeping device, we can see why seconds became our shortest (in most cases) units of time (1 Hz is 1 cycle per sec)...compare that to a computer running at 1 Billion cycles per sec, convert that to 1 cycle per Nano sec.  If the computer equivalent of a second is a Nano second, then 1 Billion Nano seconds (i.e. 1 sec) is a long time to think for a computer...it is the equivalent of saying 1 Billion seconds is a long time to think for a human (10^9 sec = ~32 years!).  Consider how much you can accomplish in 32 years!  These computers that reach our level of interconnections, with their much faster processing speed can do as much each second!  I found a link that show's we are still years away from this (I say only years again due to the exponential increases) so you have some time for that info to sink in and make peace with it haha (I'll remind my readers again, this is why I say we join them rather than try to beat them...I'll leave the details of that for a future post!).

PS  Just saw this hilarious comment regarding the last link (simulating 1 sec of 1% of the human brain)

If they had used Donald Trump's brain they could have done it in real time by now :D

EDGE: WHAT DO YOU THINK ABOUT MACHINES THAT THINK?

Sunday, January 8, 2017
The Edge is a great site, their motto is "To arrive at the edge of the world's knowledge, seek out the most complex and sophisticated minds, put them in a room together, and have them ask each other the questions they are asking themselves".  At the end of each year they publish a question and answers from some of the most influential thinkers of our time.  Last year's question touched on one of my fave topics...AI!

The responses they publish are varied enough that you see arguments/points of view from pretty much all sides of the issue. My own view is that this is yet another area of human conceit that will eventually be overturned (think back to the Sun revolving around the Earth issue...where we literally thought we were the center of the Universe haha).  What non-experts don't usually realize is that computers (even Super Computers) had a computational capacity far far less than humans for so long that when AI (e.g. machine learning) programs ran on them it was never going to be a fair contest.  Even now, by the best estimates (from Computational Neuroscientists etc) our fastest Super Computer is ~1000x slower than the avg. human mind!  But in narrow tasks, even ones now that are not so much about number crunching (like the game of GO) AI can beat the best humans on the planet!

https://www.edge.org/contributors/q2015

I'll leave off by paraphrasing my fave Future guru Ray Kurzweil, he says one reason why most people have trouble believing in his timelines (his biggest prediction is that The Singularity will occur by 2045) is that they have not been trained to think exponentially, instead the default state is to think linearly which is good enough for most day-to-day experiences but not for appreciating the pace of science & tech. He has a great metaphor/puzzle for this, imagine a pond that on day 1 has only 1 lilly pad, if we assume they double daily and that on day 30 the pond is 100% full then on what day was it only 1/4 full?  The answer of course is 28, just work backwards...on day 29 it was half full!  The point is the doubters would be looking to be in the right (that the pond will never be overrun with lilly pad's) all the way till around day 28!

PS  So many Movies and TV shows now are based on these topics, I highly recommend these below!
http://www.metacritic.com/movie/her

http://www.metacritic.com/movie/ex-machina

http://www.metacritic.com/tv/humans

http://www.metacritic.com/tv/westworld