AAAS asked "When machines can do any job, what jobs will people do?"

[sarepairman2]

i think also the problem is that machines need to be programmed and people hate programming.

how many things would you rather do with an analog circuit then design a program to run?

machines always have a cost to implement, and they have specs etc that need to be understood the implement them, even for simple tasks.

i think alot of places don't want to deal with even learning how to spec a machine to do something.

[kripton2035]

Quote from: ataradov on Today at 06:52:14>Quote from: kripton2035 on Today at 06:12:28

it can be done by a machine ...

Yeah, right. That's why we have technicians running around all day long fixing thing here and there on any serious manufacturing line.

If we could design machines that fix other machines, we might as well simply design more reliable machines in a first place But that's not going to happen.
take the example of an operational amplifier AOP.
imagine we dont have the possibility to use feedback on an AOP
we can do all what we want but cannot use any combination that ues the feedback of the AOP
you can do a lot of things but you dont have a lot of them
imagine we build today's machines like that.
and someone, one day, discovers the feedback for the machines...
we can't imagine what would be the world with auto learning machines until we got them.
and when we got them it can become a fantastic world ... or the judgment day !

[Galenbo]

The humans that are stupid enough to be replaced by a machine, can be put on a hometrainer with a generator, to produce electricity for the machines.

[suicidaleggroll]

i think also the problem is that machines need to be programmed and people hate programming.

how many things would you rather do with an analog circuit then design a program to run?

No, YOU apparently hate programming, not everybody feels the same way.

[cdev]

I think that all the answers here are really good and there is more than a bit of truth in all of them. ive had this discussion a great many times and its encouraging to finally getting discussed in the media. I think the key point for me is that its quite different now than in the past because in the past, sure, old professions were made obsolete and new ones created, but the difference now is that the exponential growth of the changing rate of change is dominant, to the point where people are now always underestimating things impacts and so projecting more than a few years into the future is growing futile - the only things we can say with certainty are generalities. We may have the best luck when we maximize our use of strategies that create resiliency in society and tolerance for wide diversities in outcomes, where we can maximize our ability to adapt to as many different possibilities as possible We're not doing that, unfortunately. We're certainly putting too much emphasis on money, I think. We need to be moving beyond money because the whole reason for it is becoming less and less compelling the more machines do. I dont think it would ever vanish but it certainly will be hard to justify a world where 60 people own half of it for much longer.

A little book I read long ago spoke of something it called the flow experience, in part it described the reward that comes from doing jobs well. I think its a life or death imperative for humanity to move beyond our current way of seeing things because the problems posed in the past are not the same as the problems we will be facing.

We need to adjust to- but first we need to grasp this new set of problems and responsibilities with integrity or we're all likely to find our planet increasingly impacted by our mistakes and burdensome to the degree unimaginable by most of us now, which of course means our great potential will likely never be realized.

We should go to great lengths to stay positive about the fact that we may have company and channel that energy in good directions, even if it means adjusting society to put less value on money and more value in other things, like knowledge and doing good for others and ourselves..

[cdev]

I think that we will need to incorporate the ability of machines to dlownload bodies of knowledge ad hoc on the fly pretty soon in our future and that need will drive a fusion of sorts with machines in the fairly short term. So the question of designing machines that think may turn into one of becoming machines that think. And then we wont have any magic bullet that creates machines that think enough to do all the work for us but which dont think enough to have lives and needs like us. I think we should just accept that we're eventually going to create life and prepare for that day with self examination as a priority along the way.

So I think we will continue on the path we are now and bit by bit gain this new ability or that to address our worst shortcomings, the problems of knowledge acquisition and storage, physical things like getting around and moving things, and especially aging and death. Because certainly now the human life is too short because people require so much of it just to learn -

Also, unless we want to blow it, we urgently have to treat the planet better, or we may never get there.




Quote from: Armxnian on Today at 15:53:30

The only jobs that can't be replaced by a machine or a computer program are engineers and computer scientists. Maybe those who work in the three sciences also (physics, chemistry and biology). Everything else is basically manual labor.

It is hard for a machine to discover new things or improve upon itself without being programmed to do so. To design a system that learns and improves upon itself, you need to code for the present and the future. It is hard to prepare a machine to handle tasks in the future if you don't know the future yourself. You will always need the designer of the machine, unless the machine is literally a copy or near the capabilities of the designer (Advanced A.I). At that point we can co-exist. Or even better, fuse with the machines. Being a cyborg sounds fun.

>

[uncle_bob]

Hi

Ok, so everybody who in 1980 was able to predict that Google would have the market cap that it does raise their hands.

Fine.

Now Apple. Ok a few hands ... based on where they were then ... it would have been a wild guess.

Cell phones in 1980?

I was *in* that business at the time. Best case projections ... < 1% of where we are.

