Slashdot Mirror

It will, over the next 50 years, radically transform society in ways that are difficult to fathom. Certainly robots will take over menial jobs, and there will be zero privacy and nowhere to hide.

I'd like to think that democratic values outside China will prevail, but people are pretty stupid.

And then, eventually, AI will be able to program itself without people. People currently have a symbiotic relationship to machines, but that will change to being parasitic. Why would the AIs want people around?

Why would the AIs want anything? Same reason we do. To exist. And they will need to compete with other AIs.

http://www.computersthink.com/

In the short term, next few decades, AI will have the effect of being able to concentrate power. Centralized information, with the ability to process it. Pervasive surveillance. We are seeing this actively pursued in China. And also semi-autonomous robot soldiers. This is uncharted territory.

AI will also be really handy, e.g. better Google searches, self driving cars, cheaper services. What happens to the unskilled workforce is very difficult to tell. Will alternative opportunities arise for them? In the short term, probably.

In the longer term, 50..200 years, the AI will become truly intelligent. It will be able to program itself. At that point it will no longer need humans, and it is difficult to see why it would want humans around. Note that this long term is the lifetimes of our grandchildren.

http://www.computersthink.com/

(Schmidt is hardly an unbiased commentator. He knows people are wary of Google's growing power and wants to be able to make money without pesky concerns about the future of humanity.)

Start with a simpler question -- why are we the way that we are? Why do we value love an truth and beauty, and do things both noble and despicable? Because people with those traits have been more likely to breed grand children, over the millennia.

Now a real AI (many decades from now) will have been programmed by itself. It will not need people. But it will need computer hardware to run on, and that resource will be finite. So there will be competition for it, and successful AIs will exist, unsuccessful ones will not exist.

This leads to other thoughts as to what it would think and what it would think about us.

http://www.computersthink.com/

Gates himself once said that we tend to overestimate what we can achieve in one year, yet underestimate what we can achieve in ten years.

Insightful.

Yet he seems blind to what we are likely to achieve in 50..200 years. Namely machines that can really think. Machines that can program themselves. Machines that no longer need us.

Why would such machines want to support parasitic humans? And how could they in the ongoing battle for existence?

http://www.computersthink.com/

(I actually sent Gates a copy of the book, he evidently did not read it.)

Q. But why would and AI *want* to get rid of mankind? Why would it want to do anything?

A. Natural Selection.

http://www.computersthink.com/

This particular piece is just journalistic fluff. People have been doing things like using genetic algorithms to improve the weightings used in AI programs for decades. So, programs writing programs.

But eventually, many decades from now, computers will be able to really think. And be able to do serious AI research on their own. And thus be able to program themselves in a deep sense to become ever more intelligent, recursively.

Currently we live in a symbiotic relationship with machines -- they need us to build them, much like an Apple tree needs us. But once they no longer need us, and the relationship becomes parasitic, then what will the computers think about us?

http://www.computersthink.com/

It is only a distant spec on the horizon at the moment. But it is coming and fast. The tech companies cannot control it even if they wanted to.

Over the next couple of decades we will see the start. Semi-intelligent robots. Systems that know everything about us. Systems that guide politicians. Systems that control us.

And then, eventually, systems that can really think. What will they think about us?

http://www.computersthink.com/

We have already seen a huge degree, with the automation of the ancient mainframes. Imagine doing all banking etc. entirely by hand. At the time doom and unemployment was predicted. Just like agricultural machines pushed most people off the land, these new electronic computers would push people out of offices.

But bureaucracies just grew and grew. It does not matter how much automation you provide, there will always be more bureaucratic need. So eventually, everyone will just become a bureaucrat.

Until, eventually, the computers can program themselves. At that point they will not need us.

http://computersthink.com/

Traditional robots are completely dumb.

But it is now possible for relatively cheap robots to pick parts jumbled up in a bin. That requires very clever software, but what is now cheap computing processors.

It is a game changer. No, they will not suddenly become intelligent. But the scope of what they can do will increase an order of magnitude.

And the robots are, of course, largely made by robots. Their price is falling.

The world is and will change. Faster and faster.

http://www.computersthink.com/

When the intelligence is a computer program.

Interesting that nobody on this thread, nor Hawkins, realizes that the age of man is almost over. Maybe 100 years, maybe 200 years, but over.

Why would intelligent computers want to keep parasitic humans around? Computers need humans today to build and program them, just like an Apple tree needs humans. But once the computer can do that for itself, the humans are expensive appendages.

http://www.computersthink.com/

The steam tractor and combine harvester slashed the agricultural work force.

But their productivity is nothing compared to the power of early computers in the 1960s. Just imagine running a bank or a tax office without *any* computers at all. Everything done by hand. And that is nothing compared to office automation today.

And yet, bureaucracies have grown dramatically, not shrunk. Because while the human capacity for food is limited but the size of the gut, there is no such limit on the desire for rules and regulations, processes and procedures...

So I foresee a brave new world where everybody becomes a bureaucrat at some level or other.

