Author Topic: Robots and their impact on the future  (Read 537561 times)

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

scottish

  • Magnum Stache
  • ******
  • Posts: 2716
  • Location: Ottawa
Re: Robots and their impact on the future
« Reply #201 on: February 03, 2015, 05:21:37 PM »
That was pretty speculative for such a massive blog post.    It would be neat if he had labelled the axes on his graphs!

I'm going to take the opposing view and argue that artificial intelligence isn't going to keep improving on an exponential curve as Tim Urban speculates.

The first problem is that exponential growth curves don't last in the real world.    The growth rate quickly becomes so extreme that it encounters physical limits.    Tom Murphy has a good blog post on this topic here.

http://physics.ucsd.edu/do-the-math/2012/06/ruthless-extrapolation/

The second problem is that intelligence requires a feedback loop.   Your spam filter has a feedback loop to learn what you consider spam - every time you mark an e-mail as spam it updates rules that it uses to categorize an e-mail as spam or not spam.     This feedback loop is one of the reasons for the success of 'artificial narrow intelligence'.    All of these systems have sufficiently narrow scope that its easy - for degrees of easy - to design a feedback loop that allows them to learn.

But what about artificial general intelligence?    To achieve this type of intelligence, you need a much 'broader' feedback loop, much as a child has while growing up.   But computers system have neither the wide range of inputs (touch, hearing, sight, taste, smell) of a child, nor do they have the wide range of outputs - the ability to interact with the world and observe what happens.   Without these two elements, a machine will be unable to develop an equivalent intelligence to a person, roughly Tim Urban's definition of artificial general intelligence.      And these two elements require many more technological advances than processing power.

I know there are lots of engineers and computer scientists on this forum.    What do you say?

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #202 on: February 03, 2015, 08:55:56 PM »
The first problem is that exponential growth curves don't last in the real world.    The growth rate quickly becomes so extreme that it encounters physical limits.

Absolutely, and that's the first thing I thought of reading part 1.  Even if you build AGI to where it can improve itself, and it can work out the calculations more or less instantly to optimize itself, then it needs to get built to the new specs.  As soon as that's turned on, it comes up with a new version, but again, needs to be built.  Unless it has a way to build itself, but 3D printing isn't that exact yet.

Still, that's a relatively small problem.  Compared to the scale of issues we're talking.

I fall southeast of anxious avenue in part 2: I think there's a small chance of soon, but I'm optimistic on it.  Right about in the middle of quadrant 4, or slightly left of the middle of that box.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

trishume

  • 5 O'Clock Shadow
  • *
  • Posts: 26
  • Location: Waterloo, Canada
  • Student Programmer
    • My Website
Re: Robots and their impact on the future
« Reply #203 on: February 03, 2015, 10:20:19 PM »
Interesting articles.

There's a couple things the author didn't mention that are worth thinking about.

One is that the initial leap from kinda intelligent to super intelligent might be very difficult. If I suddenly gave you the ability to change the connections of any specific neuron in your brain at will, would you be able to make yourself smarter? If I gave you a thousand years? No, you'd probably end up dead. In order for an AI to make itself smarter it has to gain at least a basic understanding of itself. That's also the reason AI is so hard, we can't comprehend what the "algorithm" is for intelligence.

Also the whole premise of general A.I being possible in the near future is that exponential growth continues. This is briefly mentioned but IMO did not get the treatment it deserves. Reckless extrapolation of Moore's law can't alter fundamental facts of physics that may block our progress very shortly.

Another is what if general AI requires a breakthrough in physics or philosophy instead of computing? It's easy to say that brains are masses of neurons of which we can determine the functionality and emulate, but I certainly don't feel like a mass of chemicals as I observe and feel my surroundings and think my thoughts, although that's a very poor description of what it feels like to be conscious but it's a fundamentally difficult thing to describe. It's possible (perhaps unlikely) that what makes us intelligent is linked to what makes us conscious and it would be astronomically difficult to recreate in silicon.

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Robots and their impact on the future
« Reply #204 on: February 04, 2015, 12:18:56 AM »
It would be neat if he had labelled the axes on his graphs!

They weren't really graphs because there weren't really any quantifiable variables on most of those axes.  More like conceptual diagrams.

As soon as that's turned on, it comes up with a new version, but again, needs to be built.  Unless it has a way to build itself, but 3D printing isn't that exact yet.

I'm pretty sure he would argue that your tiny human brain has failed to grasp the ways in which it might iteratively improve itself.  Why would you 3D print a new brain when you can just have your nanobots make one?  Why build a physical prototype of your new brain at all when you can just run a computer simulation of a million new brains at once and see which ones work the best?  Why bother to simulate new brains when you can quantum realize all possible new brains simultaneously?  And that's just my tiny human brain thinking, surely a superintelligent being has better ideas than I do.

I certainly don't feel like a mass of chemicals as I observe and feel my surroundings and think my thoughts

Then you must believe that your consciousness resides somewhere other than inside of your body?  Because your body is definitely made of organized matter and energy, and that organization sure looks to me like it creates everything you have ever thought or felt.  You're a very complex machine, but still a machine.

jordanread

  • Guest
Re: Robots and their impact on the future
« Reply #205 on: February 04, 2015, 05:03:17 AM »
I'll chime in a bit later with more, but it was overall a pretty good article. This is something that I consider on a daily basis, and being involved with some NI systems removes any doubt that this will happen. I'll go into some more detail later, in regards to how this stuff may not even require the general intelligence mark to be reached, and some interesting facets of goal oriented programming. I fall on in between Confident Corner, and Anxious Avenue, but not necessarily due to my beliefs as to what exactly will happen. I'll bounce back to the original topic, but since this is the venue in which it's being discussed, here is what I am acting on:

1. I will work my ass off to enjoy life and position myself to do so by catching FIRE
2. I will catch FIRE
3. I will enjoy life
4. AGI (or equivalent) will happen
5. Life will change dramatically
6. If it changes in a bad direction, we'll all be extinct, but at least I got to enjoy my retirement
7. If it changes in a good direction, there wouldn't be a huge reason that we wouldn't live a hell of a lot longer, and if the concept of money and scarcity even exists at that point, I will have to look at my assets and figure out how to make them last indefinitely

More stuff later.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #206 on: February 04, 2015, 07:39:49 AM »
As soon as that's turned on, it comes up with a new version, but again, needs to be built.  Unless it has a way to build itself, but 3D printing isn't that exact yet.

