I had occasion recently to review again the causes of the few known historical cases of sudden permanent increases in capacity growth rates in broadly capable systems: humans, farmers, and industry. For each of these transitions, a large number of changes appeared at roughly the same time. The problem is to distinguish the key change that enabled all the other changes.
There's more to reward than remuneration. People like Tim Berners-Lee and Linus Torvalds are hugely feted, if moderately paid.Bill Gates has to give away truck loads of $$$ to get that kind of kudos.
The language tells you how to build an AGI. The empirical test is simply writing the program (I have determined that 27 is the minimum number of independent concepts needed to make the language AGI-complete) then it should FOOM. (ie., initiate Singularity). Obviously a highly dangerous 'test' of logical correctness of course.
No 'existence' is too vague a concept to carry much information content. Read what I said in my last paragraph, I am not attempting a grand ontology of everything, I'm looking for an ontology of 'intelligent minds'.
Infinities are a perfectly valid mathematical concept and I have no reason to doubt their existence. The history of math shows that notions such as 'infinity' had a lot of skeptics in the early days (eg Kronecker) , but eventually most math experts came to accept the concepts (probably for good reasons). But I leave questions of 'infinity' to the logicians.
To reiterate, we don't need to start with a grand ontology of everything.
“We are looking for a language sufficiently general to fully describe the domain 'intelligent minds'. This is all we need, because this would include all thoughts that minds could ever think about reality.”
Any experiment to support this hypothesis? (As far as I know, Loglan started with similar goals, but I don't know about any increases in rationality that would result from using it.)
"In what other ways might we dramatically increase our ability to share innovations?"
- perhaps better ways to incentivize the right kind of innovation, and better metrics of success for innovators than the current profit motive.
Having toyed around with startup ideas, I found that a lot of them are unworkable despite having clear benefits to a lot of people. All sorts of annoying, random things can get in the way. The same goes for charity; a lot of the best ideas don't get much money.
Better ways of rewarding people who do good today which benefits the long-term future would be nice too.
E.g. Jimmy Wales is very poorly remunerated given his services to the current generation of children around the world who can all learn from Wikipedia, and all those who benefit from that learning.
What are those basic building blocks? Is "existence" among the basic concepts? (See my "The meaning of existence':Lessons from infinity." — http://tinyurl.com/bx2ujj2 [existence as primitive; sense-data incoherent].)
Would you agree that actual infinities would not be part of such a language? (See "Can infinite quantities exist?" — http://tinyurl.com/aqcy99w )
Academia generates a lot of valuable innovations - for example, I work in industry, and spend a lot of time reading academic papers describing algorithms and systems.
An alternative system for credentialing and funding academics, producing peer-reviewed texts, perhaps using hyperlinks instead of citations might be valuable.
It seems like a more pessimistic outlook might be that the easy ways of improving growth rates have been largely tapped out, as at this point communication between any two persons and trade are quite easy. Any improvements in this might be more marginal rather than revolutionary as in the past.
On the other hand, prediction markets or whole brain emulation might increase the rate of innovation (rather than its sharing), which I think should be seen as a separate avenue of growth, rather than a continuation of the older model.
I thinkthe answer is to be found in the earlier thread ‘Tower Of Babel Still’. By far the most effective way to shareinnovations is improved language coordination. Let me reiterate the key points I made:
“Anentirely new general-purpose 'language of thought' needs to be built up fromscratch, starting from a finite set of ontoogical 'prims', which constitute thebasic 'building blocks' of thought. Only logical pinpointing of the mostprimative categories of thought (the base categorizations used by minds) canfinally collapse the tower of babel.”
I alsoclarified that I was using the word ‘language’ in the most general sense:
“Myusage of the word 'language' is in the most general sense to mean arepresentational system of manipulating concepts and logical relations. So thiswould include mathematics, data modelling and ontology, 'languages' far morepowerful than ordinary natural languages used for speaking.”
I madeclear the aforementioned ‘language of thought’ was not intended to be a fixedontology that attempts to start from omniscience, but an open-ended one thatallows continuous modifications as ever more knowledge comes to light:
“Notone based on a fixed ontology with neccessary definitions, but an open-endedone based on prototypes.”
Finally,I pulled my greatest trick. Rather than attemptingsome grand ontology of everything (that ends up with minimal informationcontent or meaningless concepts), I suggested instead that the ultimate ‘languageof thought’ does not need to deal with all of reality, but only has to describethe domain ‘intelligent minds’:
“We arelooking for a language sufficiently general to fully describe the domain'intelligent minds'. This is all we need, because this would include allthoughts that minds could ever think about reality.”
