Jump to content
House Price Crash Forum
Guest

Endless Bear Market: Housing Drops To Nothing

Recommended Posts

I think it's pretty obvious that machines will take over more and more work from humans. In the short term that's not such a problem as it will still be possible to get rich by coming up with ideas that machines can't... but as the processing power of computers continues to grow, eventually it will be well past that of humans.

Sooner or later, humans will be obsolete. I can only wonder what the machines will decide to do with us then...

Share this post


Link to post
Share on other sites
I think it's pretty obvious that machines will take over more and more work from humans. In the short term that's not such a problem as it will still be possible to get rich by coming up with ideas that machines can't... but as the processing power of computers continues to grow, eventually it will be well past that of humans.

Sooner or later, humans will be obsolete. I can only wonder what the machines will decide to do with us then...

read "Godel, Bach & Escher: The Eternal Golden Braid" by Douglas R. Hofstadter for a fascinating thesis on the nature of intelligence & being. Complex self-referential loops - something that no machine is yet capable of replicating

thought it was worth a heavyweight reference for my 100th post!

Clearly I haven't finished the book yet, otherwise I'd be able to enter into a much more convincing ramble about why we shouldn't be too worried about this (yet)

Share this post


Link to post
Share on other sites
it is a real goal being pursued tirelessly by thousands of researchers globally, with billions of backing when the technology is ready.

Don't you think all that money and effort would be better spent on researching ways to produce clean and affordable energy (such as fusion) for the world's burgeoning population (to avoid resource wars) and eliminating starvation rather than trying to build a marginally faster computer?

Share this post


Link to post
Share on other sites
Don't you think all that money and effort would be better spent on researching ways to produce clean and affordable energy (such as fusion) for the world's burgeoning population (to avoid resource wars) and eliminating starvation rather than trying to build a marginally faster computer?

that's probably the point - we've realised that that problem is too much for our feeble human minds to overcome, so we're hoping it will be more straightforward to "invent" some artificial brains that can do it for us. Labour saving devices that mean not only are we no longer required to "do" we are no longer required to "think"!

Perhaps this is the true meaning of "ignorance is bliss"

AI - the quest for bliss?

Share this post


Link to post
Share on other sites
The poll about 20 year house price predictions got me thinking...

I know a lot about computers, robotics and artificial intelligence, and I would say that over 20 years, a large proportion of human jobs will be replaced by automation.  For an entertaining but detailed explanation, here's a pretty good fictionalised run down:

http://www.singinst.org/

I have a BSc and a PhD in artificial intelligence and I've been working in industrial AI research for 4 years. Unfortunately, or perhaps fortunately depending on your point of view, AI is nowhere near replacing humans as a viable industrial basis in most businesses. I might also add that timescales for AI predictions have massively underestimated the scale of the challenge time and time again through AI's history.

The thing is, AI has been very successful at replicating the things that humans can do which animals can't. E.g playing chess, filtering spam emails or translating business documents between German and English.

But it turns out that a lot of what we do is just like animals: fine motor control, social reasoning, nuances of communication. A lot of progress has been made on this stuff, but the problems just seems to get bigger and bigger the more you look at them.

Humans, on the other hand are cheap, pretty much look after themselves and are apparently infinitely adaptable. Even when you have a nice clever robot to assemble widgets, it turns out you need to also hire a really expensive clever human to program and maintain it as well! Since AI research is basically market driven, you can see that there isn't going to be a fast change here.

frugalista

Share this post


Link to post
Share on other sites
As a side effect, the housing humanity needs will be less and less valuable - eventually the resources will be better spent elsewhere.

This kind of thing has always interested me. Here's what I think on the subject:

Let's assume that humanity's long term aim is to spend all of our time having fun and that no one will have to do any work (and that there will still be plenty of food etc to go around). If this is going to happen, I think that we need deflation (so that everything will be free in the asymptote). This is not what happens. Here's why:

When the owner of a car plant (with 100 people working there) buys a machine that can do the work of twenty people, he should think "great now we can all do 20% less work. this would lead (in the long term) to the whole staff not having to do any work and the same productivity would be achieved. In reality the owner thinks "great, I can make 20 people redundant and make more profit for me". He then ends up paying for those people anyway (through taxation) but the whole process is much less efficient. so everyone ends up worse off.