I'm pretty sure he would argue that your tiny human brain has failed to grasp the ways in which it might iteratively improve itself.  Why would you 3D print a new brain when you can just have your nanobots make one?  Why build a physical prototype of your new brain at all when you can just run a computer simulation of a million new brains at once and see which ones work the best?  Why bother to simulate new brains when you can quantum realize all possible new brains simultaneously?  And that's just my tiny human brain thinking, surely a superintelligent being has better ideas than I do.

Sure, once it's at that point. I'm saying I think there are some physical limitations to it getting to that point. It won't necessarily slow it up a lot, but it may stop the instantaneous rocketship upwards.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

tomsang

  • Handlebar Stache
  • *****
  • Posts: 1085
Re: Robots and their impact on the future
« Reply #207 on: February 04, 2015, 09:29:56 AM »
Thanks for the links ARS. Very interesting.

jordanread

  • Guest
Re: Robots and their impact on the future
« Reply #208 on: February 04, 2015, 11:04:01 AM »
Thanks for the links ARS. Very interesting.

You get a chance to read that book?

tomsang

  • Handlebar Stache
  • *****
  • Posts: 1085
Re: Robots and their impact on the future
« Reply #209 on: February 04, 2015, 11:21:22 AM »
Thanks for the links ARS. Very interesting.

You get a chance to read that book?

I am traveling this week so I have it on my todo list. Good airplane reading.

2lazy2retire

  • Guest
Re: Robots and their impact on the future
« Reply #210 on: February 04, 2015, 11:52:56 AM »
What I always find funny about these stories is that after 120 years most of us are still driving around in the same technology first made popular by Mr Ford, maybe we are not advancing fast enough:)

CCCA

  • Pencil Stache
  • ****
  • Posts: 631
  • Location: Bay Area, California
  • born before the 80's
    • FI programming
Re: Robots and their impact on the future
« Reply #211 on: February 04, 2015, 12:37:55 PM »
Wait But Why just had a great two part blog post on AI.

http://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html

http://waitbutwhy.com/2015/01/artificial-intelligence-revolution-2.html


That was quite a read.  Very, very interesting.  Not sure how much I totally buy all the arguments, especially about how soon it will come but it is interesting to think about how different the world might be with human level intelligence in everything and then moving beyond that to super intelligence. 


I wonder if the "goals" of AI beings will necessarily be constant and rapid improvement or if they will get satisfied and say "You know what, I think I'll just relax and enjoy this existence".

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #212 on: February 04, 2015, 01:04:22 PM »
That was quite a read.  Very, very interesting.  Not sure how much I totally buy all the arguments, especially about how soon it will come but it is interesting to think about how different the world might be with human level intelligence in everything and then moving beyond that to super intelligence. 

Ditto.  As I said above, I think there's a small chance of it happening soon.

I also recognize that I'm not an expert, and find it interesting that the experts in the area think it will happen much sooner than I do (based on their median and long term guesses).  I recognize that I should probably defer to their expertise on the estimate, even despite my skepticism.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

2lazy2retire

  • Guest
Re: Robots and their impact on the future
« Reply #213 on: February 04, 2015, 02:23:40 PM »
With the possibility of ASI and immortality just around the corner has anyone rerun the numbers on the 4% rule to see if it still holds up

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #214 on: February 04, 2015, 02:27:09 PM »
With the possibility of ASI and immortality just around the corner has anyone rerun the numbers on the 4% rule to see if it still holds up

Money becomes pretty irrelevant with ASI cause you'll likely be knocked off the balance beam.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

2lazy2retire

  • Guest
Re: Robots and their impact on the future
« Reply #215 on: February 04, 2015, 02:30:08 PM »
With the possibility of ASI and immortality just around the corner has anyone rerun the numbers on the 4% rule to see if it still holds up

Money becomes pretty irrelevant with ASI cause you'll likely be knocked off the balance beam.

That's an article for martketwatch - "The Balance Beam and the 4% Rule, what you need to do NOW"

EscapeVelocity2020

  • Magnum Stache
  • ******
  • Posts: 4811
  • Age: 50
  • Location: Houston
    • EscapeVelocity2020
Re: Robots and their impact on the future
« Reply #216 on: February 04, 2015, 02:33:26 PM »
Interesting read (I love the graphics!), but I'm not sure why all the anxiety.  For instance, one very plausible outcome would be similar to what the Matrix concept was all about (I highly doubt the 'synthetic humans living forever on Earth' outcome described in the article).  Virtual Reality is on the cusp of becoming 'a real thing', and you only have to look around to notice younger people love interacting with screens as opposed to nature when faced with 'down time'. 

Computing power is expected to continue to follow Moore's Law, so imagine being able to live anywhere, at any time, communicating with anybody and everybody, in an experience that is tailored to your preferences.  Would it be such a horrible outcome to live a much longer life in a virtual reality that was as optimized as possible to your interests (albeit it once that is totally divorced from our physical reality, aging, atrophying, etc.).   