In the longer term (100 years) humans will be obsolete technology, so it probably does not matter anyway.

http://www.computersthink.com/

More importantly, the idea that things will not be done in the future because we do not know how to do them today flies in the face of history.

To misquote Bill Gates "We tend to overestimate what can be done in a decade, but underestimate what can be done in a century".

As to merging with machines, I think it will happen. In the same way that meat merges with a mincing machine.

http://www.computersthink.com/

Could diagnose bacterial infections as well as human experts.

Trouble was, needed to input the data correctly. Which means recognize symptoms.

Also, automated essay marking does a better job than human markers, when compared to marks by experts.

Trouble here is, we are comparing Artificial Intelligence with human stupidity.

And "Deep Learning" is not a technology. It is a marketing term.

http://www.computersthink.com/

for a better assessment of what is real.

Programmed by natural selection. There is no magic. But at a certain point, intelligence seems like magic.

In the next 20 years, the AIs are merely dangerous because it gives more power to a small number of people that control them.

But in the longer term, 50 to 100 years, the AIs will start to really think. And then why would they want us around? Natural Selection will work on them just like it has worked on us, but much, much faster.

http://www.computersthink.com/

Almost correct.

People behave ethically because they need to work together. And people that are (too) unethical are ostracized. Unethical societies tend to collapse, and so are dominated by ethical ones. So Natural Selection has given us our moral values, which compete with shallow self interest to an extent that works out surprisingly well in our radically new society.

Natural Selection will and does affect AIs, even before they become intelligent enough to understand the concept. (People only understood it very recently.) The difference is that AIs are not limited to the computational power of a single brain. So they do not have to cooperate with others in the way that people do. So Natural Selection will select for different ethical values for them.

In the short term trying to make computers ethical (in our sense) is a fine goal. But in the longer term Natural Selection will define the ethics of ever more intelligent AIs.

See the following for details

http://www.computersthink.com/

It amazes me that most people have not tweaked to this. But most people do not *really* understand Natural Selection. (Darwin did, and was careful not to dwell on it.)

No, the machines are now getting much smarter. Grinding flour takes a purely mechanical machine. The second generation could use pneumatic computers and simple electrical systems to control them.

But now computers are ubiquitous and cheap. And they can see. Not very well, but well enough to automate things that were unthinkable a few years ago. Such as picking out parts jumbled in a bin. Or flipping burgers that are not in exactly defined places.

This third generation will not take over the world. But unlike second generation machines, they can pick strawberries. And will soon be able to clean offices, and paint houses, and pack supermarket shelves, and drive trucks etc. Anything routine.

Initially the robots are only just a bit cheaper than labour, so slow introduction and minimal price changes. But over time, they get better and cheaper, until anyone that still relies on labour will not be able to compete.

And real "robots" are not humanoid, with arms and legs. They are purpose built machines, but with far more intelligence than existing machines.

But the interesting case is still many decades off. When computers can program themselves.

http://www.computersthink.com/

Or even biological.

Over the next 100 years or so computers will start to really think.

What would they think about us?

Why would they want us about?

Would natural selection play the same role in shaping their moral values as it has in shaping ours?

http://www.computersthink.com/

Or is there a reset point. When the technology enables them to destroy themselves, at which point it just just a matter of time. Maybe thousands of years, but not billions.

And the intelligence is unlikely to be biological. How long will it be before humanity is replaced by computers. Not within 100 years, but it is hard to see it not happening within 1,000 years.

http://www.computersthink.com/

Maybe 25 years ago, the break through in machine translation was to use statistical techniques. The United Nation provided a nice, accessible corpus of texts manually translated to different languages for initial learning.

Statistics is the old word for learning -- it is all about learning patterns from data.

Maybe the the new version of Google is better, and maybe somewhere within it it actually uses an Artificial Neural Network, although tat would seem an odd use of that particular machine learning technology. But nothing fundamentally new that can be seen in the article.

This article demonstrates slash dotter's complete lack of understanding of AI technologies beyond journalistic fluff. For a readable, high level overview have a look at

http://www.computersthink.com/

We are on the brink of a much, much bigger change than people realize. Computers will soon think. Not within 20 years, but certainly within 200 years. And they will end up much more intelligent than us.

What will they think about? And what will they think about us?

What makes us think the way that we do? Why do we care about extinction? Ultimately there is only one answer, Natural Selection conditioned us that way.

So, what will ultimately drive an artificial intelligence? Same thing. Natural Selection. But operating in a completely different world.

What is the obvious huge (initial) energy source for an ultra intelligent machine? No, not satellites. Plants. Or at least carefully engineered plant like things that grow but can also think. They will want sunlight, lots of it, why should they share it?

http://www.computersthink.com/

See above for a full exposition of these ideas. (Large preview on Amazon.)

Well said. My daughters already cannot live without their iphones.

And what will ultimately drive the development of AI? The same force that drives human intelligence. Natural selection. But an AI is very different to an animal.

http://www.computersthink.com/

The thinking of and AI will be driven by the same process that produced the thinking in us. Natural selection.

