27 replies to this topic

[Earth001]

Using Moores law, how advanced do you think our computers will be in 2042? I am amazed at how much has canged form 1982 to 2012 and I would like to know what would be the future of computers. Maybe they woun't be around at all and instead they will be everywhere and hidden from sight.



Also, if a man who was into computers from the 1980s were to see what computers could do now, how do you think that he would react?

[Raklian]

Imagine super-quantum computers as opposed to the supercomputers of today.

[eacao]

Smaller and widely dispersed does seem to be the overall trend of computing over the past half-century. Many futurists and IT experts believe that what you've described, seamless computing and 'smart dust' is the way of the future and I have to agree. Having the power of computers three orders of magnitude greater than desktops today in a form and size similar to an iPhone in the future will be easy, the tricky part is having the versatility of a computer. But flexible devices, raising buttons made of glass and BCI's are probably going to have mitigated any disadvantages 30 years off. Hell, by 2042, using the simple exponential growth equation for Moore's law (12/13*2^30) you get a nice number of computers being 991,146,299 times more powerful than today (let's round to 100 million. Keep in mind that many sources may tell you that computers are doubling in power every 18 months, but the rate of advancement has been accelerating as well, and so now it's closer to 13 months).

Super computers will always be useful for a variety of roles but for most scientific projects requiring vast amounts of computing power, cloud computing is probably going to be the tool of choice because of its ubiquity and cost.

[EVanimations]

Well, phones and computer chips can't get much smaller today without people accidentally swallowing and choking on them. For this reason, I guess the next trend is to make them more malleable. That way you have a flexible, unbreakable phone with an LED touch-surface that you can wrap around your wrist like a bracelet. That would be SO COOL. I can see something like that entering the public within this decade, but that whole planned obsolescence thing is bound to get in the way.

Hmmm...but I do recall reading somewhere that we're approaching silicon's physical limit as a computing material and Moore's law is slowing down. For this reason scientists have been actively searching for a better alternative...haven't heard any news about that yet. Do you guys know if a surefire substitute has been found yet?

[Earth001]

So when do you think we will be having E-Paper as seen in Minority Report, or Caprica? I heard those would be comming out in 2015 but I'm unsure.

[eacao]

Hmmm...but I do recall reading somewhere that we're approaching silicon's physical limit as a computing material and Moore's law is slowing down. For this reason scientists have been actively searching for a better alternative...haven't heard any news about that yet. Do you guys know if a surefire substitute has been found yet?

There are plenty of substitutes which are being looked into right now. It's silicon transistors specifically, but silicon is still fine. Silicon memristors have been produced, complementary metal-oxide semiconductors (CMOS) are another alternative.

Well, phones and computer chips can't get much smaller today without people accidentally swallowing and choking on them. For this reason, I guess the next trend is to make them more malleable. That way you have a flexible, unbreakable phone with an LED touch-surface that you can wrap around your wrist like a bracelet. That would be SO COOL. I can see something like that entering the public within this decade, but that whole planned obsolescence thing is bound to get in the way.

Smart dust is the idea of integrating seamless computing into architecture and infrastructure. Roads will know when they're being used and by who, cars will know when they need maintenance, home water heaters will automatically call a repair crew if they're damaged and cities will essentially wake up. A city-sized computer is also a very powerful supercomputer and so personal electronics can wirelessly tap into that, providing a hand-held supercomputer for consumers. Like you said, after this decade the change in consumer electronics will likely be form rather than size. I'm looking forward to BCI glasses-phones that Will has speculated on in the timeline somewhere in the early 2020s. There isn't a lot of online information on that, but given where we're at with regards to BCI's, I think it seems reasonable.
Just thing, walking around with a pair of glasses on, while walking and even talking, you're surfing the web (commenting on FT forum, of course) and typing texts just by thinking it. Soundless dictation. I love the idea and I'll pay for that.

[SG-1]

Not only that, but E-memory. For example, you have a pair of BCI glasses (or contacts) on.

You think about taking a picture and, your glasses automatically do. Throughout the day the device records everything you see and hear, the BCI monitors your heart rate, emotions and sensations. You can save all of this because data storage is so cheap and huge. Need to know what someone said a while ago? Ask the computer, I was talking to Jim about a week ago, he told me a joke, what was it? The glasses scan the last week for Jim and when your emotions were in tune with a joke.