And before you dismiss all of this as reactionary, I originally listened to Nick Bostrom on EconTalk last December discuss all of the same points (http://www.econtalk.org/archives/2014/12/nick_bostrom_on.html) and also a more calming follow-up podcast on AI with Gary Marcus (http://www.econtalk.org/archives/2014/12/gary_marcus_on.html).

I'm agnostic as to when or if Kurzweil's 'singularity' will occur, but optimistic that any outcome will be 'the right one'.  Article's like the end of Part 1 and most of Part 2 make me wonder why stir up worry?  All of Part 1 basically says this will happen in the blink of an eye and be beyond our comprehension, that it is inevitable we will either go extinct or be immortalized in exchange for our free will, so if you take that as a given, then what does worrying about it accomplish?  Unless you think all of humanity is, say, going to hit the stop button, and somehow voluntarily and forever become some sustainable, peaceful Amish super-race...

So which future would you prefer I guess:  The Matrix or The Amish?

(Sorry, this stuff is really fun to discuss and hear other people's thoughts on, so with this I'm finally subscribing to this thread!!)

EDIT:  I thought when ARS said 'south-east', that was the 'pessimistic' quadrant on the chart, but it is 'less chance of happening soon' and east was optimistic. 
« Last Edit: February 04, 2015, 02:42:32 PM by EscapeVelocity2020 »

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #217 on: February 04, 2015, 02:40:28 PM »
Interesting read (I love the graphics!), but I'm not sure why all the anxiety Rebs.

Huh?  I have no anxiety around any of this.  I clearly stated above I'm in the optimistic camp, so I don't know where you got that from.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

2lazy2retire

  • Guest
Re: Robots and their impact on the future
« Reply #218 on: February 04, 2015, 02:42:33 PM »
Interesting read (I love the graphics!), but I'm not sure why all the anxiety Rebs.  For instance, one very plausible outcome would be similar to what the Matrix concept was all about (I highly doubt the 'synthetic humans living forever on Earth' outcome described in the article).  Virtual Reality is on the cusp of becoming 'a real thing', and you only have to look around to notice younger people love interacting with screens as opposed to nature when faced with 'down time'. 

Computing power is expected to continue to follow Moore's Law, so imagine being able to live anywhere, at any time, communicating with anybody and everybody, in an experience that is tailored to your preferences.  Would it be such a horrible outcome to live a much longer life in a virtual reality that was as optimized as possible to your interests (albeit it once that is totally divorced from our physical reality, aging, atrophying, etc.).   

And before you dismiss all of this as reactionary, I originally listened to Nick Bostrom on EconTalk last December discuss all of the same points (http://www.econtalk.org/archives/2014/12/nick_bostrom_on.html) and also a more calming follow-up podcast on AI with Gary Marcus (http://www.econtalk.org/archives/2014/12/gary_marcus_on.html).

I'm agnostic as to when or if Kurzweil's 'singularity' will occur, but optimistic that any outcome will be 'the right one'.  Article's like the end of Part 1 and most of Part 2 make me wonder why stir up worry?  All of Part 1 basically says this will happen in the blink of an eye and be beyond our comprehension, that it is inevitable we will either go extinct or be immortalized in exchange for our free will, so it you take that as a given, then what does worrying about it accomplish?  Unless you think all of humanity is, say, going to hit the stop button, and somehow voluntarily and forever become some sustainable, peaceful Amish super-race...

So which future would you prefer I guess:  The Matrix or The Amish?

(Sorry, this stuff is really fun to discuss and hear other people's thoughts on, so with this I'm finally subscribing to this thread!!)

I'm on board with the "not worry about it" stance. As I see it ,my timeline suggests that personal extinction is likely to occur around the time of AGI to be followed soon by ASI and immortality, so get f@cking on with it already.
« Last Edit: February 04, 2015, 02:44:22 PM by 2lazy2retire »

EscapeVelocity2020

  • Magnum Stache
  • ******
  • Posts: 4811
  • Age: 50
  • Location: Houston
    • EscapeVelocity2020
Re: Robots and their impact on the future
« Reply #219 on: February 04, 2015, 02:43:06 PM »
Interesting read (I love the graphics!), but I'm not sure why all the anxiety Rebs.
Huh?  I have no anxiety around any of this.  I clearly stated above I'm in the optimistic camp, so I don't know where you got that from.

I added my EDIT as quick as I could!  Edit to add 'one more thing' (since no-one has responded yet, but I wish we got the same 'warning - while you were typing others have posted' for these modifications / Edits...).  I probably wouldn't have typed up a response if I had thought the consensus on ASI was 'optimistic', funny those unintended consequences :)
« Last Edit: February 04, 2015, 02:53:29 PM by EscapeVelocity2020 »

scottish

  • Magnum Stache
  • ******
  • Posts: 2716
  • Location: Ottawa
Re: Robots and their impact on the future
« Reply #220 on: February 04, 2015, 03:16:54 PM »
This:

Quote
Computing power is expected to continue to follow Moore's Law,

It's the opposite.    Computing power is reaching physical limits and will stop following Moore's law soon.   For the last several years many of the advances in computing power have been achieved by parallelism.   Which also has its limits.

Is Nick Bostrom a philosopher?

rocketpj

  • Pencil Stache
  • ****
  • Posts: 963
Re: Robots and their impact on the future
« Reply #221 on: February 04, 2015, 04:00:11 PM »
Well, if there is a chance of avoiding personal extinction through the rapid bootstrapping on an ASI then I guess I'd best make a few more trips to the gym.  It would suck far too much to be the last person to die before the species suddenly became immortal.  Like the poet who died on the last day of WWI.

