1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Chapter One: The story so far.

Discussion in 'Eventually Everything will become free.' started by amusicsite, 9 Apr 2018.

  1. amusicsite

    amusicsite dn ʎɐʍ sᴉɥ┴ Staff Member

    Location:
    UK
    It’s easy to think of automation replacing jobs as a modern problem but it’s far from it. To some extent we would not have the life we have now without using tools and we have been using tools for a very long time. It all started way back when we learnt to use stone tools, spears and arrows to improve our efficiency with hunting, preparing food and making clothes. Along with the agricultural revolution we went from being animals who spent most of our time staying alive, to a society that communally looked after each other. One of the big advantages of this was it gave us the ability to support certain people in our communities that did not have to do anything and could spend their time thinking. Ultimately this lead to people who would look at ways to improve how we did things. Replacing stone tools with metal, improving how they are made or working on creating new tools that make tasks more productive.

    There is a current trend at the moment where people are talking about the ‘exponential growth of technology’ and this also is nothing new. If you went back 10,000 years you would probably have to wait 5,000 years for things to get twice as productive as they were then. Then 2,500 years for it to get twice as good again. This seems to be a rough trend throughout history and if you follow it through you get 1,250, 625, 312, 156, 78, 39, 20, 10, 5, 2.5 years… Then 18 months, 9, 4.5 months then before long you have productivity doubling every month or even every day. This trend seems to hold fairly true over time. Look at the world 100 years ago and the technology of the day would probably be good for 50-80 years. 50 years ago the technology of the day was good for 15-20 years. While 10 year old technology could be out of date in less than 5 years. These days some people think a 2 year old phone is old technology.

    Not all things change at the same rate or at the same time. Something like a chair did not change much for a long time, then new tools change how they are made from time to time. Precision tools lead to more elaborate wooden chairs or new materials like plastics allowed chairs to be made cheaper and lighter. So the wooden stool you may had got 10,000 years ago, you can probably still buy something similar today and was dominant for thousands of years. Then you had a few thousand years of simple wooden chairs, a few hundred years of elaborate wooden chairs. Last century you had a host of improvements. Chairs made from wood, metal, plastic even cardboard all manufactured at an incredibly fast rate. By the end of the 20th Century we had the choice of hundreds if not thousands of different chairs we could buy. Over the next decade we could see 3D printed chairs that use less resources and are stronger or made of new materials we have not developed yet.

    There are a few reasons why the advancements get quicker over time. The first is the increase in global population. When there were only a few million humans there may only have been a few dozen people in the world who had the time, resources and brain power to radical change technology. When you have a population of a few billion there there are millions of people working in R&D or coming up with new ideas. So the more the population increases, the greater the chance of discovering new breakthrough technologies. Another thing is the tools we have. You could not make a computer chip with stone tools, in fact some of the technologies you need to make a modern manufacturing plant for the current computer chips has only been around for about 50 years or less because you need a computer to design and build them.

    The steam engine enabled us to power more machinery, electrical generation lead to the electronics revolution and computers allowing us to enter the information age. Today it’s technologies like 3D printing, deep learning A.I. and autonomous robots that are taking us to the next level.

    There are people now saying robots and A.I. are going to replace some jobs. I’d go even further and say they are going to replace every job. Just like the other obsession of mine which has been getting us off fossil fuels and moving to electric vehicles. I think it’s inevitable and only a question of how long it takes. People are planning for a world where all vehicles run off electricity and they are probably going to mainly be self driving. Full automation replacing all jobs is a very similar thing but at the stage electric cars were 10 years ago. People now are saying it’s not going to happen, new technologies creates new jobs, people need work, along with a host of other reasons that are given as to why full automation will not work in practice. They are people looking at the past, how we have done things so far or cynically trying to protect their own interest by keeping things running as they are now because they know how to maintain their status in the current environment.

    Predicting the future is always tricky but one thing that always seems true is, ‘If you can offer something identical for less money people will buy it off you instead of a the more expensive option that does the same’. The pressure to make things cheaper is not likely to go away and the best price is free. Another rule of business is the established corporations don’t innovate much. Most new advances come from either a big corporations diversifying or a new upstart totally changing the rules. One innovative company could lead us to having a world where everything cost nothing...