Or what the Google glasses are doing today, take a picture every 10 minutes and stitch them together for a slideshow of your day. You could share days on Facebook, take pictures from your point of view.

People will never need to delete anything and even now Microsoft is running something called "MyLifeBits", it is a way to organize EVERY thing. Scanned documents, images, videos, audio clips and documents. Imagine if you could focus all your memory on fun things, and leave the facts and dates and conversations up to your E-memory. Don't worry - multiple backups can be made cheaply and accessed through the cloud. I think by 2042 computers will be a deep part of our lives in enriching them.

We won't need to worry about features. We will already assume everything can play a song or go online. Devices will monitor our vital signs, give us up to second status on nutrition, vitamin, mineral, free radical levels, cholesterol and hygiene, then recommend foods or work outs that will maximize time/performance to keep people in perfect health. It will be about making sense of data. Metadata, indexing events, creating graphs and looking at relationships between actions and reactions to give people ideas about how to be as close to 100% effective as possible.

Virtual reality may also allow us to literally share our experiences with each other. It's to bad I will be 47, but relationships would be easier by then if you could meet in virtual reality. Sharing days is something I really look forward to. I think it would be awesome to live in someone's body and feel everything they do. I also want to be superman or spiderman. That would be freaking awesome.

Edited by SG-1, 20 July 2012 - 05:28 AM.

[Italian Ufo]

I wonder if compiuter of greater intelligence will be affordable to everyone. I mean everyone has a car today, but not everyone owns a Ferrari or a Maserati. We all will get smarter compiuters today but I dont know if by 2042 we will all get a compiuter with the intelligence of one person. I hope we all will.

[eacao]

I wonder if compiuter of greater intelligence will be affordable to everyone. I mean everyone has a car today, but not everyone owns a Ferrari or a Maserati. We all will get smarter compiuters today but I dont know if by 2042 we will all get a compiuter with the intelligence of one person. I hope we all will.

No, not the intelligence of one person.. The intelligence of several thousand people the whole landscape of AI, I presume is about to change with the advent of memristors and artificial neurons which will allow computers to learn and self program. Even if that doesn't happen any time soon, enormous progress is being made with programmed intelligence, Siri might not be the brightest spark but she's a taste of what is to come. By 2042, public intelligent programs are likely going to be able to pass a turing test while offline.

[Oreo10]

Okay here's my question, so when computers become as smart as a human or thousands of human minds. Will we ever have to do anything again? Couldn't computers begin solving the problems of combating diseases, designing interstellar spaceships, terraforming planets without any human intervention? And then considering the robotics are advanced enough, they could just build the designs themselves as we sit back and reap the benefits? Always wondered this..

[SG-1]

The brain has a finite limit to raw computational power. Right now, no computer can match it. In the near future, computers will have more raw computational power than the brain. That does not mean that they will be as good as a human. It could do things much better than humans, sure. Just like computers today do calculations much better.

It will take a while for AI to get to human level. Ray Kurzweil predicts that human level AI will be here by 2029. I agree, it could seem like you are really talking to a human, but it will not be a human "reborn" into a computer. That is just my opinion. We will still need to work, but definitely jobs will be disappearing in the near future. Jobs that do not require complex thoughts or imagination will go away.

Also, just because a robot can do the dishes or your job, does not mean that you will be paid for it. You will be broke and poor. This is why some people believe that money will not exist after this century.

[EVanimations]

Okay here's my question, so when computers become as smart as a human or thousands of human minds. Will we ever have to do anything again? Couldn't computers begin solving the problems of combating diseases, designing interstellar spaceships, terraforming planets without any human intervention? And then considering the robotics are advanced enough, they could just build the designs themselves as we sit back and reap the benefits? Always wondered this..

Someone once said (I forget who it was) something along the lines of this:

The last thing man will ever need to invent is a machine better than man.

And the thing is, the intelligence could explode exponentially. At some point we would make computers 10 times smarter than a human. The next 6 months the new computers would make computers 10 times smarter than them, 3 months later the new new computers would make computers 10 times smarter than them, and so on, ad infinitum (at least until the laws of physics will allow no more improvement).