Innovation (and innovation of innovation, …) is facilitated within an environment supporting replication (such that advances tend to be preserved), a small but significant degree of stochastic mutation (such that entrapment within local maxima can be avoided but without excessive disruption), and novel recombination of loosely coupled functional structures, exploiting synergies self-similar over increasing scale.
For humans, oral and written language greatly facilitated replication. Mutation was a result of noise in all channels of communication, interaction, learning, etc. Recombination was an indirect result of interaction across gradients of geography, resources, etc., and these structures propagated over increasing scale from the domestic group to the tribe to the regional organization.
For farming, methods were discovered, communicated and replicated. Mutation again played a part due to variations in climate, the indigenous ecology, instability of weather and resources, etc. Recombination arose out of specialization of crops and related trade, and synergies were available to be discovered at increasing scale of land, manpower, storage, etc.
For industrialization, again, methods were discovered, communicated, and replicated with increasing efficacy, stochastic variation in quality of replication, quality and availability of resources helped bump the system out of ruts, recombination of loosely linked functional structures abounded in the domains of material, know-how, markets, etc., and methods could be applied with self-similarity over increasing scale.
Now, as for "whole-brain emulations" as a driver of innovation, questions arise. Replication is a given. Mutation, however, would seem to be intentionally limited in order to preserve the integrity of the individual brain. (Emulated human brains, in contrast with anticipated novel forms of artificial intelligence, would lack such intrinsic extensibility.) Recombination could take place *within a culture* of such integral brains but not between the emulated brains themselves, as opposed to much more scale-friendly "hive-minds", or even better, a fractally distributed intelligence. So it seems to me that Robin's whole-brain emulations would at best play an essentially static, temporary functional role but not be a paradigm changer like the examples of human culture, farming, industry [and information technology.]
Prediction markets too seem to lack the evolutionary attributes described above, offering intensional but hardly extensional improvements.
Languages don't tell you things, they just express thoughts you already have.
There's more to reward than remuneration. People like Tim Berners-Lee and Linus Torvalds are hugely feted, if moderately paid.Bill Gates has to give away truck loads of $$$ to get that kind of kudos.
You seem to have confused a necessity condition with a sufficiency condition.
The language tells you how to build an AGI. The empirical test is simply writing the program (I have determined that 27 is the minimum number of independent concepts needed to make the language AGI-complete) then it should FOOM. (ie., initiate Singularity). Obviously a highly dangerous 'test' of logical correctness of course.
No 'existence' is too vague a concept to carry much information content. Read what I said in my last paragraph, I am not attempting a grand ontology of everything, I'm looking for an ontology of 'intelligent minds'.
Infinities are a perfectly valid mathematical concept and I have no reason to doubt their existence. The history of math shows that notions such as 'infinity' had a lot of skeptics in the early days (eg Kronecker) , but eventually most math experts came to accept the concepts (probably for good reasons). But I leave questions of 'infinity' to the logicians.
To reiterate, we don't need to start with a grand ontology of everything.
“We are looking for a language sufficiently general to fully describe the domain 'intelligent minds'. This is all we need, because this would include all thoughts that minds could ever think about reality.”
Any experiment to support this hypothesis? (As far as I know, Loglan started with similar goals, but I don't know about any increases in rationality that would result from using it.)
"In what other ways might we dramatically increase our ability to share innovations?"
- perhaps better ways to incentivize the right kind of innovation, and better metrics of success for innovators than the current profit motive.
Having toyed around with startup ideas, I found that a lot of them are unworkable despite having clear benefits to a lot of people. All sorts of annoying, random things can get in the way. The same goes for charity; a lot of the best ideas don't get much money.
Better ways of rewarding people who do good today which benefits the long-term future would be nice too.
E.g. Jimmy Wales is very poorly remunerated given his services to the current generation of children around the world who can all learn from Wikipedia, and all those who benefit from that learning.
What are those basic building blocks? Is "existence" among the basic concepts? (See my "The meaning of existence':Lessons from infinity." — http://tinyurl.com/bx2ujj2 [existence as primitive; sense-data incoherent].)
Would you agree that actual infinities would not be part of such a language? (See "Can infinite quantities exist?" — http://tinyurl.com/aqcy99w )
Academia generates a lot of valuable innovations - for example, I work in industry, and spend a lot of time reading academic papers describing algorithms and systems.