I know that this is a very simplistic example but I think it illustrates a very important weakness in the capital system: because it's essentially adversarial, we spend a lot of our time competing where, if we worked together, we could all be better off.

Here endeth the lesson!

Share this post


Link to post
Share on other sites
He then ends up paying for those people anyway (through taxation)

Not necessarily, as he can always move his car plant to Slovenia or somewhere and pay much less tax (and cheaper wages).

a very important weakness in the capital system: because it's essentially adversarial, we spend a lot of our time competing

Actually competition is the only reason he sells his cars for £9000 rather than £900,000 (and also becuase you come up against what consumers can afford). If he had a monopoly (ie no competition) I can assure you car prices would be much higher.

Share this post


Link to post
Share on other sites
In reality the owner thinks "great, I can make 20 people redundant and make more profit for me"

You're assuming, of course, that factories would continue to exist, and that AI and nanotech and similar future technologies that are at least reasonably feasible wouldn't allow people to build their own cars at home from scrap iron and dirt.

Power is increasingly decentralising. In many areas, the old left-wing goal of 'workers owning the means of production' has already become trivial, and I expect that to continue: for example, in my spare time I work on low-budget movies, and in the last decade we've gone from shooting on film that costs maybe fifty pounds a minute with the help of huge film labs and expensive post-production companies to shooting HDTV on tapes that cost two pounds an hour and doing all the post-production (editing, sound-mixing, etc) on a PC or Mac in a bedroom.

I do agree on the timescales, BTW: I don't see human-level AI before 2050, if that. But given how little intelligence is required for a lot of jobs, maybe 2020 isn't too far off for replacing them.

Edited by MarkG

Share this post


Link to post
Share on other sites
This kind of thing has always interested me. Here's what I think on the subject:

Let's assume that humanity's long term aim is to spend all of our time having fun and that no one will have to do any work (and that there will still be plenty of food etc to go around). If this is going to happen, I think that we need deflation (so that everything will be free in the asymptote). This is not what happens. Here's why:

When the owner of a car plant (with 100 people working there) buys a machine that can do the work of twenty people, he should think "great now we can all do 20% less work. this would lead (in the long term) to the whole staff not having to do any work and the same productivity would be achieved. In reality the owner thinks "great, I can make 20 people redundant and make more profit for me". He then ends up paying for those people anyway (through taxation) but the whole process is much less efficient. so everyone ends up worse off.

I know that this is a very simplistic example but I think it illustrates a very important weakness in the capital system: because it's essentially adversarial, we spend a lot of our time competing where, if we worked together, we could all be better off.

Here endeth the lesson!

This really is not the way that free market economics (aka capitalism) works. The fact that despite being invented 260 years ago by my compatriot Adam Smith it is still misunderstood in this way is a great mystery and source of problems for the world.

What actually happens is that the staff now freed from the drudgery of car building are able to do something else useful thereby increasing everyone's net wealth (or, looked at another way, decreasing the amount needed by an individual person to generate a unit of wealth). The economy has become more productive.

In what way is buying and selling freely undertaken (whether shares, cars, labour or whatever) adversarial? In fact, the only alternative to the free market, which lets the market participants themselves determine what is bought and sold and for how much, is the unfree market - where someone else does the determination. That really is adversarial.

edit confusion about Adam Smith's age

Edited by BoredTrainBuilder

Share this post


Link to post
Share on other sites
In what way is buying and selling freely undertaken (whether shares, cars, labour or whatever) adversarial?

Because a lot of people seem to spend a lot of time and effort trying to get you to buy their version of a product rather than on improving the product. (ie advertising and gimicks).

