Good point but when monitoring Moore’s law, we look at Intel processors. Intel has gotten stuck many times, and it has repeatedly bounced back with vigor. They do not feel “stuck in their ways”.
jldsays:
No, well designed systems don’t fall for local extrema.
I suggest you read more about simulated annealing and (re?)read Ashby’s Design for a Brain http://www.rossashby.info/biography.html#D4B
I do not think we know how to design a self-improving AI system that does not hit diminishing returns.
jldsays:
Agreed, but IMHO the “diminishing returns” will have more to do with P!=NP than with local extrema.
BTW, in case you haven’t read it, Design for a Brain isn’t really about designing a brain or other AI but about the properties of ALL self organizing systems, the system always “wins” but is not what you would expect…
I disagree, at least a little, about transferable skills. The most valuable skills are intangible and transferable. Specific, objective, codified knowledge is not transferable, but it’s also easily commoditized. Knowledge of Java, per se, is a commodity. Knowledge of how software projects work is less tangible, more transferable, and more lucrative.
What I mean by being transferable is that if you know English then you might think that this helps you with Latin. I think you allude to more generic or general skills as opposed to transferability. But even then, I would caution people that skills and knowledge are often less generic or general than it seems. And that’s a problem because people have the illusion of “knowing it all” when, in reality, their knowledge and expertise is far more limited. Someone may know a lot about how to build software for a web startup, but that might prove of little help when building software for embedded devices in an Internet-of-things setting or when building software for high performance computing. I think that no matter at what level you work in, it takes effort to remain “young” (avoid “cognitive rigidity”). To make matters worse, it may often make not economical sense to fight cognitive rigidity.
I agree that it takes effort to remain flexible, and there are (short-term) economic incentives to not be flexible.
It’s another efficiency-robustness trade-off. Specialization is efficient but not robust. In biological terms, a species that is too well adapted to its environment is likely to go extinct when the environment changes.
Simonsays:
“There is very little difference between spending 5 years programming Java and spending 20 years doing the same.”
Do you have any stats to back this up? Having been a programmer for 35 years, I have seen some big ‘aha’ moments in my development as a programmer… But spaced out by 5, 10, or 15 years at a time. I can only speak for me, but I feel as if my productivity, achievements, and enjoyment have continued to increase as time has gone by.
You write about people assuming that the way things are must naturally be for the best (I paraphrase), however, I would argue strongly that there is a “scientific bias” (also caused by our only focussing on what we know we know), that we can somehow know best. This leads us into such things as environmental damage, growing monocultures (and then using vast quantities of pesticide and herbicide) and other short sighted practices.
There may be advantages in editing the human species to be longer lives, less prone to disease etc. There may also be disadvantages. We will surely discover those as we start editing, just as we have started to discover the shortcomings of editing our environment.
Byron Allen Blacksays:
Condet, East Jakarta is a traditional ‘Betawi’ area (the Betawi are thought to have been the original inhabitants of what is today a 20-million-strong megapolis) in the southeast corner of the city. I’ve lived here since 1988.
In my early years I was amazed to meet energetic, apparently intelligent 18-year-olds who had practically never ventured beyond these local neighborhoods – and who saw nothing unusual about this. Their entire world was encompassed in the ‘kampung’ and this would naturally prove to be a titanic limitation in life, when asked to comprehend the world at large. For instance, Indonesia has millions of ‘TKI’, ‘guest workers’ in Hong Kong, Malaysia and the Arab east, and they must quickly learn to deal with an extremely alien environment. Many are abused, underpaid or driven to desperation (and crime) when they find themselves helpless in such a situation. Later I extrapolated this to attempt to understand the mindset of my fellow ‘hard-shell Baptist’ Texan kinfolks and acquaintances, people who had never bothered to travel or learn another language. I had personally had the blessing of being an ‘Air Force brat’, and lived in Bavaria during the Occupation in 1946 and 1948. I traveled extensively from my teens and attempted to integrate into alien European and Asian cultures. Later it was travel through psychedelics, including LSD – another experience impossible to describe to others or for them to attempt to comprehend. Somewhat reluctantly I attempted to fit into academia but also found that scene self-limiting, territorial, resistant to leaps of concept and generally suspicious of any proposition or experience which does not look backward. The current overt disinterest in and reluctance to accord credence to mystical experiences or extra-terrestrial sightings / encounters stands as another huge example of this (check out the YouTube video on retired USAF officers and UFOs interfering with nuclear weapons bunkers, in a presentation given at the National Press Club). Who is prepared to come face to face with alien beings? It’s going to be the ultimate test of what Daniel terms ‘cognitive rigidity’. I apologize for the long-winded exposition; this essay excited me because of my ongoing interest in and research into the phenomenon of ‘uncertainty avoidance vs. uncertainty tolerance’.
Though I’m now at the point of retiring, I feel so lucky to have earned a very good living in an industry that has undergone massive changes and which has given me the opportunity to delve into so many interesting areas of research and technology and yet never become a part of any of them. I’ve talked at length to experts in almost every field of science and industry, and have travelled the world to record the things they’re working on. I trust therefore that I can say honestly that I’ve managed to keep an open mind, but have seen examples of what you’re talking about. And the more I’ve learned the more I’ve realised how little we know and the more there is still to learn.
I’m a film maker who now grows trees and who, with my wife, provides holiday accommodation in barns I’ve converted myself.
