AI and the rise of China: this is the New New World

Economics, Politics, Philosophy, History and Culture
Forum rules
Always add something of value to the discussion and have fun. Mind your language, no porn, no libel, no trolling and no personal attacks.

Please note, views expressed on the forum do not necessarily represent the views of Mises UK. the Mises UK Council, Mises UK Faculty or Mises UK members. Membership of this forum is open to anyone world wide who is interested in civil discussion.
User avatar
Clayton
Posts: 97
Joined: Sat Mar 31, 2018 4:24 pm

AI and the rise of China: this is the New New World

Post by Clayton » Fri May 04, 2018 1:58 am



This video is pure gold. Lots of fluff at the beginning of the video, so skip to 5:00 to start the meat of the talk.

The advent of AI is widely misunderstood - the masses misunderstand it, the market misunderstands it, the media misunderstands it and governments misunderstand it. AI is to the human brain what the internal combustion engine was (is) to human/animal muscles. AI doesn't mean you never need to think again, it just means you need to engage in much less laborious/tedious thought. Deep thinking becomes something more like going to the gym. You read Euclid's Elements or play a game of chess to make your brain stay in shape but you don't need to engage in that kind of heavy lifting in order to earn a living.

The danger of AI lies almost completely in how misunderstood it is. AI is not magic. AI will not automatically solve our problems. AI will be ultimately dependent upon humans for the foreseeable future, not the other way around. All the same, AI is far more powerful in real applications than most non-experts really understand. Here are a few 2-minute paper videos on AI applications that have surprisingly broad implications for the future:







As you can see, we have several key ingredients converging at this time in history:

- Cheap computational parallelism (e.g. desktop GPUs being wired into mesh supercomputers)
- Massive, distributed, real-world data streams
- Push-button (more or less) Deep Learning architectures that can convert mass, raw data into actionable information, often in real-time
- A virtuous circle where neural networks are facilitating faster development of themselves, as well as their own base layers (i.e. the underlying computational substrate)

Integrated circuits that have a combination of analog circuitry and digital circuitry (e.g. digital-signal processors) are called "mixed-mode designs". What I am asserting is that we are entering a new, post-technological phase of history where technology and un-technology are going to become increasingly intertwined - we are becoming a mixed-mode society. In the end, technology is only valuable because of what it enables us to really do. The latest developments in AI are taking technology development out of the hands of a cloistered few ("engineers") and placing it into the hands of an increasingly broader audience. In short, those who do not value technology for its own sake are going to find technology increasingly unavoidable and they are going to find fewer reasons to avoid technology as it becomes ever more seamless - even ruralites will start wearing "smart clothing" before long. By the same token, technological aptitude will eventually become less valuable as a standalone skill. What really matters is not what a robot can be made to do on the testbench but, rather, how the robot can be made to serve human ends.

Humanoid robots are coming sooner than you probably think. The old "pistons, gears and levers" approach to robotic design is already obsolete. Rather, robots are going to be designed by Deep Learning methods, allowing the computers to churn through countless possibilities in order to settle on the optimal dynamics. We still don't have an AI smart enough to make the robot solve useful tasks? No matter, there are many disabled people (or even able-bodied people) who would be thrilled to earn a living by donning a VR helmet and piloting a robot around or remoting in when the robot has reached its destination and is ready to perform some task.

Buckle your seatbelts, folks, Stage I of this rocket has just detached and Stage II is about to ignite...

User avatar
Physiocrat
Site Admin
Posts: 260
Joined: Mon Mar 12, 2018 11:25 am

Re: AI and the rise of China: this is the New New World

Post by Physiocrat » Fri May 04, 2018 7:32 am

This is fascinating. I very much agree with your vision of how AI will interact with humans. I think in a way it will be more of a cyborg rather than an AI takeover future. Do you have any links to articles or books which predict future changes when advanced digital technology further blends with ordinary human life?
The atoms tell the atoms so, for I never was or will but atoms forevermore be.

