Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
One of the biggest arguments people make against the singularity is Moore's law will end in the early 2020's. While that is true the integrated circuit will end when it reaches about 5nm it will not mean the end of computers advancing exponentially as we will move to the next paradigm, 3D self organizing molecular circuits. When I get information on the new paradeigm I have been posting it here and this morning I got the latest one on the self organizing part. This is good in multiple ways. Obviously for this and it will be used in medicine, in this instance to fight cancer.
This is from PHYS.org:
Self-assembling anti-cancer molecules created in minutes
Researchers have developed a simple and versatile method for making artificial anti-cancer molecules that mimic the properties of one of the body's natural defence systems.
The chemists, led by Professor Peter Scott at the University of Warwick, UK, have been able to produce molecules that have a similar structure to peptides which are naturally produced in the body to fight cancer and infection.
Published in Nature Chemistry, the molecules produced in the research have proved effective against colon cancer cells in laboratory tests, in collaboration with Roger Phillips at the Institute for Cancer Therapeutics, Bradford, UK.
Honestly I will admit Ray Kurzweil is the expert on this field not me so I suggest you read the Singularity is Near by Ray Kurzweil. He goes into a lot more detail on how software is not stuck in the mud and is in fact advancing exponentially.
What I can point to is real world examples on how information technology continues to advance on a ever increasing exponential rate like the article I posted above. People are starting to merge with computers now but by the mid to late 2020's it will be normal why 2030 is called life 3.0 or the nanotech revaluation.
I've read Kurzweil's book. At first it seemed convincing, but now not so much, for the reason I gave in my previous post.
I think at this point, all that can really be said is that at any given time in the last several decades, some parameters of technology are increasing exponentially and others were flat or very nearly so. You can certainly make an inductive argument that this should continue for the next several decades. However, that does not make a convincing case for a singularity without some way of determining which technological parameters are relevant and which ones aren't.
Otherwise it's just guesswork and I may as well just roll a die.
I've read Kurzweil's book. At first it seemed convincing, but now not so much, for the reason I gave in my previous post.
I think at this point, all that can really be said is that at any given time in the last several decades, some parameters of technology are increasing exponentially and others were flat or very nearly so. You can certainly make an inductive argument that this should continue for the next several decades. However, that does not make a convincing case for a singularity without some way of determining which technological parameters are relevant and which ones aren't.
Otherwise it's just guesswork and I may as well just roll a die.
Its not guess work and its more then just Ray Kurzweil. I have seen many people talk about Moore's law and the long term implications. I know thinking exponentially is not easy but the singularity will happen by 2045 and life 2.0 will happen around 2020 and life 3.0 will happen around 2030. I have studied this enough from both the con and pro side enough so I can say that with 100% confidence.
Have you read the book abundance is out future or the second machine age?
Last edited by Josseppie; 08-11-2014 at 11:56 AM..
Intel, Moore's Law Still Ticking With 14nm Broadwell Chips
Quote:
Originally Posted by ncole1
I've read Kurzweil's book. At first it seemed convincing, but now not so much, for the reason I gave in my previous post.
I think at this point, all that can really be said is that at any given time in the last several decades, some parameters of technology are increasing exponentially and others were flat or very nearly so. You can certainly make an inductive argument that this should continue for the next several decades. However, that does not make a convincing case for a singularity without some way of determining which technological parameters are relevant and which ones aren't.
Otherwise it's just guesswork and I may as well just roll a die.
Here is more proof that computers have not stopped advancing exponentially. By the time integrated circuits end in 10 years we will have moved on to the next paradigm, 3D self organizing molecular circuits, and computers will keep advancing exponentially.
SANTA CLARA, CALIF. —Moore's Law keeps ticking along and Intel's latest batch of computer chips, processors code named Broadwell and designed for fanless systems like tablets and 2-in-1s, cram more transistors into a smaller, thinner package than ever before.
The key to Broadwell, which will first roll out in System-on-a-Chip (SoC) products dubbed Core M, is Intel's successful ramp of its next-generation 14-nanometer process technology for fabricating microprocessors. The chip giant has lagged a little bit behind its traditional, "tick-tock" pace in getting 14nm up and running, but Intel's Broadwell team on Monday announced that "Broadwell Y" parts are now being produced at volume at Intel fabs in Oregon and Arizona, with a third 14nm fab scheduled to come on line in Ireland in 2015.
Location: where you sip the tea of the breasts of the spinsters of Utica
8,297 posts, read 14,182,127 times
Reputation: 8105
Quote:
Originally Posted by ncole1
I've read Kurzweil's book. At first it seemed convincing, but now not so much, for the reason I gave in my previous post.