Didn’t Intel lose the mobile revolution just as Microsoft did? Afraid I’m not old enough to remember Intel ever re-inventing itself.
Good point but when monitoring Moore’s law, we look at Intel processors. Intel has gotten stuck many times, and it has repeatedly bounced back with vigor. They do not feel “stuck in their ways”.
No, well designed systems don’t fall for local extrema.
I suggest you read more about simulated annealing and (re?)read Ashby’s Design for a Brain http://www.rossashby.info/biography.html#D4B
I do not think we know how to design a self-improving AI system that does not hit diminishing returns.
Agreed, but IMHO the “diminishing returns” will have more to do with P!=NP than with local extrema.
BTW, in case you haven’t read it, Design for a Brain isn’t really about designing a brain or other AI but about the properties of ALL self organizing systems, the system always “wins” but is not what you would expect…
I disagree, at least a little, about transferable skills. The most valuable skills are intangible and transferable. Specific, objective, codified knowledge is not transferable, but it’s also easily commoditized. Knowledge of Java, per se, is a commodity. Knowledge of how software projects work is less tangible, more transferable, and more lucrative.
What I mean by being transferable is that if you know English then you might think that this helps you with Latin. I think you allude to more generic or general skills as opposed to transferability. But even then, I would caution people that skills and knowledge are often less generic or general than it seems. And that’s a problem because people have the illusion of “knowing it all” when, in reality, their knowledge and expertise is far more limited. Someone may know a lot about how to build software for a web startup, but that might prove of little help when building software for embedded devices in an Internet-of-things setting or when building software for high performance computing. I think that no matter at what level you work in, it takes effort to remain “young” (avoid “cognitive rigidity”). To make matters worse, it may often make not economical sense to fight cognitive rigidity.
I agree that it takes effort to remain flexible, and there are (short-term) economic incentives to not be flexible.
It’s another efficiency-robustness trade-off. Specialization is efficient but not robust. In biological terms, a species that is too well adapted to its environment is likely to go extinct when the environment changes.
“There is very little difference between spending 5 years programming Java and spending 20 years doing the same.”
Do you have any stats to back this up? Having been a programmer for 35 years, I have seen some big ‘aha’ moments in my development as a programmer… But spaced out by 5, 10, or 15 years at a time. I can only speak for me, but I feel as if my productivity, achievements, and enjoyment have continued to increase as time has gone by.
You write about people assuming that the way things are must naturally be for the best (I paraphrase), however, I would argue strongly that there is a “scientific bias” (also caused by our only focussing on what we know we know), that we can somehow know best. This leads us into such things as environmental damage, growing monocultures (and then using vast quantities of pesticide and herbicide) and other short sighted practices.
There may be advantages in editing the human species to be longer lives, less prone to disease etc. There may also be disadvantages. We will surely discover those as we start editing, just as we have started to discover the shortcomings of editing our environment.
Condet, East Jakarta is a traditional ‘Betawi’ area (the Betawi are thought to have been the original inhabitants of what is today a 20-million-strong megapolis) in the southeast corner of the city. I’ve lived here since 1988.
In my early years I was amazed to meet energetic, apparently intelligent 18-year-olds who had practically never ventured beyond these local neighborhoods – and who saw nothing unusual about this. Their entire world was encompassed in the ‘kampung’ and this would naturally prove to be a titanic limitation in life, when asked to comprehend the world at large. For instance, Indonesia has millions of ‘TKI’, ‘guest workers’ in Hong Kong, Malaysia and the Arab east, and they must quickly learn to deal with an extremely alien environment. Many are abused, underpaid or driven to desperation (and crime) when they find themselves helpless in such a situation. Later I extrapolated this to attempt to understand the mindset of my fellow ‘hard-shell Baptist’ Texan kinfolks and acquaintances, people who had never bothered to travel or learn another language. I had personally had the blessing of being an ‘Air Force brat’, and lived in Bavaria during the Occupation in 1946 and 1948. I traveled extensively from my teens and attempted to integrate into alien European and Asian cultures. Later it was travel through psychedelics, including LSD – another experience impossible to describe to others or for them to attempt to comprehend. Somewhat reluctantly I attempted to fit into academia but also found that scene self-limiting, territorial, resistant to leaps of concept and generally suspicious of any proposition or experience which does not look backward. The current overt disinterest in and reluctance to accord credence to mystical experiences or extra-terrestrial sightings / encounters stands as another huge example of this (check out the YouTube video on retired USAF officers and UFOs interfering with nuclear weapons bunkers, in a presentation given at the National Press Club). Who is prepared to come face to face with alien beings? It’s going to be the ultimate test of what Daniel terms ‘cognitive rigidity’. I apologize for the long-winded exposition; this essay excited me because of my ongoing interest in and research into the phenomenon of ‘uncertainty avoidance vs. uncertainty tolerance’.
I know where you’re coming from.
Though I’m now at the point of retiring, I feel so lucky to have earned a very good living in an industry that has undergone massive changes and which has given me the opportunity to delve into so many interesting areas of research and technology and yet never become a part of any of them. I’ve talked at length to experts in almost every field of science and industry, and have travelled the world to record the things they’re working on. I trust therefore that I can say honestly that I’ve managed to keep an open mind, but have seen examples of what you’re talking about. And the more I’ve learned the more I’ve realised how little we know and the more there is still to learn.
I’m a film maker who now grows trees and who, with my wife, provides holiday accommodation in barns I’ve converted myself.
Thanks.
If your health is good, consider delaying retirement.