    …Well despite the title of this book, that will not quite be true. There will still be things that do cost money or at least you only get if you have status. There are the penthouse apartments in New York, an original copy of something historic, items that are scarce and things like that but the day to day stuff you need to survive will all be free… Even beer!

    The common way of looking at these technological advancements is looking at ‘disruptive technologies’, which are technologies that have totally changed how the world works. The big things like agriculture, tools, weapons, language, banking, the printing press, electricity, the car, aeroplanes, computers, the internet and mobile phones. The disruptive technologies of the last 100 years have made a drastic impact on the world within a decade. Cars replacing horses within a decade, electricity replacing burning oil, computers replacing paperwork and the adoption of mobile phones freeing us from being tied to a table and chair. All have had a massive decade where they have taken off, though this is often not their first decade of being around. To start with these all had their periods of being a novel new idea, along with the people saying this will be the next big thing while other will say it’s never going to work in practice. There is often a period of making it better, more cost effective and safer, till it gets to a point where it’s ready for the mass market. Then it has the massive leap forward to a point where everyone who wants it and can afford it has one. The time it takes to become marketable is dropping really fast too. Computers took 50 years till they reached the point where most people in the developed world could afford one. Mobile phones 25 years and SpaceX has revolutionised launching rockets into space becoming the dominant launch vehicle within a decade.

    Full automation is in the just about to become good enough and cheap enough for mass adoption. We could technically do it nowadays but it’s going to be very expensive, take a while to get set up and sort out the problems. The price is dropping fast and so is the time it takes to train robots and computers to do the jobs we do. Every part of the process needed is being improved for either an automation process or some other unrelated task. Therefore it’s only a matter of time till we hit this golden age of automation where all the bits we need are freely available and people start rolling it out on mass. At this point we could see full automation of almost everything happen within a very short time. Potentially by sometime in the 2030s almost all jobs could be done by a technological solution better than we humans can do them.

    Chapter Two: From the ground up.
     
    Last edited: 22 Mar 2020
  2. classic33

    classic33 Veteran Geek

    Huge questions now being asked over AI capable of "thinking for itself" and programming itself, based on what it's "learnt".

    How far would you be willing to trust it? Would you allow a "self taught" AI surgeon operate on you? Some have likened it to the industrial revolution, only this time the machines may require no watching. No human intervention if things don't work as planned.
     
  3. amusicsite

    amusicsite dn ʎɐʍ sᴉɥ┴ Staff Member

    Location:
    UK
    Would you have gotten into one of the first cars, that were more than likely to explode? Would you get into a modern car?

    Would I let a "self taught" AI surgeon operate on me, probably though I would be a lot more likely in 30 years time when I know they have a 20+ year history of doing it better than a human. Actually one of my friends had one of the first remote surgery operations a little while ago. It was still performed by a human but carried out by a machine and the operation went fine. It's only one more small step to replace that remote human. It's the same as humans. Would you trust a 16 year old person to perform that operation if you knew they would one day be the world's best surgeon? Probably not but ten or twenty years later when they have the experience you would.

    AI nowadays is not quite good enough for most things but it's about 100x better than it was a decade ago. So in another decade's time I'd say it would be good enough for me to trust it in most circumstances. At the moment the AI systems are based on what they have learnt over a relatively short period but give them another 10 years and they will have added to that knowledge all the things they have learnt since that initial learning process. Just like humans, they need to occasional fail and then learn from their mistakes. The difference is that, unlike humans, they should not get old and start forgetting what they have learnt. So at the moment I see them like brash teenagers who can spend a few weeks learning something and think they are the masters of it. After years of experience they will look back at their teenage self and say I'm much better now. AI can get that experience much quicker too, as you can clone the software and get it working on hundreds, thousands or million of instances and lessons learnt can be fed back to all the other AI systems.

    The question to me is not is AI good enough but will it one day get there. I think it will, it's just a matter of how long that takes.
     
    classic33 likes this.