I agree that the free market does not have to be adversarial but I think that it usually ends up that way because of human nature (the same reason that communism fails).

Share this post


Link to post
Share on other sites
...

Whatever happens, anything that is significantly more intelligent than we are, that has been designed by humanity, will hopefully be intelligent enough to grasp our aspirations and fears and either help us to match itself, or else set up a cosy utopian "retirement" deal with sw^nky surroundings, eternal youth, endless entertainment, and lots of sex!

...

I,Robot - the machines will take us over in order to save us from destroying ourselves.

Share this post


Link to post
Share on other sites

Hasn't the present silicon-based processor technology just about reached it's limit (in terms of speed & memory density)? I seem to remember a recent Horizon programme examining the quest for carbon-based technology (organic?).

I must say, from my, fairly limited, appreciation of moden digital electronics, the organization of the cpu & associated memory etc. seems to simulate brain structure, albeit in a very crude way - I think you can certainly make analogies, if nothing else.

I believe the real difference is sheer complexity; perhaps the brain consists of countless interlinked, faily simple processing machines which can network in a myriad ways (apologies, I'm knackered & this is coming straight off the top of my head (brain)). I remember reading an interesting book by Dennet which concluded that consciousness was most likely distributed throughouit the brain.

So perhaps there are ensembles of ensembles of ensembles of networked processors making up one brain.

Hang on, these (brains) are now themselves networked via the internet!

Something that occurred to me a while ago: are we building the next stage in our evolution??

Meanwhile, back at HPC ...

Share this post


Link to post
Share on other sites
All the trends point to 2020-30 for human level processing power.  Once it's doable, at price, it will go from nowhere to everywhere.  Look at the spread of outsourcing as another example.

Absolutely not. This is just utterly conceptually flawed. What is missing here is the fact that the difference between what human minds can do, and what computers can do, is not a matter of scale. There is a fundamental conceptual difference between the two that we as yet have no adequate grip on. It doesn't matter how many supercomputers you link up, computers have no intentionality - they have no purpose. Humans set out to do things, computers don't. Give a child a pen and a piece of paper and watch it go - turn on your computer and watch it sit there with the cursor blinking away.

I have a BSc and a PhD in artificial intelligence and I've been working in industrial AI research for 4 years. Unfortunately, or perhaps fortunately depending on your point of view, AI is nowhere near replacing humans as a viable industrial basis in most businesses. I might also add that timescales for AI predictions have massively underestimated the scale of the challenge time and time again through AI's history.

Three useless degrees in Philosophy - BA, MA and PhD - and yes, absolutely. People talk blithely about AI and enhanced intelligence and so on when what they mean is enhanced ability to solve defined problems by rule-governed/algorithmic processes. Human "intelligence" is of a radically different order than this and putting aside the sci-fi "predictions" the truth is we've practically no idea how to conceptualise what human minds can do, let alone replicate it.

Share this post


Link to post
Share on other sites
Absolutely not. This is just utterly conceptually flawed. What is missing here is the fact that the difference between what human minds can do, and what computers can do, is not a matter of scale.

I was just about to despair of the Waco element on the board.

Finally someone is talking sense.

Nice one Munro.

Share this post


Link to post
Share on other sites

I’m from the electro-mechanical age. When I started work we were seriously promised that automation would be doing the repetitive tasks and the human race would have an abundance of leisure on it’s hands. It was predicted that by the year 2000 the average working week would be 10hours. The rest would be lounging in the sun. The millennium is now nearly five years old, and where’s that leisure time, and how many hours are we expected to work? 48

Share this post


Link to post
Share on other sites

Still waiting for the promise from some 15/20 years ago that technology would lead to unlimited leisure time.

With an income to utilise it.

The future?

Never turns out as predicted.

If you want AI to behave as Human "intelligence" expect anything.

Just as with Humans.