I think at this point, all that can really be said is that at any given time in the last several decades, some parameters of technology are increasing exponentially and others were flat or very nearly so. You can certainly make an inductive argument that this should continue for the next several decades. However, that does not make a convincing case for a singularity without some way of determining which technological parameters are relevant and which ones aren't.
Otherwise it's just guesswork and I may as well just roll a die.
Well, in a sense it's guesswork about which advances and bottlenecks will happen (including financially, what continues to get funded and what not - increasingly it seems to depend on DARPA interests). But generally speaking scientific knowledge has adhered to Moore's Law, expanding exponentially like nuclear fission (or yeast in a jar of sugar water), we can make a good educated guess about where it will end up. It doesn't matter whether the parameters exactly match with what we see now, scientists work around roadblocks in creative ways. I suppose it will end up more like Kurzweil's predictions than most would think.
Well, in a sense it's guesswork about which advances and bottlenecks will happen (including financially, what continues to get funded and what not - increasingly it seems to depend on DARPA interests). But generally speaking scientific knowledge has adhered to Moore's Law, expanding exponentially like nuclear fission (or yeast in a jar of sugar water), we can make a good educated guess about where it will end up. It doesn't matter whether the parameters exactly match with what we see now, scientists work around roadblocks in creative ways. I suppose it will end up more like Kurzweil's predictions than most would think.
So far he has been pretty spot on. I can look back on my life and see the changes as I have been around computers since the mid 1970's. I know how it looks and sounds but I also know it will happen. Plus even he admits that post 2045 its impossible to make predictions as we will be advancing so fast.
Look at the computer of 1970 versus the computer of today. My I Phone and Google glass have thousands of times more processing capability then the largest super computers in 1970. That is a huge change. We will see that kind of change in the next 16 years....Because of life 2.0 and 3.0 the majority of us will not die anytime soon. Getting old is just not something our generation has to worry about.
Again, I really do love your optimism, but you're making a fundamental mistake in thinking that increased computing power equals increased capabilities in the real world. Sadly, it just doesn't work that way. Computing power has increased by many thousands of times since the 1970's but this hasn't translated into vastly more capable aircraft, cars, spacecraft, medical technology, or food production. Machines, industrial processes, and other ways of getting things done simply don't break down into mathematical computations all that well. This is especially true of the human brain; no matter how much computing power increases, AI remains as elusive as ever. This is because the brain is not a simple engineering or mathematical problem; it's something far more nebulous and defies any computing algorithm.
I have no doubt your prediction that computational power will increase thousands fold in the next 16 years will come true. But this computing power will not deliver the kind of revolutionary technological change you're hoping for. What will happen is technology and medicine will continue to slowly advance, building on incremental changes and improvements much as they have done since 1970. Yes, computational power will be thousands of times greater, but this will only be partly responsible for humanity's progress. Instead, most of the credit will go to the slow but relentless march of science and good old fashioned human experience (organizations and companies slowly learning how to do things more efficiently). In other words, humanity will advance in much the same fashion it always has.
I'll concede that the rate of advancement is going to be greater as computational power increases, because this increased computing ability acts as a "force multiplier" making engineers, scientists, heck everyone more efficient. But you quickly hit a point of diminishing returns where humans simply cannot advance being a certain pace because we ourselves are the bottleneck. We can only advance so fast so matter how good the tools we have to work with. The inherent limitations of humanity will keep the pace of technological and medical advancement fairly modest compared to the predictions of Kurzweil and other futurists.
At this point you'd no doubt point out that AI will come along and blow past humanity's puny limits and start a cycle of exponential advancement. This might actually happen if a real AI could ever be invented, but that remains doubtful because as I mentioned earlier, AI isn't an issue of computational power. If it were, we'd already have created it.
Again, I really do love your optimism, but you're making a fundamental mistake in thinking that increased computing power equals increased capabilities in the real world. Sadly, it just doesn't work that way. Computing power has increased by many thousands of times since the 1970's but this hasn't translated into vastly more capable aircraft, cars, spacecraft, medical technology, or food production. Machines, industrial processes, and other ways of getting things done simply don't break down into mathematical computations all that well. This is especially true of the human brain; no matter how much computing power increases, AI remains as elusive as ever. This is because the brain is not a simple engineering or mathematical problem; it's something far more nebulous and defies any computing algorithm.