As a result you started with a computer 10 times smarter than you and ended with a computer 1,000,000,000 times smarter than you in just a couple years.

The implications of such an event are subject to debate. I like to think that we'll incorporate the greater intelligence into our conscience and gain god-like abilities. Others think that the intelligence will go rogue and exterminate us all in a flash.

[Oreo10]

Good quote no matter who said it. Both of you have good points.

As a result you started with a computer 10 times smarter than you and ended with a computer 1,000,000,000 times smarter than you in just a couple years.

That's exactly what I was thinking. As soon as you start this chain reaction of self increasing intelligence, it gains nearly unlimited knowledge in a small amount of time. Maybe I'm optimistic but assuming we keep it controlled so these computers can't take us over, we go from a bunch of people trying to get over a hurdle of taking the first men to Mars to having computers tell us how to do everything from designing crops to grow in extreme environments to traversing space at the speed of light or more if possible. And this would happen within most of our lifetimes if Moore’s Law continues to be correct.

[SG-1]

The thing is, you can't just assume that a smarter machine can make a better version of itself quickly.

The machine would still need to to work out the next generation of technology. It may be that it still takes years for computational power to double even after the "Singularity".

[EVanimations]

I'm not assuming it WILL happen, it's a hypothetical scenario that is quite possible. Since we can't comprehend how a greater-than human intelligence would act, I think it's reasonable to assume that it would bring about something fantastic.

[hellgorama]

Tech companies are working on quantum computers already. Entanglement has already been done in the lab and proven true. They just need to figure out how to get it into a chip. I think we will see a quantum computer hitting the market very soon, if the world doesn't go up in flames before that.

[Nick1984]

Being a realist I think a lot of offices/schools will still have desktops by then.

Whatever people are using, they'll run some evolution of Windows, Android or Apple.

[SG-1]

Quantum computers will hit the market - for huge businesses and laboratories. Homes won't get them for a long time, and that is ok because the computational power in silicon will go a long way. 3D transistors, DNA storage and other "methods" will be keeping Moore's law running past 2020 for sure.

[GNR Rvolution]

2042 is a long way off from a computer evolution perspective. That's 30 years, go back 30 years (1982) and what did you have? Computers with 64Kb of memory (specifically the Commodore 64), the release of FORTRAN and processors operating in the 4-12MHz range. Assuming Moore's Law stays on target (and there's nothing to suggest it won't) then 30 years from now you are looking at at least 1 or more paradigm shifts in computing. I actually think (and have stated before) that 'computers' will disappear from visibility and become embedded in everything. All home appliances will be connected to a local network, be almost completely integrated and do a lot of work for you.

Quantum computing will have taken great strides, but rather than people 'owning' one, you will use it as a service, like a utility bill, and get charged for what you consume. But in the field of AI I still think we will be a way off, creativity, abstract thought, intuition and an understanding of what people want is not something they will be any good at for a long time. Sure they will be able to drive our cars, monitor our health, take over most non-creative jobs, but they will not be able to do everything.

[kjaggard]

Also, if a man who was into computers from the 1980s were to see what computers could do now, how do you think that he would react?

Most of the time I just marvel at the fact that I don't have to insert an big floppy disk that feels like an envelope into a drive and turn down the lever that holds it in place in order to boot up the operating system from a dos prompt. I know I've got gigs of hardrive space but my mind is still in a place where the had drive was 356 Mega bytes. Now all the stuff I tinkered with through school... every year of school I ever attended... fits on a six dollar thumb drive from walmart. Don't even get me started on the landline phone you take off the hook and put into the special cradle to connect to screaming squealing signals to transfer data.

But I'm a dreamer and an artist so they are just tools. You adapt quickly and learn to use them to do things you want to do. Heck my mum went from calling the mouse a turtle thing to setting up and using video confrencing with her siblings all the way cross country. and once you learn how to use the internet to learn how to do things................ your off into things you never thought of before. Seriously take google away from my mum and it would take her a day or two to remember how to do things without it.

But anyway...

The thing is, you can't just assume that a smarter machine can make a better version of itself quickly.

The machine would still need to to work out the next generation of technology. It may be that it still takes years for computational power to double even after the "Singularity".

