Tuesday, January 31, 2006

Computronium it is

There does not seem to be any reason that the super-intelligences of the future would not place an exigency on computronium.

Even if super-intelligences are not strictly goal-oriented (see "Goal - a temporarily useful construct"), they will likely be active in attempting to unravel whatever is not yet understood about the universe, including how to relocate off the Earth prior to the implosion of the sun (luckily there is some time for this thorny problem, about 4-5 billion years).

Super-intelligences could follow three routes to try to understand what is not yet understood about the universe: 1) directly by their own experimenting, experiencing and thinking, 2) indirectly by perceiving the experiences and thinking of others (with varying levels of sentience) living biologically and electronically and 3) by simulation.

To leverage their resources and inject a diversity of new ideas faster, super-intelligences would likely take advantage of Options 2 and 3, creating or otherwise engendering wetware and software lifespaces/simulations.

To generate and assimilate all of the new information, super-intelligences will clearly require AMCAP (as much computronium as possible).

Sunday, January 29, 2006

Finally, the impending demise of traditional media companies!

Apple and Google's video distribution deals for television content are FAR MORE REVOLUTIONARY than is currently contemplated. The market wants on-demand device-portable video content and there is no reason to pay $40-60+ per month for cable or satellite's non-portable mediocrity.

Current price points of $1.99 per episode easily accommodate regular monthly viewing of 8 programs for $40, less than a satellite TV subscription. On-demand TV program video purchases from iTunes or Google Video can be supplemented with Netflix for recent movie releases and prior season television episodes. News and weather are more expediently obtained on the Internet. With two thirds of US broadband homes being DSL subscribers, there is little to hold them back from cutting the cable!

The current possibilities are wonderful but the future ones may be even better as Apple, Google and other entrants start adding additional audio, video and other content created by traditional entertainment companies, new media and prosumers/individuals. Even right now there is a smorgasbord of TV, music video, short film, video blog, music, podcast and audio book content available. The next level of personalized on-demand information and entertainment will include much more diverse content pulling from meetings, conferences, lectures and classrooms, and also, personal life recorders. No more need to take notes in any situation. The DVR should remind of subscribed to events, like LongNow seminars, Xerox Parc talks or Future Salons, giving the option for live participation and automatic recording.

Wednesday, January 25, 2006

Goal - a temporarily useful construct

Is the concept "goal"” just a temporarily useful human construct? Why would goals not be potentially useful in the future for super-intelligence? Goals are a mechanism of motivation and emotional reward for humans. Since emotion, e.g.; self-actualization and happiness, is probably not the governing driver of super-intelligence, goals are not useful.

If goals did not exist, what would direct behavior and activity? Would there be behavior and activity? Should there be? Yes, there should be activity to avoid entropy, unless the second law of thermodynamics has been re-tuned or otherwise no longer applies. If not goals, what would determine and direct the activity of super-intelligence?

Absent survival imperatives and despite the cliché, the main thing to occupy super-intelligence will be the unfinished business of understanding everything that is not yet understood. As in (currently) the full physics of this universe (and others), dark energy, dark matter, black holes, number, location, purpose and use of dimensions, the Higgs boson, the human brain, protein folding, etc.

Goals will not be needed to motivate super-intelligence to investigate and understand that which is not yet understood, just knowing that it is not yet understood will be sufficient enough to impel activity.

Tuesday, January 24, 2006

AI vs. synthetic intelligence

Augmented human intelligence in the future may still be weak compared to AI and synthetic intelligence. Will the true power players, AI and synthetic intelligence, fear each other and try to exterminate each other? If either is a survival threat to the other, attempted extermination would have to be assumed.

Does exterminating the perceived enemy or peaceful co-existence occur at the next upward levels of intelligence? Fear and the extermination objective will have to exist until intelligence gets away from historical animalistic tendencies. When and if animalistic survivalist tendencies are no longer useful, they will be eliminated.

Barring a clean break with historical Earth intelligence and depending on the timing of the alternative intelligence development and if emotion is a component of AI and synthetic intelligence, it may take a few battleground cycles for co-existence.

Thursday, January 19, 2006

Why do we want next-gen intelligence?

The next generation of intelligence, super-intelligence from our current viewpoint, might not just be computer-related but in the form of one or more of these three things:

1. AI - Artificial Intelligence, computers that have greater than human intelligence, grounded in software and perhaps hardware, not biological wetware.

2. Non-human biologically evolved intelligence from genetic therapy applied to existing biological substrates and/or synthetically created biology.

3. Augmented human intelligence from gene therapy, nanotechnology, implants and/or prostheses applied to humans to evolve super-intelligent humans.

The issue is what happens to us lowly humans in the advent of one or more of these three potential future occurrences. The last case is not a big deal, early adopters and eventually the masses use the technology and become super-humanly intelligent while a small corps of protestors and luddites remain behind.

The first two cases give a little more cause for concern. Non-human super-intelligence will be stratospherically removed from current human intelligence. Dispensing with Kurzweilian optimism, Bill Jovian pessimism and Lanierist quasi-sequitorism and thinking logically, three possibilities exist:

1. Non-human super-intelligence will be indifferent to human intelligence (perhaps most likely but risky since human intelligence could be inadvertently quashed Douglas Adams-like in a minor accident),

2. Non-human super-intelligence will exterminate human intelligence due to perceived competition for resources or some other threat, and/or

3. Non-human super-intelligence will be kind to human intelligence and wants to keep alive or preserve all or part of it. Note the important distinctions between "keep alive and preserve" and "all or part." Again, we cannot possibly imagine superhuman selection parameters for such an exercise, perhaps the individuals with a particular gene are keep alive for further study or some other bizarre situation arises; a welcome invitation to science fiction explore more fully.