Share this post


Link to post
Share on other sites

Three useless degrees in Philosophy - BA, MA and PhD - and yes, absolutely. People talk blithely about AI and enhanced intelligence and so on when what they mean is enhanced ability to solve defined problems by rule-governed/algorithmic processes. Human "intelligence" is of a radically different order than this and putting aside the sci-fi "predictions" the truth is we've practically no idea how to conceptualise what human minds can do, let alone replicate it

It has moved on a little bit from that, it is possible to set up modelling techniques which quite effectively match what small parts of the brain accomplish. For exacmple if say you wanted a system to tell the difference between bird calls you could do it - you wouldn't have to set out an algorithm to model all the calls you just need to set up a self-learning system that has sufficient complexity and a means of getting the original data into the system from which it can then learn by example. This is what a new-born does from the outset - except the complexity is hugely greater and has been tweaked over millions of years - enough processing power and the right configuration to make sense out of a whole load of jumbled up signals and a means of making some very complex outputs - movement, speech, reasoning etc and after a few years you have a sentient human being. Bottom line is wall that whirring round in our heads is little more than electrical impulses firing off from one node to another, just on a massive scale.

Impossible to tell how long it would take to get to anything like human level inteligence (or even motor and observational skills) but the thing is once you have cracked it, it is now almost immediately infinitely reproducible - the world could potenitally change in a few years.

Share this post


Link to post
Share on other sites
Absolutely not. This is just utterly conceptually flawed. What is missing here is the fact that the difference between what human minds can do, and what computers can do, is not a matter of scale. There is a fundamental conceptual difference between the two that we as yet have no adequate grip on. It doesn't matter how many supercomputers you link up, computers have no intentionality - they have no purpose. Humans set out to do things, computers don't. Give a child a pen and a piece of paper and watch it go - turn on your computer and watch it sit there with the cursor blinking away.

Three useless degrees in Philosophy - BA, MA and PhD - and yes, absolutely. People talk blithely about AI and enhanced intelligence and so on when what they mean is enhanced ability to solve defined problems by rule-governed/algorithmic processes. Human "intelligence" is of a radically different order than this and putting aside the sci-fi "predictions" the truth is we've practically no idea how to conceptualise what human minds can do, let alone replicate it.

Missed this as I was thinking and posting.

I agree.

Emotions, they're the really tricky aspects of the Human mind.

What makes you love or hate someone, or enjoy some activity and not another?

And the Eureka moment of inspiration?

Share this post


Link to post
Share on other sites

I remember reading in a science journal some months back .......

- the Human Brain has (according to them!) a processing power of 2-3 Teraflops ......

- the predicted increase in processing power in computers (they reckon!) should reach the 2-3 Tf region within the next 20 years .......

Now, some "professor types" argue that what we call "human consciousness" (out of which all our emotions, innovations, aspirations, even sense of humour derives from) is nothing more than sheer processing power !!

Therefore AI reaching and surpassing human beings becomes a certainty ..... with only a matter of time as the obstacle ......

However, this is assuming these "professor types" are correct ..... cos other "professsor types" theorise that the workings of the human mind are vastly more complex and beyond the scope of any inorganic artificial creation ......

Therefore technology will continue advancing but despite all the hype, will never the less, change little from the status-quo ....... ie, everybody still working for a living ...... just in differant industries and with differant skills with differant methods etc etc ......

The big question is ..... what exactly is "human consciousness" ...... and the big answer is ...... nobody really knows ..... :(

Until this is figured out, the future is ...... errrr ....... anybodies guess ......

Personally though, I'm hoping for the first scenario as it seems more interesting and exciting :D (even though it could (possibly!) mean the demise of the human race - oh well, it was nice while it lasted :) )

Edited by Perfectionist

Share this post


Link to post
Share on other sites

I dont have a degree, but did watch a lot of Red Dwarf so know a fair bit about nanotechnology and AI. Not sure when this will all happen (humans replaced by bots, etc) , but my best bet would be around the time of the fourteenth series.

Edited by BoringMike

Share this post


Link to post
Share on other sites
The poll about 20 year house price predictions got me thinking...