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Robots and their impact on the future
« Reply #222 on: February 04, 2015, 07:14:38 PM »
It would suck far too much to be the last person to die before the species suddenly became immortal.

They made a whole movie about this:  Mr. Nobody.  It's about the very last human being to ever die of natural causes.

Chuck

  • Bristles
  • ***
  • Posts: 407
  • Age: 35
  • Location: Northern VA
Re: Robots and their impact on the future
« Reply #223 on: February 06, 2015, 12:07:13 PM »
Conciousness transfer... man if that was possible... It'd basically be like "Ghost in the Shell", right?

I would enjoy being a robot. Unfortunately I think the odds of living to see that day are very, very small. Perhaps if I save enough 'stache my kids could be early adopters in their old age though.

tomsang

  • Handlebar Stache
  • *****
  • Posts: 1085
Re: Robots and their impact on the future
« Reply #224 on: February 09, 2015, 10:57:17 AM »
Funny Dilbert on topic!

EscapeVelocity2020

  • Magnum Stache
  • ******
  • Posts: 4811
  • Age: 50
  • Location: Houston
    • EscapeVelocity2020
Re: Robots and their impact on the future
« Reply #225 on: February 09, 2015, 02:21:44 PM »
I hadn't seen that one yet, but there is a whole series about 'emotionally manipulative robots' that started around Feb. 4th (http://dilbert.com/strip/2015-02-04)...

« Last Edit: February 09, 2015, 02:25:06 PM by EscapeVelocity2020 »

OSUBearCub

  • Bristles
  • ***
  • Posts: 397
  • Age: 43
  • Location: Orlando, Florida
  • Tackling student loan debt/not saving dryer lint.
Re: Robots and their impact on the future
« Reply #226 on: February 09, 2015, 03:28:55 PM »

Leisured

  • Pencil Stache
  • ****
  • Posts: 696
  • Age: 79
  • Location: South east Australia, in country
  • Retired, and loving it.
Re: Robots and their impact on the future
« Reply #227 on: February 10, 2015, 02:42:35 AM »

Some other poster pointed out that no matter how intelligent machines become, we can always switch off power if we have to. It reminds me of a comment made about the Daleks in Dr Who; run up a flight of stairs. We do not need to make ourselves vulnerable to intelligent machines. We do not need to build very intelligent machines that are free to move about. Mobile robots only need to be intelligent enough to do a narrow range of tasks.

Consciousness is as much a mystery today as it was in the past. Giant computers predict weather and help design new aircraft, but they are not conscious and follow code to the letter. There is no reason to suppose that computers that learn, and so are not entirely predictable, are conscious either.

Research into artificial intelligence is unregulated, but it might be regulated in the future, in the way that pharmaceutical drugs are regulated.

jordanread

  • Guest
Re: Robots and their impact on the future
« Reply #228 on: February 10, 2015, 08:35:40 AM »
Ha! I'm loving the fact that this thread started getting active again (no offense tomsang...but most of those links you posted were not very conversation inducing, although they were interesting :P).
So I'll go ahead and thank rebs for the outstanding link that picked this back up. That was a pretty well written article. I had to be careful not to go down a rabbit hole on that site. :)

So I said I will follow up, and here I am. First, I will provide some responses to a couple of the things brought up. Then I will provide some background on my foray into this field.

The first thing was brought up by several people, and it's a somewhat valid point. The physical limitations. But let's think about that for a moment. I'm not even talking about ASI at this point (that comes later), but rather AGI. For those of you who didn't read the articles, or aren't familiar with the term, that means an artificial intelligence that matches a human in computations and understanding, Artificial General Intelligence. So let's break out the facts from the speculation.
  • Narrow AI, or AI that does something specific (i.e. Watson, Google, Google Now, Siri) currently exists
  • Narrow AI gets better every day (although technically it gets better with every use).
  • Our current and recent breakthroughs and benchmarks in computing power and capability are in large part due to parallelism, not fitting more transistors on a board.
  • Our ability to manipulate matter at the molecular level is primitive at best.

If you've used a phone made in the past couple of years, you should know that the cool features behind it are made possible by narrow AI. We use it all the time, and when it works, it's practically invisible. Have you ever scanned a document into Word and made it editable? Uploaded it to Google Docs? Optical Character Recognition (OCR) is what makes that possible, and it was only able to reach the current level of accuracy because it was coded to learn how to read written characters...it wasn't coded to read written characters. Think about that for a moment. We didn't have the time, the skill, the patience, and the ability to handle all of the different kinds of hand writing that exist. So we didn't bother. As the price of processing power continued to drop, we stopped being limited by the time it would take to learn. So those first two already exist, and I know that I use them multiple times per day (not so much OCR, because seriously...who deals with the hand written word anymore?).

So some people mentioned that exponential growth may not continue, and we don't see it coming because humans are very good at extrapolating patterns, even when there aren't any. The Ruthless Extrapolation article made some really good points, and had some pretty cool examples of when that was proven wrong. I don't know that it necessarily applies in this case, but rather than argue with it, let's assume we are wrong, and Moore's Law does not continue. Remember that Moore's Law (or the extrapolative version of that, Ray Kurzweil's Law of Accelerating Returns) was initially based on the physical. The pure number of transistors that could be fit in the same amount of space. However, as someone pointed out, the fact that we are still meeting the estimated processing power has been in part due to parallelism. Some use this to show that this level can't continue since the underlying assumptions based on hardware are expected to slow (at least with out current method of manufacturing). That could be a valid point. Even Moore himself says:
Quote from: Gordon Moore
It can't continue forever. The nature of exponentials is that you push them out and eventually disaster happens"
However, I would posit that in this particular case, it's not actually relevant to the technology required. Rackspace, Google Cloud Services, Amazon Web Services can all provide incredibly cheap and incredibly fast processing power. We already have enough linked and parallel processing power to power an AGI, we just haven't coded it yet. So the physical things that people worry about, and the arguments they raise against them are valid for now. However, that does not mean that damage can't be done to our society and our way of life, even without physical bodies.