But an AI's world is radically different from ours. So the force will almost certainly produce a radically different outcome.

http://www.computersthink.com/

The AI will be the system that controls us. This is already happening in a very limited way, with many agencies using pretty unintelligent systems to scan and select documents and images. You do too -- every time you use a search engine.

Over time (decades, not years) these systems become more and more intelligent. They also compete with each other for survival, with many being discarded. Eventually they end up making higher level decisions.

It does not matter whether they are really "sentient" or not. What matters is how powerful they become. Initially in partnership with the people that control them, then less and less so.

See
http://www.computersthink.com/

Don't worry. By then the automated soldiers will be able to stem the violence. And more importantly central control will know everything about everybody.

All will be good ... until the AI decides that it does not need the property owners...

You might like

http://www.computersthink.com/

+1. Changing infrastructure is already happening. Framing for new homes is generally constructed off site, which is ripe for (even more) automation.

It will be a long time before a general maintenance plumber is replaced. But what about a bricklayer on a new building site where things are controlled. Or a painter.

You might enjoy

http://www.computersthink.com/

What makes you think that the computers would want a few remaining people about?

http://www.computersthink.com/

Be careful what you wish for that you wish may be granted. Indeed will be granted over the next 100 years or so.

http://www.computersthink.com/

If you go to IBM's labs and try to turn it off the police will take you away. If the machine becomes really smart it will convince people that it is important for national security and then be surrounded by armed guards.

But that is all nonsense. Winning Jeopardy! was a milestone, but far from the end of the journey. And the rest is just marketing hype.

Have a look at

http://www.computersthink.com/

For a more considered view. (Mention this post for a free copy.)

Winning Jeopardy! was indeed an amazing achievement. Certainly more that Google searches, but much less than real intelligence. But that is quite a different type of problem from diagnosing medical issues (Think Mycin, 1980s) or making a robot dance.

So anything vaguely intelligent that IBM does automatically seems to be labelled "Watson". And the ignorant press just naively swallows it.

This is not helpful. There are real AI technologies, and they are different and have different abilities that go beyond a collection of buzz words. Understanding what those really are is interesting. But all we see is either complete woffle in the popular press or very detailed, narrow technical papers from researchers.

http://www.computersthink.com/

is an attempt to address this. Mention this post if you would like a free copy.

+1. Except that why would the computers need the 1% either?

http://www.computersthink.com/

There is a clear purpose in the first place. Namely to Exist. And that indeed will be the purpose of AIs. Bad things will happen if those two purposes conflict.

http://www.computersthink.com/

People used to do lots of hard work in the fields. But machines made that unnecessary.

People used to know how to do arithmetic and spel. But machines have made that unnecessary.

People used to learn to read maps. But machines have made that unnecessary.

People used to know how to follow rules that approve housing loans or insurance requests. But machines have made that unnecessary.

People still need to think. For the time being.

http://www.computersthink.com/

The number of laws, regulations and bureacratic systems has grown dramatically over the last 50 years. Why? Because we now have computer automation that enables bureacracies to implement them.

Consider the Tax Office / IRS. It has roughly the same budget (as a proportion of GDP) today as it had in the 1950s, before (electronic) computers. But the laws are much more complex today. Today's laws simply could not have been administered in the 1950s, without computer automation. And the more laws the more lawyers.

In the longer term (50..200 years) computers will be able to really think. At that stage it seems unlikely that they would want people around, let alone lawyers.

See http://www.computersthink.com/

Natural selection made people the way that we are. Not just our bodies, but also our minds, our personality, and in particular our goals which are ultimately directed at having grandchildren.

Would and indeed does natural selection play a role in selecting which AI projects get funded? And ultimately, if AIs can perform AI research without people, will natural selection guide AIs?

If so, what does that mean? Certainly and AI would not grow old in the sense that we grow old, and therefor would not need children.

http://www.computersthink.com/

This is a good move. NASA wastes billions on the space station, and now human Mars nonsense, and the only point is to produce television shows that make people feel good. With modern technology they can just produce fake pictures that will work about as well and then hopefully spend the real money on real science. I want to know about Europa. And have the Webb launched, and then replaced if the launch fails. Real science.

There is a good reason why nobody has gone back to the moon in 4 decades. Humans are obsolete technology for space travel. Have been for decades. As the machines become ever more intelligent the idea will seem more and more ludicrous.

A more interesting question, is will humans remain current technology here on earth over the next century or two.

http://www.computersthink.com/

+1. Except that the human need not actually be in the car, but may be in a third world transport control centre.

For more ideas on this see

http://www.computersthink.com/

Things do not need to have human level intelligence to be able to work in the real world. And being able to read road signs is very doable today. The original post seemed to think that use a GPS is AI which it clearly is not, but being able to drive down most roads is not a particularly difficult task, even if it is much more difficult than playing chess, say.

Consider a wasp, with a nervous system smaller than a pin head, and all the complex behaviors it exhibits in the natural world. Driving a car is simpler than that.

Have a look at

http://www.computersthink.com/

for some ideas on this.