An alternative system for credentialing and funding academics, producing peer-reviewed texts, perhaps using hyperlinks instead of citations might be valuable.
It seems like a more pessimistic outlook might be that the easy ways of improving growth rates have been largely tapped out, as at this point communication between any two persons and trade are quite easy. Any improvements in this might be more marginal rather than revolutionary as in the past.
On the other hand, prediction markets or whole brain emulation might increase the rate of innovation (rather than its sharing), which I think should be seen as a separate avenue of growth, rather than a continuation of the older model.
I thinkthe answer is to be found in the earlier thread ‘Tower Of Babel Still’. By far the most effective way to shareinnovations is improved language coordination. Let me reiterate the key points I made:
“Anentirely new general-purpose 'language of thought' needs to be built up fromscratch, starting from a finite set of ontoogical 'prims', which constitute thebasic 'building blocks' of thought. Only logical pinpointing of the mostprimative categories of thought (the base categorizations used by minds) canfinally collapse the tower of babel.”
I alsoclarified that I was using the word ‘language’ in the most general sense:
“Myusage of the word 'language' is in the most general sense to mean arepresentational system of manipulating concepts and logical relations. So thiswould include mathematics, data modelling and ontology, 'languages' far morepowerful than ordinary natural languages used for speaking.”
I madeclear the aforementioned ‘language of thought’ was not intended to be a fixedontology that attempts to start from omniscience, but an open-ended one thatallows continuous modifications as ever more knowledge comes to light:
“Notone based on a fixed ontology with neccessary definitions, but an open-endedone based on prototypes.”
Finally,I pulled my greatest trick. Rather than attemptingsome grand ontology of everything (that ends up with minimal informationcontent or meaningless concepts), I suggested instead that the ultimate ‘languageof thought’ does not need to deal with all of reality, but only has to describethe domain ‘intelligent minds’:
“We arelooking for a language sufficiently general to fully describe the domain'intelligent minds'. This is all we need, because this would include allthoughts that minds could ever think about reality.”
#winning
"copying entire minds"
Useless to us (the individuals reading this).
It would just mean new competition and potential enemies.
"In what other ways might we dramatically increase our ability to share innovations?"
It is already here: The internet.
I only miss 2 things:
1) Ownership of ordinary sustainable capital that frees me from work.
2) Marginal improvements in the quality of digital fun.
Innovation (and innovation of innovation, …) is facilitated within an environment supporting replication (such that advances tend to be preserved), a small but significant degree of stochastic mutation (such that entrapment within local maxima can be avoided but without excessive disruption), and novel recombination of loosely coupled functional structures, exploiting synergies self-similar over increasing scale.
For humans, oral and written language greatly facilitated replication. Mutation was a result of noise in all channels of communication, interaction, learning, etc. Recombination was an indirect result of interaction across gradients of geography, resources, etc., and these structures propagated over increasing scale from the domestic group to the tribe to the regional organization.
For farming, methods were discovered, communicated and replicated. Mutation again played a part due to variations in climate, the indigenous ecology, instability of weather and resources, etc. Recombination arose out of specialization of crops and related trade, and synergies were available to be discovered at increasing scale of land, manpower, storage, etc.
For industrialization, again, methods were discovered, communicated, and replicated with increasing efficacy, stochastic variation in quality of replication, quality and availability of resources helped bump the system out of ruts, recombination of loosely linked functional structures abounded in the domains of material, know-how, markets, etc., and methods could be applied with self-similarity over increasing scale.
Now, as for "whole-brain emulations" as a driver of innovation, questions arise. Replication is a given. Mutation, however, would seem to be intentionally limited in order to preserve the integrity of the individual brain. (Emulated human brains, in contrast with anticipated novel forms of artificial intelligence, would lack such intrinsic extensibility.) Recombination could take place *within a culture* of such integral brains but not between the emulated brains themselves, as opposed to much more scale-friendly "hive-minds", or even better, a fractally distributed intelligence. So it seems to me that Robin's whole-brain emulations would at best play an essentially static, temporary functional role but not be a paradigm changer like the examples of human culture, farming, industry [and information technology.]
Prediction markets too seem to lack the evolutionary attributes described above, offering intensional but hardly extensional improvements.
This suggests that when looking for what might cause future increases in growth rates, we also look for possible better ways to share innovations.
For some information about new possibilities that people are trying:
http://creativecommons.org/
http://opensource.org/
http://www.arduino.cc/
http://www.reprap.org/wiki/...