Yours sincerely,

Physiocrat

User avatar
FvS
Posts: 121
Joined: Mon Apr 02, 2018 3:03 am

Re: AI and the rise of China: this is the New New World

Post by FvS » Sat May 05, 2018 1:53 am

He takes her hands. We see a childlike excitement in his face. He
wants so badly to share the almost orgasmic thrill of discovery,
the satisfaction of creation.

DYSON
Baby, this thing is going to blow 'em all away.
It's a neural-net process --

TARISSA
I know. You told me. It's a neural-net
processor. It thinks and learns like we do.
It's superconducting at room temperature.
Other computer are pocket calculators by
comparison.
(she pulls away from him)
But why is that so goddamn important, Miles?
I really need to know, 'cause I feel like I'm
going crazy here, sometimes.

DYSON
I'm sorry, honey, it's just that I'm thiiis
close.

He holds up his thumb and index finger... a fraction of an inch apart.
She picks up the prototype. It doesn't look like much.

DYSON
Imagine a jetline with a pilot that never makes
a mistake, never gets tired, never shows up to
work with a hangover.
(he taps the prototype)
Meet the pilot.

Image
"Most whites do not have a racial identity, but they would do well to understand what race means for others. They should also ponder the consequences of being the only group for whom such an identity is forbidden and who are permitted no aspirations as a group." - Jared Taylor

User avatar
Physiocrat
Site Admin
Posts: 260
Joined: Mon Mar 12, 2018 11:25 am

Re: AI and the rise of China: this is the New New World

Post by Physiocrat » Sun May 06, 2018 3:52 pm

Clayton wrote:
Fri May 04, 2018 1:58 am

The advent of AI is widely misunderstood - the masses misunderstand it, the market misunderstands it, the media misunderstands it and governments misunderstand it. AI is to the human brain what the internal combustion engine was (is) to human/animal muscles. AI doesn't mean you never need to think again, it just means you need to engage in much less laborious/tedious thought. Deep thinking becomes something more like going to the gym. You read Euclid's Elements or play a game of chess to make your brain stay in shape but you don't need to engage in that kind of heavy lifting in order to earn a living.
Clayton, do you have links or articles which have a similar view of AI meshed with human future?
The atoms tell the atoms so, for I never was or will but atoms forevermore be.

Yours sincerely,

Physiocrat

User avatar
Clayton
Posts: 97
Joined: Sat Mar 31, 2018 4:24 pm

Re: AI and the rise of China: this is the New New World

Post by Clayton » Mon May 07, 2018 3:15 am

Physiocrat wrote:
Sun May 06, 2018 3:52 pm
Clayton, do you have links or articles which have a similar view of AI meshed with human future?
https://www.theguardian.com/technology/ ... k-life-3-0

https://youtu.be/Gi8LUnhP5yU?t=1h10m0s




User avatar
Physiocrat
Site Admin
Posts: 260
Joined: Mon Mar 12, 2018 11:25 am

Re: AI and the rise of China: this is the New New World

Post by Physiocrat » Mon May 07, 2018 6:16 am

Excellent. Thanks Clayton
The atoms tell the atoms so, for I never was or will but atoms forevermore be.

Yours sincerely,

Physiocrat

User avatar
Jon Irenicus
Posts: 149
Joined: Sun Mar 25, 2018 9:36 pm

Re: AI and the rise of China: this is the New New World

Post by Jon Irenicus » Mon May 07, 2018 11:02 am

I'm not sure Kurzweil's divide between mammals and non-mammals in this context makes much sense. Having done a quick search on the topic, this came up:

https://en.wikipedia.org/wiki/Triune_br ... _the_model


Pretty interesting talk nonetheless.
Former overlord of the original Mises forum.

User avatar
FvS
Posts: 121
Joined: Mon Apr 02, 2018 3:03 am

Re: AI and the rise of China: this is the New New World

Post by FvS » Tue May 08, 2018 5:22 am

Have I just seen too many movies to think AI is a good idea?