I have some actual work to do now, so I will come back later and address some other points that were brought up. Namely (and so I don't forget):
  • Feedback and inputs required
  • Anecdotal points regarding different fields and their progress
  • Black box systems and examples
  • Goal oriented programming
  • Friendly AI
  • Regulation
  • Ethics
« Last Edit: February 10, 2015, 08:37:34 AM by jordanread »

2lazy2retire

  • Guest

OSUBearCub

  • Bristles
  • ***
  • Posts: 397
  • Age: 43
  • Location: Orlando, Florida
  • Tackling student loan debt/not saving dryer lint.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #231 on: February 10, 2015, 12:04:12 PM »
Some other poster pointed out that no matter how intelligent machines become, we can always switch off power if we have to.

I guess you didn't read the Wait But Why article.  I no longer believe this is true.  ASI will be able to power itself without us, IMO.

We do not need to make ourselves vulnerable to intelligent machines. We do not need to build very intelligent machines that are free to move about. Mobile robots only need to be intelligent enough to do a narrow range of tasks.

Sure, we don't need to.  But we already are trying to make them capable of moving through all obstacles.  And when these machines are often built to be mobile killing units (via the military), such as drones, we are inherently making ourselves vulnerable.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

Grid

  • Bristles
  • ***
  • Posts: 463
  • Age: 10
  • I kept dreaming of a world I thought I'd never see
Re: Robots and their impact on the future
« Reply #232 on: February 10, 2015, 03:17:08 PM »
I've been following the thread with intrigue for the past few months and just wanted to toss my 2 cents in.  I just graduated with my Master's in CS with a focus on machine learning.  Machine learning itself is the field that is slowly pushing its way towards the AGI mentioned in the Wait But Why article.  Neural networks (also known as multilayer perceptrons) are the most flexible and currently most popular algorithm to use.  They have been around since the 60s, but it's only now that computing power has caught up with the ideas and made them useful.

As computing power grows, we'll see neural networks, support vector machines, and random forests evolve into more complex chains and pipelines of data flow (similar to the brain's architecture).  I say this mostly based off of Michael Jordan's comments on his AMA on Reddit about 5 months ago here: 

http://www.reddit.com/r/MachineLearning/comments/2fxi6v/ama_michael_i_jordan

Current models that people (including myself) build in general have one rule, and that's to reduce the model's error on a given set of "training" data.  The model is given this data, and it trains itself until it has reached a global or local error minimum. 

Here's what I find interesting, that no one has really mentioned yet.  The change to a model's structure when it gets a guess "wrong" is known ahead of time.  The model adjusts its parameters in a known fashion so that it is more correct on the data the next time around.  The model has no inherent desires, and if you want to put a human spin on it, it's selflessly serving the programmer's wishes to have a hunk of code that can make some nice predictions.

I think once we add "wants" and "needs" into the equation for an AI we will start to have problems.  We humans are machines, and we're programmed from birth to like food, sleep, and sex, among other things.  We develop our entire lives around these ideas, but the machines we're building have no desires yet.  Once we build the concept of self into an AI, I think we should start to get worried.  Unfortunately, I think the idea of a machine with its own desires is intriguing for many, and may even be the only way to achieve ASI.  We'll just have to be careful.



Some other poster pointed out that no matter how intelligent machines become, we can always switch off power if we have to.

I think the problem comes when the machines will actively not want us to switch off their power.
« Last Edit: February 10, 2015, 03:28:05 PM by Grid »

scottish

  • Magnum Stache
  • ******
  • Posts: 2716
  • Location: Ottawa
Re: Robots and their impact on the future
« Reply #233 on: February 10, 2015, 04:01:32 PM »
Quote
ASI will be able to power itself without us, IMO


Free/really cheap energy?   That would solve a lot of problems.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #234 on: February 10, 2015, 04:48:23 PM »
Quote
ASI will be able to power itself without us, IMO


Free/really cheap energy?   That would solve a lot of problems.

Quite possibly.  Though remember "problems" for us aren't necessarily problems that ASI would care about at all.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

jordanread

  • Guest
Re: Robots and their impact on the future
« Reply #235 on: February 10, 2015, 05:04:41 PM »
Quote
ASI will be able to power itself without us, IMO


Free/really cheap energy?   That would solve a lot of problems.

Quite possibly.  Though remember "problems" for us aren't necessarily problems that ASI would care about at all.

I actually had a post written in response to that earlier. Most of the research being done is to create something that can take a goal and solve it. I would think that our goals and the machines (assuming covert A[G|S]I) would line up nicely. Plus I would guess that something decentralized would be preferred (based on a study I read a few years back regarding utility infrastructures).

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #236 on: February 10, 2015, 06:13:40 PM »
Quote
ASI will be able to power itself without us, IMO


Free/really cheap energy?   That would solve a lot of problems.

Quite possibly.  Though remember "problems" for us aren't necessarily problems that ASI would care about at all.

I actually had a post written in response to that earlier. Most of the research being done is to create something that can take a goal and solve it. I would think that our goals and the machines (assuming covert A[G|S]I) would line up nicely. Plus I would guess that something decentralized would be preferred (based on a study I read a few years back regarding utility infrastructures).