I know a lot about computers, robotics and artificial intelligence, and I would say that over 20 years, a large proportion of human jobs will be replaced by automation. 

No they won't. As others have said, we can get computers and robots to do things that are programmable. Anything that requires actual creative analysis and thought computers cannot do *at all* at the moment. The nearest thing to creative thought I've seen is Hofstadter's work, but that is ludicrously simple stuff.

People look at computers playing chess and think "wow, it's really thinking". It may be in some sense thinking, but certainly not in the way a human does when playing chess.

I believe computers will one day model the brain, but we are a long long long way away from that at the moment.

Share this post


Link to post
Share on other sites

I agree with Munro/Frugalista...I will point out at this point that I was involved for two years with the European Space Agency looking at 'biomimetic' AI and how it could be applied to space exploration - Subsumption, Basis behaviours, evolved connectionist models and the like. Whilst I am not an AI expert (space engineering is my background) I did get a feel for the area, and did some of my own research into using a basis behaviour system to allow self-assembly in space (you can take a look at this on):

http://www.esa.int/gsp/ACT/biomimetics/tes...esearch_SAA.htm

I feel I have to wade in when ill-informed 'futurists' who read too much Moravec start beating on about machines replacing us in 20 year time. The idea is nice that they could, but as Frugalista said, the 'intelligence' that machines display is of a wholly different type to our own. Big Blue beat Kasparov not because it was smart, but because it used a quite simple searching algorithm (iterative deepening alpha-beta search) combined with A LOT of processing power - i.e. it simply took a brute force approach made possible by it's ability to evaluate millions of positions a second, compared to Kasparov's 2 or 3! Wow, it must be smart everyone says...no, it was just fast. In fact people like Rodney Brooks (father of subsumption) have convincingly argues that the big leaps in processor speed over the lat decades have perverted AI research and allowed the illusion of progress. I agree with this.

Traditional AI is very good in restricted domains where the problem is well defined, the world is largely deterministic etc., and expert systems can display very good levels of 'intelligence' within their specific domain (just don't ask them about anything else), but as the Frug says, it is HOPELESS when it comes to replicating the kind of things that humans and animals do as a matter of course: walking, visual processing and recognition, visio-motor stuff, anything like that...

AI has notoriously 'underdelivered' when compared to all the hype, a lot of it peddled by enthusiasts who do not appreciate how bloody difficult it is. Don't get me wrong, there are some very impressive applications of AI (google etc. are wonderful tools), but I just don't see intelligent robots or anything like that anytime soon. But of course I could be wrong...IMHO, the best bet for obtaining intelligence is to get it as a emergent quality rather than trying to explicitly program it...this will require going back to the first principles used by nature (evolution, neural networks, co-evolved control and morphology) and trying to tap into these sorts of processes that have given rise to complexity and intelligence in us. Just don't expect anything major anytime soon :blink:

Share this post


Link to post
Share on other sites
other "professsor types" theorise that the workings of the human mind are vastly more complex and beyond the scope of any inorganic artificial creation ......

But humans would say that, wouldn't they?

Remember, it's only about 30 years ago that people were saying that a computer could never win a chess game against a human, and not so long ago that people were saying that a computer would never win a chess game against a grand-master. Now both have been done, people just shrug it off and say 'well, chess doesn't really involve _thinking_, does it?'

Quite frankly, IMHO there are very few people who understand enough about the human brain to make any kind of authoritative comment on whether a computer will ever be able to achieve the same. And most of them seem to believe it's just a matter of time.

'Professor types' can argue all they want that humans are special, but while they spend their time arguing, AI gets better and better.

Share this post


Link to post
Share on other sites
Anything that requires actual creative analysis and thought computers cannot do *at all* at the moment.

Define 'creative analysis' in such a way that a human can do it and a computer can't.

Let me give you an example. Years ago when I was at school, we used to play a sci-fi game called 'Traveller'. One spin-off from that was a game called 'Trillion-Credit Squadron': basically each player got a trillion credits (the unit of money in the game), and used the game rules to design a fleet of spaceships and fight against each other.