I have no doubt your prediction that computational power will increase thousands fold in the next 16 years will come true. But this computing power will not deliver the kind of revolutionary technological change you're hoping for. What will happen is technology and medicine will continue to slowly advance, building on incremental changes and improvements much as they have done since 1970. Yes, computational power will be thousands of times greater, but this will only be partly responsible for humanity's progress. Instead, most of the credit will go to the slow but relentless march of science and good old fashioned human experience (organizations and companies slowly learning how to do things more efficiently). In other words, humanity will advance in much the same fashion it always has.
I'll concede that the rate of advancement is going to be greater as computational power increases, because this increased computing ability acts as a "force multiplier" making engineers, scientists, heck everyone more efficient. But you quickly hit a point of diminishing returns where humans simply cannot advance being a certain pace because we ourselves are the bottleneck. We can only advance so fast so matter how good the tools we have to work with. The inherent limitations of humanity will keep the pace of technological and medical advancement fairly modest compared to the predictions of Kurzweil and other futurists.
At this point you'd no doubt point out that AI will come along and blow past humanity's puny limits and start a cycle of exponential advancement. This might actually happen if a real AI could ever be invented, but that remains doubtful because as I mentioned earlier, AI isn't an issue of computational power. If it were, we'd already have created it.
Anything that has to do with information technology is thousands of times more advanced then it was when I was a kid growing up in the 70's and 80's and that is directly related to computational power. The same thing happened to medicine when the genome project was complete why things like 3D printers and genetics are now advancing faster then every before. However the changes we saw since 1980 till today will pale in comparison to the kind of changes we will see in just the next 16 years. That is directly related to how advanced information technology will become.
We have not had "good" AI simply because computers have not been that advanced. Yet. I put the good in quotes because we actually have a lot of AI around us and a lot of the things we have and take for granted are AI. However now that computers are at the knee of the curve they will quickly become more advanced passing up humans by the early 2020's. Then you will see AI surpass us and humans will be a bottle neck unless we merge with the computers and that is why we will have to do it by 2030 if not sooner.
Look I know how it sounds but all I can say is get ready because society will advance so fast and so fundamentally that you will not be able to recognize it in 2030. Once 2030 is here we will look back on 2014 as being primitive in comparison.
Last edited by Josseppie; 08-11-2014 at 11:32 PM..
Its not guess work and its more then just Ray Kurzweil. I have seen many people talk about Moore's law and the long term implications. I know thinking exponentially is not easy but the singularity will happen by 2045 and life 2.0 will happen around 2020 and life 3.0 will happen around 2030. I have studied this enough from both the con and pro side enough so I can say that with 100% confidence.
Have you read the book abundance is out future or the second machine age?
No, but you are continuing to ignore the problem I pointed out. I do not have a problem understanding exponentials (I am a physics grad student after all if that matters!). Nor am I saying that there aren't exponential improvements in technology.
What I am saying is, you don't have a non-arbitrary, objective way of deciding which measures of technology are relevant and which aren't. Storage capacity per square centimeter may indeed be exponential over the past 60 years. I am not denying that, and if you think I am, you are straw-manning my stance on the issue. At the same time, the heat tolerance or max operating temperature is the same as it was 60 years ago. The CPU speed of a single-core Intel processor has been flat for 10 years. The time it takes a Windows machine to boot and be fully available for use after switching on has been roughly flat for 15-20 years.
So, in other words, some things have improved exponentially (which I am not denying) while at the same time, other things (I just gave some examples) have been flat.
You believe that machine intelligence will accelerate because (a) you believe it legitimate and reasonable to infer that it will behave like the things which have increased exponentially, and (b) You have an inductive case for continuing the trend seen in the last few decades.
But one could equally well make the converse argument, namely that (a) It is legitimate and reasonable to infer that machine intelligence will be subject to the same things that cause technological progress to halt or stumble; and (b) There is an inductive case for thinking that things which have been flat for 10, 20, or 60 years (see my examples, above) will remain flat for the foreseeable future.
How does one know which of these is more reasonable? The only thing that affects whether you conclude exponential growth or a flattening of machine intelligence, is what parameter you choose to measure technology by. Some parameter choices correspond to exponential growth. Others are flat or nearly so, or have decelerated (in the case of CPU speed was formerly exponential and now flat). Which parameters are relevant, and how do you know that without pre-supposing what you want to conclude? Merely pointing to something that's exponential begs the question against the "flatness" perspective - you need to make a convincing case that machine intelligence's limiting factors are more akin to those parameters that grow exponentially than they are to those that remain flat. Otherwise, you simply have no convincing reason to believe what you claim about a technological singularity involving exploding machine intelligence.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
Intel, Moore's Law Still Ticking With 14nm Broadwell Chips
