Smart robots will take over a third of jobs by 2025, Gartner says

What you're really arguing here is that "thinking" is something fundamentally special, that it is not an emergent property of the underlying physical properties or states of the material system. This is dualism. I was assuming a monist position where building a smarter-than-human machine would ultimately be an engineering problem.

The dualism vs. monism debate goes back thousands of years.

 

I don't see how dualism can be taken seriously as an idea at all.

There's an entire thread discussing this in terms of free will and determinism in General Discussion.

 

No I didn't.
Thinking is not special.
I said that the biological system, brain AND body, is way more adaptive to changing situations than a programmed (whatever way) robot computer and mechanical parts.
Why everytime changing what I said?
Stop bicycling around what was said.

 

Sorry Pirocco, I must have misunderstood you.

I pointed out that we have been able to engineer machines that exceed the limitations of human biology in other areas: "Is it impossible for humans to build a machine that can travel faster than humans? Is it impossible for humans to build a machine that is stronger than humans?"

So, if thinking is just an emergent property of biology then why should it be impossible to build a machine that is smarter than humans? I say it is therefore possible and that we can (even though it is a tremendously difficult).

If the brain is just a (biological) machine that provides a substrate for the process of thinking, then why can't we engineer a non-biological substrate that is equivalent or better?

Now we're back to where you were originally with the argument that "a brain... is way more adaptive to changing situations than a programmed (whatever way) robot computer and mechanical parts."

Your key objection seems to be based on the assumption that knowledge and responses embodied in a computer program are set in stone at the time of creation through a set of rules. But this assumption is incorrect.

As mentioned earlier, machine learning systems are built without such constraints. They are systems for learning new rules, new knowledge, and new responses rather than following explicitly programmed rules.

We already have systems that are faster and more adaptive to changing situations in extremely narrow domains. Over time we can build systems that are more generally adaptive and more generally intelligent.

 

The Wright Brothers built an aircraft before fully understanding the science of aeronautics. But the point is not that we can't build it now, but that it should be theoretically possible sometime in the future given everything we currently know.

 

Its interesting to speculate on that point and where it will take us. Technology can be viewed as an intelligence amplifier that multiplies our potential.

"Machines don't just replace what we do, they change the nature of what we do our tools change us. By extending our capabilities they set new expectations for what's possible"

Broad changes have already been noted thanks to being able to "outsource" storage and retrieval of knowledge.

 

The advantage that humans have in "designing" a "better" brain compared to nature is that we can experiment far more easily. Just like evolution, you only need to hit on a design/system that works once. In my very general opinion, at heart it's probably more about flexible learning routines that can interact with it's own hardware in new, unpredictable ways. Both of these elements already exist independently and I think it's just a matter of the hardware getting better and better.

I think it was Richard Dawkins who discussed a processor that Intel (or someone) built that had flexible gates that could be modified into thousands of possible hardware configurations. They overlaid evolutionary-style software on top where a task was given based on a starting configuration. The software then created random "mutations" of the starting configuration, selected a couple that moved it closer to achieving the task and randomly "mutated" those and kept repeating for a few hundred or few thousand iterations until the task had been achieved. From memory, of the final configurations of the processor that solved the task, a couple were already known. One was noticeably more efficient than the researchers had ever devised, but once they saw the solution they knew how it worked. And then there was a special one that worked but (at the time) they had no idea why as <in my crappy layman's terms> a pathway that should exist from the input signal to the output simply didn't exist. In the book it was postulated that there was some form of electro-magnetic connection across different parts of the chip that enabled it to function. Whatever the exact details, it was an example of a method that provided not just a better solution by experimentation and learning, but also provided a solution that would never have been guessed as it was <seemingly> dependent on a physical property intrinsic to the hardware itself in the presence of the input that solved the task that probably wouldn't work with a slightly different processor (or input).

If anyone else knows this story, please correct/elaborate

 

I can't wait to get my cyborg chips implanted. What a step change that would be

 

Robots will never take over human labour entirely.

It's counter productive. In fact my guess is that robots will be reduced in years to come.

You can increase productivity all you like with robots, you can cut cost by reducing the human work force to a minimum.

It's the human with a job that needs to purchase products.

 

Even when robots do, say most of a medical operation there will still be human supervisors. Robots will have to be all but human in order to do everything that humans are capable of. Entertainment is a good example. Humans will provide far better entertainment products than any robot for quite some time. And just look at the way that the entertainment industry has expanded in the last 50 years and how many jobs have been created as a result. That's because a lot of menial tasks have been automated allowing for people to take on higher level work for better pay and have more money to spend on such products thereby creating entire industries.

 

This topic makes me think about the space age predictions of the future.
It's now nearly 50 years later and look at the 'sissy part of the predictions' that actually happened, and what remains of even that sissy part.
That's what happens if you focus and ignore the rest / the whole of reality.

 

Put clothes on a robot and it looks creepy:

[youtube]http://www.youtube.com/watch?v=tFrjrgBV8K0[/youtube]

And then there's this f*(ker that throws concrete besser blocks:

[youtube]http://www.youtube.com/watch?v=2jvLalY6ubc[/youtube]

 

Forget about that 1776 Tea-Party.

 

I would question why a free society would need this application of technology ? And what about those mandated ethics subjects at university?

 

Law enforcement and keeping the community safe is the most obvious application:

 

(btw: I hope there's no one who is unfamiliar with The Onion)

 

I have the movie

 

Kissing robot available for sale now:



At least, I think its a kissing robot.