!Warning! - Violence and a topless robot.



"Most whites do not have a racial identity, but they would do well to understand what race means for others. They should also ponder the consequences of being the only group for whom such an identity is forbidden and who are permitted no aspirations as a group." - Jared Taylor

User avatar
Jon Irenicus
Posts: 149
Joined: Sun Mar 25, 2018 9:36 pm

Re: AI and the rise of China: this is the New New World

Post by Jon Irenicus » Tue May 08, 2018 5:20 pm

On the topic of AI:

https://www.telegraph.co.uk/science/201 ... ing-robots

Arachnophobes, beware. You've been forewarned. :P
Former overlord of the original Mises forum.

User avatar
Clayton
Posts: 97
Joined: Sat Mar 31, 2018 4:24 pm

Re: AI and the rise of China: this is the New New World

Post by Clayton » Wed May 09, 2018 2:48 am



Goertzel's approach is based on meshing the AI (tech) and logistics (market) into a single, distributed platform that allows countless, independent actors to cooperate to create an AI network. This approach - not the monolithic "Big Data crunching" of Google, Facebook, et. al. - is where I predict we will see the medium-to-long-term power of AI. You can only make speech synthesis so good (and then humans can't perceive the improvements any further), you can only make image-recognition so good, and so on. All these truly exciting AI tasks that the Big Data crunchers are taking on are still just interface-level tasks, they will never give rise to general-purpose reasoning. The long-run future lies in the holistic approaches to AI, whether the theoretical approach associated with people like Juergen Schmidhuber and Marcus Hutter, or the craftsman approach of Ben Goertzel.

To better understand SingularityNET, look at the material economy as a metaphor. Suppose you start a business selling pancakes. Are you going to grind your own flour? Then you need to acquire a flour-grinder and raw wheat. Unless you plan to build your own flour-grinder and grow your own wheat, in which case, you will need to acquire tooling and fabrication, farm equipment and hire farm hands and buy land, and so on. Obviously, each business chooses to specialize and focus down on producing a good from a sufficiently narrow slice of the production-structure as to be manageable. No one can operate a business that spans the entire structure of production from raw materials all the way to consumer goods.

Goertzel envisions the emergence of a global, general-purpose super-intelligence to be analogous to the material economy but where the goods and services produced along each stage are intangible - services like data-processing. Obviously, this model doesn't fit with the monolith approach that corporations employ. But the more I think about it, the more sense it makes to me. Suppose I produce digital art prints suitable for display in poster art galleries. A lot of the tools used by digital art these days is benefiting from AI algorithms. But the goal of art is to create, so I don't necessarily want to use the same AI filters that everybody else is using. I might have some completely novel idea but lack the idea of how to implement it. I could hire a team of software engineers to build my tool for this one art project but that might not be profitable. What I really want is a tool that I can somehow describe the kind of filter I want and the tool itself just "figures it out". So, you need some kind of AI translator that can digest natural language into something like formal requirements. Then you need an AI dispatcher that can send out these requirements to AI modules that can satisfy them (it has to know who to call). And so on and so forth. But who's going to pay for all of that to happen? Well, the idea is that you have a kind of "intelligence economy" of AI modules that can be activated in exchange for some kind of fee. You send the module some cryptocurrency and the description of what you want it to do (in some common language) and it performs its task for you. So, the digital artist might drop $10k into his AI arts software tool which will then use this cash to drive the rest of the global AI brain to create the desired art filter. Just as with human software development, the artist might have to go through several stages of alpha-testing and beta-testing before he gets the desired AI software module that performs the transform he wants. We can think of each "AI module" on the network as a kind of "factory" that takes orders and produces goods, but where this factory is more like a math function that takes information inputs and transforms them into information outputs.

At this point, this is still very blue-sky stuff. Nobody even knows how to do all of this. But I think the key insight, here, is that some group of people has to get the snowball rolling down the hill in order to get it to build up.

Post Reply