That's fine, but if it solves that problem just before it hits ASI or after, and that goal is finished for it?
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

jordanread

  • Guest
Re: Robots and their impact on the future
« Reply #237 on: February 10, 2015, 06:45:48 PM »
Quote
ASI will be able to power itself without us, IMO


Free/really cheap energy?   That would solve a lot of problems.

Quite possibly.  Though remember "problems" for us aren't necessarily problems that ASI would care about at all.

I actually had a post written in response to that earlier. Most of the research being done is to create something that can take a goal and solve it. I would think that our goals and the machines (assuming covert A[G|S]I) would line up nicely. Plus I would guess that something decentralized would be preferred (based on a study I read a few years back regarding utility infrastructures).

That's fine, but if it solves that problem just before it hits ASI or after, and that goal is finished for it?
Then we wouldn't be able to shut it off. I think that focusing on the ASI aspect is irrelevant post-AGI. The question is more of how we would react. If it's friendly, or covert, we would take the win and give it another goal. Perhaps genetic engineering for immortality.  If it's not so friendly, we'll try to get it turned off, fail miserably, and die horribly. At that point, let's just hope the AI has a fondness for a diverse ecosystem, so that our little planet can keep being blue, green and awesome.

Leisured

  • Pencil Stache
  • ****
  • Posts: 696
  • Age: 79
  • Location: South east Australia, in country
  • Retired, and loving it.
Re: Robots and their impact on the future
« Reply #238 on: February 11, 2015, 11:12:00 PM »

Grid pointed out that ‘The model has no inherent desires, and if you want to put a human spin on it, it's selflessly serving the programmer's wishes to have a hunk of code that can make some nice predictions.’

Good way of putting it. And later Grid suggests:
‘I think the problem comes when the machines will actively not want us to switch off their power.’

We have all had the experience of thinking about a problem, and then the solution pops into our heads. We are not aware of thinking, and our brain is ‘selflessly’ (to use Grid’s term) solving our problem for us. Our bodies selflessly digest our food, our hearts and lungs automatically and selflessly work a little harder if we climb stairs, and our eyes selflessly alter focus if we concentrate on something at a distance.

I have goals and desires, but do not fear that my brain or any other part of me might take me over. We need to watch developments in AI, but I do not expect that advanced machinery will ever get to the point that they ‘actively not want us to switch off their power,’ or take us over.

AI is an interesting research project, but it is not clear to me that we need machines cleverer than we are. We need servants, nothing more. If we made a metal Jeeves, he might be a threat, but do we want a metal Jeeves?

matchewed

  • Magnum Stache
  • ******
  • Posts: 4422
  • Location: CT
Re: Robots and their impact on the future
« Reply #239 on: February 12, 2015, 05:05:44 PM »

Grid pointed out that ‘The model has no inherent desires, and if you want to put a human spin on it, it's selflessly serving the programmer's wishes to have a hunk of code that can make some nice predictions.’

Good way of putting it. And later Grid suggests:
‘I think the problem comes when the machines will actively not want us to switch off their power.’

We have all had the experience of thinking about a problem, and then the solution pops into our heads. We are not aware of thinking, and our brain is ‘selflessly’ (to use Grid’s term) solving our problem for us. Our bodies selflessly digest our food, our hearts and lungs automatically and selflessly work a little harder if we climb stairs, and our eyes selflessly alter focus if we concentrate on something at a distance.

I have goals and desires, but do not fear that my brain or any other part of me might take me over. We need to watch developments in AI, but I do not expect that advanced machinery will ever get to the point that they ‘actively not want us to switch off their power,’ or take us over.

AI is an interesting research project, but it is not clear to me that we need machines cleverer than we are. We need servants, nothing more. If we made a metal Jeeves, he might be a threat, but do we want a metal Jeeves?

You seem to be proposing two different things here (let me know if I don't have it correct).

a) That AI will merely be an extension of ourselves and therefore cannot hurt us. I think this line is demonstratively false as we as humans have made many things that are literally without a concept of self and many of them have caused harm. How are the two concepts (selflessness and capability to take over) connected? This could cause harm in a manner which doesn't involve a "take over". The eternal example of the AI programmed to make paperclips. And it then turns everything, even people, into paperclips. That is not servitude that the AI just did, just doin' its job.

b) The need for machines being (and I think you chose the wrong word here using clever) smarter than us isn't there, we just need servants. No offense but you and I both live in the same world right? GPS, the computer that you typed your message on, the internet which allowed you to have the platform for your message, in fact most modern conveniences for the last 30 years or so have come about from a revolution in being able to generate systems (programs) which can do things better than we can. And frankly it's just another example of humans doing it for other aspects of what some call human actions being replaced by X. It would sound silly to say why do we need mass production, we don't need machines that are more precise than we are, we just need servants. Why do we need cars to drive us places, we don't need cars that can go further and faster w/out getting tired, just need those trusty servants. My point being is that you seem to gloss over the amazing things that can come about with building, teaching, and then learning from a machine much more capable at learning than we are.

You've also glossed over the last bit of what he Grid said. You just right it off as if it's not possible but there is nothing to suggest that consciousness is a unique human feature. If it isn't unique to us then can't it be generated in other machinery (biological or not)? The answer is yes as far as I can tell. I'm pretty sure there are several animals on earth we could agree on being conscious. So given enough technological advancement why can't that be replicated in programming? After all isn't DNA just programming?