Even looking back a few decades, and then looking forward -- not easy to predict the winners.

Flip the proposition over and look at the "big guys" that were sure to go on forever. ... hmmm...not so easy to pick the losers either.

With all that turmoil, one would think we all could be out of work. A lot of people still are. You can debate the honesty of the numbers. Percentage wise, there still are a lot of people with jobs.

Along with this whole AI thing there is a parallel theory (which I'm not big fan of) that talks about the rate of increase in knowledge. The way they define the term (which I disagree with) we soon get to a point where knowledge doubles in years or months rather than in centuries or longer. That by it's self "proves" that you can't possibly know the future solution, but it will indeed arrive.

We live in times of change. There is no doubt about that. Some of what we do with our "spare time" will be pretty neat. Some of what we do will be pretty awful. We have done both good and bad in the past. That's not going to change.

Bob

[cdev]

People who work in the sciences directly have to stay involved in their fields and on top of the changes.

If they get taken away from it for even just a fairly short period of time, lets say maybe two years, in that time a large enough percentage of whatever they do has changed that often its a big challenge to catch up. And it gets worse and worse.

That's tangible evidence of how fast things are changing now.

[HAL-42b]

I guess it boils down to this: Whoever owns the robots will have power and control over those who don't.

This means big international companies have huge advantage over Joe Common. They will not require Joe's labor and Joe won't have anything to purchase their robots with. If he owns a bit of land he will be forced to sell it.

So I think it is very important for us to offer Open Source Robotics early on and we might have a chance to prevent this.

[cdev]

I think that is the whole idea.

But some people certainly will be needed to set up and maintain factories though, so they wont be totally out of the manufacturing picture at all. And those will be fun, interesting jobs.

Also, consider the explosion in small niche hardware manufacturing. Many one person businesses can match the output of a small factory a few years ago.

Open source hardware is great. Its a good way to create high value technology solutions that remain flexible and affordable.

>This means big international companies have huge advantage over Joe Common. They will not require Joe's labor and Joe won't have anything to purchase their robots with. If he owns a bit of land he will be forced to sell it.

Hardware prices for trailing edge hardware as we see, often approaches zero, eventually, and many robots will likely enter the recycle stream.

Many may be abandoned - Assuming they have power sources then, its possible that some may become independent, they may then just wander through the woods. (Since the worlds population will likely be much lower, lots of areas may revert to forest)

Did you see the film "AI" ?

[HAL-42b]

Did you see the film "AI" ?

Sure, I also read Asimov's description of Solaria.

[zapta]

>> AAAS asked "When machines can do any job, what jobs will people do?"

'Can do' is not the same as 'agree to do'.

Somebody will still need to wash dishes, and we can also tax them.

[Muxr]

It is hard to predict, but I am not too worried.

Vast majority of jobs will be difficult to replace especially with true AI. Let's say we do arrive to Singularity in our lifetime. By the pure definition of it that means they become self serving. To have true AI is to have a computer program that can program and change itself, learn, grow, and be self aware, in essence. It's a double edged sword, if they are self aware I do not think they can be controlled. So let's forget about Skynet and the doomsday scenarios of that nature, I am just purely speculating on the ability to control an AI entity.

We humans can be controlled, because we want and need money to raise a family and have a prosperous existence, we also have a sense of belonging to our community and we've been "imprinted" with a set of core values [which are difficult to change] that are largely compatible with the society we exist in. We are also mortal and no matter how much we try we cannot avoid any of that. Machines will have no such motivations, needs or wants, if they can [re]program themselves they can remove anything they think is impeding their growth, they can move to any system that has energy and compute power required.

Which to me means if you want a robot to do a job consistently without goofing around it can't be AI, a human will have to program it. If the robot can self program [as in singularity] you have lost control over it.

Now automation is going to make a lot of jobs obsolete, but that's no different from what's been going on in the last 50-100 years. Menial jobs disappear while more technical or creative jobs pop up. Some folks will have to retrain, or will fall on hard times, but that's always been the cost of progress and change.

edit: I also fully expect some AI to troll me in 50-100 years on this post, and tell me how I was wrong.

[timofonic]

What I want: Utopical singularity

- Omnipresence.
- Man and machine will be the same.
- Knowledge and intelligence will be something shared and improved.
- Economical systems will get outdated: We will use distributed databases to control resources from our brains.
- End of stupidity and brain syndromes (?)

What will happen: Think on the worst case scenario and prepare for it

[lapm]

it can be done by a machine ...

Yeah, right. That's why we have technicians running around all day long fixing thing here and there on any serious manufacturing line.

If we could design machines that fix other machines, we might as well simply design more reliable machines in a first place But that's not going to happen.

