ROBOTS.

EXAPTED SAYS: The way in which technology is currently combined with organizations to increase productivity does not usually involve direct replacement of humans with equivalent task-processing by machines. The economy gets more and more deeply integrated with technology, including artificial intelligence. Current AI agents are generally less flexibly intelligent than very stupid humans, although for some tasks (often unique tasks that humans never did anyway) they perform much much better. Companies try to combine machines with humans in the most productive way possible. They do try to cut labor costs, as they try to cut all kinds of costs; but machines are expensive and not really very smart. The humans with the least intelligence and skill get sidelined in this process, but usually new types of tasks are exapted and new skillsets are required all the time. The humans with the least amount of raw talent which is (virtuously) combined with education can often lose their spots, but the number of required spots for very intelligent and skilled humans often increases at the same time. It is unclear whether human IQ will be able to increase fast enough for this process to continue for much longer, since the economy and technology progress exponentially and it could be too fast or too great for humans.

I imagine two “sweeps”:

(1) The technology-integration sweep:
As technology evolves, specific human intelligence tasks are one by one more productively performed by narrow AI technology in human-machine organizations, at a much faster pace than ever before. The pace of change is so great that we are not sure if human IQ levels can keep up. We could see a situation where humans are still slightly more generally intelligent than machines, making them indispensable in the economy, but machines are just much cheaper. Whether most humans would be employed or not depends on a lot of different factors. (That would be one sort of limit or culmination of the technology-integration sweep.)

(2) The general-intelligence sweep:
Machines become generally more intelligent than humans. Eventually costs for AGI machines will come down to the point that it is counter-productive for an organization/system to integrate with a human.

I don’t think we’ve gotten to (2) yet. I think we’re at (1). It seems a bit premature to be convinced that the labor pool will be shrinking (compared to population or job-seekers) within, say, the next 10-20 years. I think if the labor pool does shrink, it could be due to other (non-“technology”) factors (corruption, hegemony, poor management, etc.).

 

HGRAYSTON: I think we need to differentiate between a world where robots and computers do all the work, and a world where everything is free. Those are two completely different things, and the degree to which these are routinely used in the same sentence demonstrates how little careful thinking is going on with regards to these subjects.

A completely automated world does not mean a completely free world. From an economic perspective, a completely automated world hypothetically could come about if the economics of automation were simply more favorable than the economics of employing humans.

There are many examples of heavily automated or even completely automated factories and business processes today. None are free. And the projection that smarter computers will make everything free is specious, to say the least. The exact opposite is far more likely, that smarter computers will make things at least as expensive as they are today, because smarter computers are clearly tough to achieve and represent very sophisticated (and still deeply theoretical) technology.

Therefore, how do you respond in the scenario where all humans are unemployed, but money is still very much around, and things still cost about what they do today? Is it still a utopian fantasyland, or is it a dystopian displacement scenario? It seems hard to imagine it still being utopia, when every time the utopian case is made for complete human displacement it rides on the coattails of the completely unrelated “everything is free” thing.

ME:

If you need money in order to obtain anything, and you need a job in order to obtain money, but all the jobs are performed by machines, this is clearly a system that is broken. The best thing to do would be to sever the link between jobs and wages. Pay everyone a basic wage, no exceptions. During a period of time when only some job types are taken over by automation, the basic wage might be be enough to get by on, but not enough to afford luxuries. This might provide an insentive to go out and get what work is available, but if you find it difficult to get employment (which will be increasingly the case as the tide of automation rises) at least you will not be in dire straits.

Once the job market is completely taken over by robotics, Moravec suggests taxing the hell out of the robotic industries, and using the funds to finance a lifelong retirement for humans.

JABELAR: Anyway, the problem I see is that humans are hampered by some serious evolutional baggage related to “rare” resources that are becoming abundant. The obesity problem is an example — our technology has allowed us to have access to ice cream and potato chips all day long, while our evolution still considers those things rare and so gives many people an overwhelming urge to stock up body fat.

So to truly progress, we will somehow need to overcome our strong urges (which become addictions) related to things that used to be rare:
– food abundance is already causing obvious problems.
– sex abundance is becoming a problem (already we can see more naked women in an hour than most guys would see in a lifetime 30 years ago), and will become a true abundance when robotic sex companions get good enough.
– material abundance is becoming a problem (just watch shows like “Hoarders”, just look around the average suburban house, or look at shopping addiction).
– power abundance is becoming a problem (in virtual worlds you can be god-like and so people are getting addicted and basically forgoing “real” life in order to spend time in the virtual environment).

If we don’t overcome the above, I am pretty sure almost 100% of us will end up addicted to immersive virtual environments that satisfy the urges.

I do believe that there are other “higher” urges (probably also evolutionary baggage) like seeking understanding, spiritual connectedness, etc. But those too can potentially be fullfilled in virtual environments.

Maybe there’s nothing wrong with humanity integrating into urge-fulfilling virtual worlds. But somehow it seems a bit of a shame to end up all connected to orgasmatrons or whatever to fulfill some unnecessary evolutionary imprints.

PREDICTIONBOY (commenting on my statement that robot diversity will be greater than that of biological life): I would say that the degree of intelligence incorporated into a product would actual reduce – perhaps immensely – the need to have a distinct offering based on shape or size in the first place.

An excellent example that we can imagine is in the case of robots targeted at the consumer. In this case we must restrain our raw imagination and include basic human predilections, as well as economic considerations.

For example, the Roomba is a special purpose robot. It does one thing – vacuums your floor. If we look at this example and simply extrapolate to the future, we could imagine special purpose robots that do other limited tasks. One that clears the table. Another that washes the dishes. Still another that dusts the furniture. Etc.

In the near term, there may be just such a proliferation. It’s hard to tell, because after a decade the Roomba is still the only widespread example we have. Except that there is a robot offering that mows the lawn, that has had some limited success, but far less than the roomba.

Now, at this point let’s corral our raw imagination, and introduce human preferences. In this case, it could be readily suggested there will most definitely be a limit to which most consumers are going to want a small army of special-purpose robots milling around the house, cluttering things up. In our age of consumerism there’s already a huge problem with clutter, and these will add significantly to this.

So there’s that to consider. Now let’s again constrain our raw imagination with simple economics. For you, clutter is not a problem, you have plenty of room for all these droids, so you purchase them. Each of these have their own manufacturer, maintenance issues, etc, etc. Each of them will have an upgrade cycle, new models being introduced, which is yet another recurring cost.

In short, at some point this proliferation of all these special purpose droids becomes a physical nuisance and an economic burden. Defeats the purpose of having them.

What’s the alternative? Forget the droids, just do the work yourself? Yes, that is one alternative, but there is at least one more.

That is, combining all of these various droids into one droid, that can perform any or all of these tasks at any time. What would enable this to be possible? The low-hanging apple falls from the tree: intelligence. Intelligence is what could allow one platform to learn and perform many different tasks, some of which it has never done before.

Therefore, greater intelligence actually REDUCES the need profoundly to have a plethora of different shapes and sizes in the first place. Which is the exact opposite of Extropia’s point.

Now for the next question: if you were to purchase just one droid, what would its form factor likely be? Do I need to say it? It should be obvious what the choice of most humans would be.

Advertisements
This entry was posted in Secondary Thoughts.. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s