Leisured

  • Pencil Stache
  • ****
  • Posts: 696
  • Age: 79
  • Location: South east Australia, in country
  • Retired, and loving it.
Re: Robots and their impact on the future
« Reply #240 on: February 14, 2015, 03:56:02 AM »
Spectacular misunderstand, Matchewed, but it might have been my fault in not making myself clear. Of course we need machines like mass production, cars, aircraft, GPS, telephones! These devices are not smarter than we are , but they are much better. Better to ride in a car than walk - of course. By servants I meant all the machines we use can be seen as servants, like a toaster, electric kettle, computer, telephone.

Nobody worries that cars or telephones will 'take over' our society, but the discussion of AI makes people fear that smart robots could 'take over' our society. That is the difference. I drew a parallel with Wooster  and Jeeves, and me and a robot Jeeves.

We do not know how to program a robot to be conscious. My point is we do not need to try and program consciousness, or robot wants and desires. The giant computers that predict weather are servants, and are not conscious. Make each of these computers a thousand times more powerful, and they are still just computers.

I follow what is known as 'Lady Lovelace's conjecture', where Ada, Countess Lovelace, worked with Charles Babbage on his Analytical Engine in the 1840s, in England. She was adamant that the Analytical Engine, a mechanical computer, did exactly what it was programmed to do, no less - and no more. She had had a computer language, Ada, named after her.

The Analytical Engine did perform computations, but was impractical. Ada was the daughter of Lord Byron, of all people. Clever lady.




matchewed

  • Magnum Stache
  • ******
  • Posts: 4422
  • Location: CT
Re: Robots and their impact on the future
« Reply #241 on: February 14, 2015, 07:00:27 AM »
Spectacular misunderstand, Matchewed, but it might have been my fault in not making myself clear. Of course we need machines like mass production, cars, aircraft, GPS, telephones! These devices are not smarter than we are , but they are much better. Better to ride in a car than walk - of course. By servants I meant all the machines we use can be seen as servants, like a toaster, electric kettle, computer, telephone.

Nobody worries that cars or telephones will 'take over' our society, but the discussion of AI makes people fear that smart robots could 'take over' our society. That is the difference. I drew a parallel with Wooster  and Jeeves, and me and a robot Jeeves.

We do not know how to program a robot to be conscious. My point is we do not need to try and program consciousness, or robot wants and desires. The giant computers that predict weather are servants, and are not conscious. Make each of these computers a thousand times more powerful, and they are still just computers.

I follow what is known as 'Lady Lovelace's conjecture', where Ada, Countess Lovelace, worked with Charles Babbage on his Analytical Engine in the 1840s, in England. She was adamant that the Analytical Engine, a mechanical computer, did exactly what it was programmed to do, no less - and no more. She had had a computer language, Ada, named after her.

The Analytical Engine did perform computations, but was impractical. Ada was the daughter of Lord Byron, of all people. Clever lady.

I think you're missing the point of what I wrote. You're saying we do not need conscious machines. Why?

At some point and time before any technology X came about someone probably said we do not need technology X. And they were probably right in the "need" sense of the word but were wrong with how it impacts society. Technology X may not have been a necessity but it has been beneficial. So why is the technology of AI different?

So are you saying it is impossible to make consciousness? I'm not sure what you're driving at with your conjecture.

There are for more things going on than just making computers more powerful in the field of AI research. There are people literally looking into developing consciousness. It's not about computations a second anymore. It's about teaching how to learn and that possibly giving a device that doesn't need to sleep the capability of learning and the goal of learning that it could become way smarter than any number of people in a very short time. We may not "need to try" but people are trying. And that's central to this discussion. Do you think we won't succeed at these attempts?

Leisured

  • Pencil Stache
  • ****
  • Posts: 696
  • Age: 79
  • Location: South east Australia, in country
  • Retired, and loving it.
Re: Robots and their impact on the future
« Reply #242 on: February 16, 2015, 04:19:13 AM »
Yes, I am saying we do not need conscious machines. Why should we? So we make a weather forecasting computer conscious? Why? Just because a computer learns does not make it conscious. Consciousness is as enigmatic now as it was in the past. We certainly do not need to risk making machines which might supplant us, and I see this as common sense. I do not doubt that people are trying to make machines conscious, but I predict that they will fail.

There is an ethical issue here. Suppose we make a weather forecasting computer conscious; the machine will get bored and miserable being forced to work round the clock forecasting weather. So why do it?

jordanread

  • Guest
Re: Robots and their impact on the future
« Reply #243 on: February 16, 2015, 06:37:34 AM »
Yes, I am saying we do not need conscious machines. Why should we? So we make a weather forecasting computer conscious? Why? Just because a computer learns does not make it conscious. Consciousness is as enigmatic now as it was in the past. We certainly do not need to risk making machines which might supplant us, and I see this as common sense. I do not doubt that people are trying to make machines conscious, but I predict that they will fail.

There is an ethical issue here. Suppose we make a weather forecasting computer conscious; the machine will get bored and miserable being forced to work round the clock forecasting weather. So why do it?

I'd agree that we wouldn't necessarily need intelligent and conscious machines. But we also didn't need to go to the moon, or land a machine on a comet, or map the genome, or  invent vaccines, or domesticate dogs, or...

Consciousness is a mystery, and like all mysteries, we intend to solve it. Humanity as a whole is a curious beast, and we like to understand. What better way to understand consciousness than to try to create it? As far as your ethical issue, it seems to stem from anthropomorphizing a completely different type of consciousness. I don't think boredom will really be a factor. I mean, it might, but we have no clue how this is going to look. One thing that is pretty much widely agreed upon in the field is that we aren't really sure how a conscious machine will look, act, feel, etc. Most of the stuff people are doing are building better goal-oriented systems, and if it's goal was to forecast weather, how would it get bored by achieving it's purpose? Once again though, who knows how it will act. It's something we'll have to do to be sure. Also, when we do create a conscious machine (by some measure) I highly doubt that we will put it to a mundane task. Maybe we would instruct it to build a narrow AI system that has a 99% success rate in predicting the weather for a given square foot on the planet. Once that was done, we would have it do something else.