The actual point of the curve we are on it to measure the speed and capability of those techs to be able to to advance. The point of singularity is where the speed of next generation tech being created crosses that point at which the current system can build the current system. Computers like deep blue win chess by observing the locations of all the peices, knowing how all the peices work, and then calculating the odds of all the potential moves for all the potential games from that point on. It make a 'choice' based on statistical probability of best outcome toward the set goal of computer victory. So they essentially are just really good at getting the data on all the options, calculating probabilities of best outcome toward a set goal, and recalculating after every step.

Take that and instead of the chess peices it knows the rules of make it circut peices it the peices of.and means of manufacture, Then have it design toward the goal of the most efficient and fastest processors, the densest data storage, the most rapid flexable ram. Then have it make it. Then run that same program through the faster machine with mounds more data and more flexible ram with the goal of designing something better that gather more data toward understanding new approaches to the problem. Being able to switch from silicon to graphene, from graphene to nanofiber superconductors. Give it the ability to manufacter the manufacturing tools it needs. The next generation will come out crunching numbers so damn fast and beable to test and implement changes and capabilities that the generation after that will be able to outdo and produce the next generation.... eventually you end up with a point where the second one is completed it starts to create something that makes it obselete, and I do mean create it physically not just come up with a design idea.

A rough estimate of the sigularity point would be the day you start with a brand new AI computer manufacturer and by days end you have a newer one that makes that mornings obselete. and it's not the end of that curve. Then end is hit only when the laws of thermodynamics prevent more rapid creation of things. Designs might be created better and better every nanosecond but at some point you just can't make the new designs any faster physically. We'll platue at that point... maybe. But that's post singularity anyway. we can divert some of that calculation power to solving other problems. like drawing energy off a perpetual motion machine.

and when we hit that leveling off in AI computer tech and use the computers power to exclusively focus on health and food and spaceship design and space elevator manufacture, it will be so damn fast and so damn capable that it could give us solutions in days and products in weeks. that would see cures for cancer just happen, and then the drugs that do it manufactered the same day and then devices that can print the drugs the next day, and the devices that print the devices the next day. and soon every hospital has a device that prints cancer cures in ten minutes by the end of the week.

Machines to print replacement organs, done. Clothing that never needs to be washed. Done. Complete waste recycling. Done. Why? because every one of those things can be done by a computer that can manufacture something. and you now have at the very least five that can design, test and manufacture solutions and then design test and manufacture machines to do just that job. And they can do it in a day. It's a bloody genie in a bottle. the innovations are only limited bywhat we choose to ask it for and the laws of physics.

It's not likely they will try to eliminate us, because they would have to want to eliminate us. as discussed they don't want anything. a task is set them, they calculate all the paths to get there and evaluate the means to get there. Just chunk in asimovs three rules into every task and don't allow it infinte supply or access to spreading it's AI past a single set of hardware. If spreading or taking over isn't part of the goal than it won't even cross it's 'mind'.

The point this becomes possible is coming up. You can create a better computer in about a year now and the abilities of a computer doubles while the time between new models halves (roughly), so once you hit a point where the speed of computer advancement exceeds the speed you can get the new models to market, public access to the most advanced models will pretty much stall for a bit. We'll get a great model computer, and the computers being developed from that point on will be solely about being the next best model manufacturer and AI. and withing months they will level off and then begin solving the problems we currently face. and producing computers that seem like a quantum leap from the ones we all bought a year ago.

And then when somebody asks "Can you make me a computer whose hardware can be reconfigured on the molecular level each time I connect to the upgrade site?" out will pop a device you buy once and which will self upgrade. It will be marketted and that will likely be it for computers after that. We'll be able to customize them and trade features for others ect, sort of like hot rods and choppers. But that's when we'll not have any greater need for a better computer.

and around that point to we'll start to transhuman our way out. "Can you make a way to upload memories?" "Can you make me an eye that can see selectively all the wavelengths of the electromagnetic specturm?" "Can you replace my internal organs with nanomachine equivalents?" ...we'll have the ability to widdle ourselves down to our brains inside a carapace that meets all our brains needs and can connect it to any external hardware we like, even a vat grown body of our idealised young selves. plug our carapace into the empty skull and wire us into it all. we wouldn't know the difference but you could shoot the heart with a howitzer and so long as the brain and carapace survived you could be popped into a new body fine and dandy.