Someone got the idea to program a computer to play the game. They programmed in the rules and left it playing against itself for a long time, trying different strategies and combinations of ships until it had the best fleet and strategies it could come up with.

That year he entered it in the international tournament that the game company ran. The human players laughed at the bizarre fleets that it chose to use.

And it beat every single one of them.

The next year he entered an improved version.

And it beat every single human player again.

The program retired undefeated, because the people running the tournament said that if he entered the next year, they'd shut the tournament down.

But, you know, beating hundreds of human players in a game isn't that impressive, even though, if a human had done that well, we'd say 'Wow! What a great thinker!'. Here's what is:

According to an article I read around the time, the computer came up with a tactic that wasn't explicitly programmed into it and which no human player, to my knowledge, had ever used before.

In certain circumstances, _IT DESTROYED ITS OWN SHIPS_

You see, if you read the rules carefully and thought it through, you'd realise that in some situations, if one of your ships was badly damaged, you were better off if you destroyed it yourself than if you tried to keep it alive. Humans playing the game thought 'well, that's a starship, there are people on board and it's got guns and stuff, I have to keep them alive'. The computer thought 'I win more games if I destroy ships that are badly damaged'. So it did.

Now, personally, if I'd played against a human and they'd used that tactic and kicked my ass, I'd think they were pretty damn creative. Why isn't the software that came up with the tactic just as creative? What's special about goop-brains that silicon brains can't also do?

Share this post


Link to post
Share on other sites
But humans would say that, wouldn't they?

Remember, it's only about 30 years ago that people were saying that a computer could never win a chess game against a human, and not so long ago that people were saying that a computer would never win a chess game against a grand-master. Now both have been done, people just shrug it off and say 'well, chess doesn't really involve _thinking_, does it?'

Quite frankly, IMHO there are very few people who understand enough about the human brain to make any kind of authoritative comment on whether a computer will ever be able to achieve the same. And most of them seem to believe it's just a matter of time.

'Professor types' can argue all they want that humans are special, but while they spend their time arguing, AI gets better and better.

Yes MarkG, AI is getting better and better, but only in very restricted areas. Take Chess for example...it is a COMPLETELY discrete, COMPLETELY deterministic area of activity: AI can handle areas like this fairly well because there is no ambiguity, no randomness, no partial observability etc. Yes, machines are very bloody good at Chess now (but not at GO it would seem). But ask any of those machines to perform a task outside a chess board....ask it to pass you a glass of orange juice. What Happens?

F all, that's what.

Indeed, noone really understands the brain, and faced with this object of unbelievable complexity, many AI researchers turned towards operating at another level of abstraction/analogy: behavioural AI - i.e. extracting principles from behaviourism and behavioural psychology to engineer AI systems...with some success as well.

As I said, I think that emergence is the way to obtaining real intelligence: the brain is an emergent system....somehow, a clump of several billion interconnections (within the context of a body able to interact fruitfully with the world, which is a crucial component of intelligence) is able to write music, paint, solve mathematical problems, laugh, cry etc. I think that probably to achieve human-like intelligence, we would have to recreate the brain in some form, or at least the complexity and nuance of the brain. Don't forget that nature is parsimonious....all the features of natural systems like the brain are there for a reason...because if they weren't, nature would not bother with them.

By 'Professor Types', I assume you mean people who know what they are talking about? Don't criticise them...it is the Professor types who are improving AI...jeez, you make 'Professor Types' sound like Luddites. They are simply knowledgeable people who APPRECIATE the complexity and magnitude of the problem...looking at the pretty pictures in FOCUS magazine and then deciding that true AI is just around the corner isn't going to make it so.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

  • 302 Brexit, House prices and Summer 2020

    1. 1. Including the effects Brexit, where do you think average UK house prices will be relative to now in June 2020?


      • down 5% +
      • down 2.5%
      • Even
      • up 2.5%
      • up 5%



×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.