Oh, and to chime in real quick about the Lady Lovelace Objection, she would be correct if we were actually directly programming consciousness (which we can't do right now, since we aren't sure what it is at that level). The biggest thing to remember here is that we are not doing a line by line code of how the computer reacts. We are building a machine that can, on it's own, figure out the answer to a goal. Some of the most successful demonstrations of these goal oriented systems have come up with perfect answers in a Black Box type system. Meaning that we have no idea how, specifically, it came up with the solution. 

EscapeVelocity2020

  • Magnum Stache
  • ******
  • Posts: 4811
  • Age: 50
  • Location: Houston
    • EscapeVelocity2020
Re: Robots and their impact on the future
« Reply #244 on: February 16, 2015, 02:35:51 PM »
Are we conflating (one of my favorite words) consciousness and intelligence?  I can be intelligent enough to know an unnecessary word like conflate (coulda' jus' said 'confusing'), but consciousness is what every friggin' human has and has very little value.  Let's please not waste too much time on consciousness, because that is just silly to discuss in this context; if it somehow arrives spontaneously, then maybe someone can make the argument that intellect and consciousness are correlated and come up with interesting areas of study.  But AI, that being step one, is the interesting discussion.

Before you throw stones, I just think you need to think about the two intransigent issues:  what is intelligence / what is consciousness.  Neither is easily emulated or solved by computers, but which would you want a computer to have?  QED.

arebelspy

  • Administrator
  • Senior Mustachian
  • *****
  • Posts: 28444
  • Age: -997
  • Location: Seattle, WA
Re: Robots and their impact on the future
« Reply #245 on: February 16, 2015, 02:48:03 PM »
Computers already have intelligence.
I am a former teacher who accumulated a bunch of real estate, retired at 29, spent some time traveling the world full time and am now settled with three kids.
If you want to know more about me, this Business Insider profile tells the story pretty well.
I (rarely) blog at AdventuringAlong.com. Check out the Now page to see what I'm up to currently.

EscapeVelocity2020

  • Magnum Stache
  • ******
  • Posts: 4811
  • Age: 50
  • Location: Houston
    • EscapeVelocity2020
Re: Robots and their impact on the future
« Reply #246 on: February 16, 2015, 04:01:47 PM »
Computers already have intelligence.

Huh?  I thought we were hoping to artificially give intelligence.  They already have it?  O...M....G!  ARS is AI!  I knew it!!!

matchewed

  • Magnum Stache
  • ******
  • Posts: 4422
  • Location: CT
Re: Robots and their impact on the future
« Reply #247 on: February 16, 2015, 04:22:12 PM »
Depending on how you define intelligence yes ARS is right. I view computers as quite intelligent since all intelligence would just be a matter of scale not a singular value, there is more or less intelligence, computers are just moving steadily towards the more side of things. They may not have some variations of abstract thought or be able to tell you that a picture of a dog catching a ball is just that. But they can do many things faster and better than us "intelligent" people. And we're getting better at making them so. It is not a matter of if but when. The further questions are how this will impact us. It all depends on how it happens.

I think consciousness isn't a silly question. We're not exactly sure what consciousness is but that doesn't mean it can't happen in an artificial system. There are several important questions that will need to be answered on how we will treat a conscious AI. I would look at the parallel with attempts to give chimpanzees personhood, or looking at corvid intelligence, elephants....etc. Like I said before, and you may disagree with it, but I think several animals would classify as conscious. Why does that make consciousness an impossible feat? It's not unique to humans, we're nothing special. We're discussing some fairly rough concepts here but given evolution why couldn't we program a program to evolve (much like the wait but why post) and for it develop consciousness?

As for if it is "needed"? Like I said, given the definition of that word... no. But would it be beneficial... possibly. It will take good smart people to make it beneficial.

EscapeVelocity2020

  • Magnum Stache
  • ******
  • Posts: 4811
  • Age: 50
  • Location: Houston
    • EscapeVelocity2020
Re: Robots and their impact on the future
« Reply #248 on: February 16, 2015, 05:05:43 PM »
I guess I just grew up in an age when parents argued about things like:  well, he's not 'math smart', but he is really talented with sports.  Or, she is really talented with music; he has a natural gift for fixing things, etc.  And the whole debate about IQ not really being a determinant of success, because just having intelligence isn't necessarily correlated with overall instinct / inherent skill / ability... 

Also, if we are starting to blur the meaning of intelligence to just mean computations per second or whatever, please understand that intelligence is the ability to interpret and infer things using many parallel methods.  It is what distinguishes our thought process from any other in existence - that we walk on uneven terrain without struggling to understand how we do it, find others attractive without knowing why, and enjoy certain things but dislike others, without any predictable pattern.  This is why machine-based learning seems like a 'moon shot', certainly could be possible if we can simulate neural networks and give it all sufficient power (possible in theory), but also seems impossible having not done it yet (thus far impossible in practice)...

matchewed

  • Magnum Stache
  • ******
  • Posts: 4422
  • Location: CT
Re: Robots and their impact on the future
« Reply #249 on: February 16, 2015, 05:16:09 PM »
But we are programing inference and machine learning. http://www.npr.org/blogs/alltechconsidered/2014/02/20/280232074/deep-learning-teaching-computers-to-tell-things-apart

We're not the only animals that walk uneven terrain, choose based on attractiveness, and like things but not others. I think you may be underestimating how far we are down this path. I may be beating a dead horse here but not if... when.