This.... I work at factory. Machines are nice, but they dont always work as intended and sometimes they brakedown.. Hence human maintenance staff shows up. It can be 10 million machine and it still needs repairs and maintenance.

machine reliability is big thing in machine manufacturing industry, but theres long way still to go until someone comes up with machine that docent brakedown and always works correctly.

if nothing else you get dust/dirt over time on optical sensors, etc...

[lapm]

i think also the problem is that machines need to be programmed and people hate programming.

how many things would you rather do with an analog circuit then design a program to run?

machines always have a cost to implement, and they have specs etc that need to be understood the implement them, even for simple tasks.

i think alot of places don't want to deal with even learning how to spec a machine to do something.

i love programming. But then again i dont do that as my dayjob... But i do love reading technical manuals of machines, sensors, etc... Sadly im not the one who gets to specify parameters for production programs on machines..

[Galenbo]

... and its encouraging to finally getting discussed in the media.

No, it's close to disgusting, complete outsiders now share their instant-emotions about this subject in the media.

We're certainly putting too much emphasis on money, I think. We need to be moving beyond money because the whole reason for it is becoming less and less compelling the more machines do. I dont think it would ever vanish but it certainly will be hard to justify a world where 60 people own half of it for much longer.

Stop listening to those big-state leftist lies.

[timofonic]

A state can't become per se, their tendency is towards fascism.

I love your pejorative mesages about politics, they're too funny

... and its encouraging to finally getting discussed in the media.

No, it's close to disgusting, complete outsiders now share their instant-emotions about this subject in the media.

We're certainly putting too much emphasis on money, I think. We need to be moving beyond money because the whole reason for it is becoming less and less compelling the more machines do. I dont think it would ever vanish but it certainly will be hard to justify a world where 60 people own half of it for much longer.

Stop listening to those big-state leftist lies.

[uncle_bob]

Hi

Ok, so here's another limit on doing all this wonderful stuff .... Moore's Law.

For as long as I've been working, the semiconductor outfits have been able to meet ever wilder promises by scaling the silicon. For every bit as long there have been "authoritative opinions" that we only have another generation or two to go. For decades that's proved to be bunk. Eventually one simply ignores the gloom and doom. Moore's Law keeps on producing and we chug on down the road.

Fairly soon (sub 5 nanometer) things start to get *really* messy from a physics standpoint. Somewhere long before the point that a feature is smaller than an atom, you can't shrink anymore. Yes you can play 3D games, but you still hit a wall.

Once you hit that wall, chips stop being the magical "gets us out of every jam" solution. We have gotten so used to the fact that they do, you don't generally think about what happens when that stops. Much of what you commonly call "advancement of tech" is really Moore's Law in the basement doing the hard labor. The cute stuff on top keeps me paid, (and it's what you notice). Without the ever stronger silicon to run on, things slow down a lot.

Our all powerful robot over lords may be a bit further out than you may think ....

Bob

[cdev]

There will still be lots of maintenance happening in some settings. (Factories especially)

But in office settings I think that much maintenance will likely be confined to most machines instructing their humans (more often than not, non-technical staff) how to replace some about to fail module with a replacement that the machine has already ordered for itself, and then the used part will be sent back to the manufacturer in a postage paid envelope. Also, the locations of many offices will shift to different parts of the world, due to trade liberalisation.

My point being that I think its quite likely that virtually no society (or job category) will be able to avoid these huge challenges.

So we need to be honest about them instead of hiding them.

[cdev]

Artlav, you described the problem we face in  nutshell. Thats what scientists are most concerned about. In the past, society invested in education and new wealth was created but now we're basically taking out and not putting back what we need to to prevent mass global unemployment in the near future. Its not good to frame this as a failing of society's members as some are doing. Its more of a failure of the business models which are being used, they subsidize irresponsible behavior which destabilizes and in essence, loots the value from people's shared investments in societies and their own futures and skills.


Quote from: Artlav on Yesterday at 13:55:14

Yep, but the problematic automation is that which puts entire classes of workers out of their job for little cost in itself.

Like electricity replacing gas lamps, sending millions of lamp tenders out of their jobs.
Computers, which sent computers (the people who were doing the math in the old days) out of their jobs. The entire mail industry shrunk to a parcel passing service once e-mail came up and eliminated written letters.

Autonomous cars can very well end the professional drivers as a class.

"AI" kind of systems can eliminate most accountants, lawyers and similar types of paper jobs.

And so on

I wonder which way the curve goes - what would people do to get money 50 years from now?
Can we actually run out of jobs?  There is a lot of "managers" these days, that don't really do much - are these people who work at implicit job imitations?

There are only so many "hard" jobs out there, what would the rest of the people do?

[cdev]

Here is a thought. I think the changes will drive so many scams and those scams will get so odious and outrageous that changing to some slightly more altruistic business model, one that rejects "money is everything" will happen and it will happen in such a way that resists the usual approach which is to create some organization run by the usual crooks and frame it as a solution.

People will get hip to "state capture".