Not curiosity and technology lust but rather a broader and better organized agenda should be driving humans to reach this super-intelligent and possibly uncontrollable future.

Wednesday, January 18, 2006

Would you join a collective intelligence?

Thought experiment: Would you join a collective intelligence?

As with anything, many might answer "Yes" assuming that there is a benefit to the individual in doing so and that the benefits outweigh any drawbacks. Most would probably insist on at least:

-being able to selectively turn participation in the collective intelligence on and off and/or fully withdraw participation at will

-being able to segregate the individual brain (both processing and memory) into different permissioned tiers of access

It is a subtle and important issue that the individual must derive a benefit to him/herself as an individual by participating in the collective intelligence.

A pre-assent demo would also be a nice means of encouragement.

What would be some of the potential benefits of participating in a collective intelligence? Some of the potential benefits could include extending one's own and the collective intelligence's capability, experiencing a greater degree of actualization, and meaningful contribution and belongingness to something greater than the self.

Exactly because a collective entity or intelligence is collective implies that organization and perhaps specialization is inherent. The organization may be in hierarchical tiers such as that described in Greg Bear's "Blood Music," in flatter loosely coupled architectures like a peer-to-peer application with resources stepping in as needed or in any variety of other possible and later evolved configurations.

What if you could log your brain into a resource which immediately deploys your mental capabilities in the most appropriate and efficient way? Our true path is found immediately. In the current world, we self-organize, struggling around to find or define our unique place in the world. There are certainly some benefits and growth in the exploration process but it is inherently frictional and inefficient. It would be interesting to have an additional option and experience to be efficiently deployed in one or more collective intelligences. All humans have a need to feel belongingness to something larger than the self. So far humanity has had religion and some other primitive answers to what a collective intelligence could be in fulfilling this human need.

There are several potential objections to joining a collective intelligence; three of the main ones are below. Counterarguments easily refute the concerns.

1. Loss of Freedom - The human illusion and obsession with freedom and fear of the loss of [the illusion of] freedom. The potential losses of freedom in a collective intelligence are loss of volition in thought and behavior and in organizational rigidity, but these are the same in the current world. Assuming value and usefulness, there is no reason not to have volition in a collective environment. Adding skills or knowledge at the individual level would also overcome organizational rigidity and allow tier jumping, the more flexible evolving requirements of a collective intelligence would also likely trigger kaleidoscope re-sorting and re-application of groupings within the intelligence.

2. Loss of Privacy - The twin human fear along with loss of freedom. David Brin's "Transparent Society" is in various stages of being upon us. It is inevitable and not necessarily related to collective intelligence. The idea is to get adjusted to transparency as quickly as possible and enjoy the benefits, one of which would be participating in a collective intelligence. Permissioned access tiers are contemplated as an early prerequisite for one being willing to participate in a collective intelligence. As an interesting example, see Arthur C. Clarke and Stephen Baxter's "The Light of Other Days" exploration of complete privacy loss and the resulting dystopian devolution of humanity.

3. Forced specialization and/or the obsolescence of the human individual. Say collective intelligences take off. What does that mean for the future of human growth and development? One problematic possibility is that because it is easier to free ride on the collective, there is no point in self-development and therefore individuals could become obsolescent. This is also not a worry because the interests of the collective and the individual are aligned. A benefit of the collective intelligence should be that it senses, perhaps better than individuals, the unique strengths and talents of individuals and encourages their further development. Of course this is a selfish motive for the collective as prior to emergent meta properties, the collective is dependent on the growth of the individual minds for its own growth, but individuals also benefit via greater self-actualization.

The limited format of the human individual

Human individuals are limited in two ways, physically and mentally.

The physical limitations are easy to identify, at minimum:

1. Humans individuals are carbon-based which triggers decay, aging and mortality, not to mention the problem of not being able to have a storable backup copy.

2. Human individuals are Newtonian physics law-based; we are generally rooted in space and time (considering the physical parameters and exigencies of the universe, humans are only NOT rooted in comparison to plants).

The mental limitations of human individuals are more interesting:

1. The first level of mental limits is that humans are a predictable result of their experiences, genes and value systems. Any behavior is imminently rational not in the sense of utility per economics but when examined against the backdrop of an individual's value system, beliefs, experience and personality. Free will is a persistent illusion. These limits are extendable - if we want a different behavior, thought or feeling, we just recode past experience, value systems and goals with any of a number of techniques including Neuro-Linguistic Programming.

2. The second level of mental limitation is that humans are individuals, not part of a collectivity. Presumably some interesting future experiments in this area will be coming in the next several decades. Doug Engelbart and successors have long been focused on models of collective intelligence, the launch and impact of which will hopefully start to be seen soon. The next blogpost will address collective intelligence in more detail.

Sunday, January 15, 2006

Yes, but can that Star Trek eyewear read text?

Head-mounted displays (HMDs) are swirling in the market and starting to approach adequate functionality but are still too expensive, too low resolution and strange looking for public use. At MacWorld last week, Massachusetts-based MicroOptical announced that their $269 Myvu glasses for viewing iPod video content will be available in the spring, though limited by the low video resolution of the iPod. The product has already been launched to Orange mobile phone subscribers in France. Other electronics accessory manufacturers already have headset display models on the market of varying cost and type (e.g.; VGA, 3-D, LCD, one-eye/two-eye, $500-800+), such as Virtual Realities and iCuiti with the DV920, 3dVisor, eMagin and others.

Video is nice, but a ubiquitous application for the mobile head-mounted display should be text reading of web content, ebooks and any other kind of digital files. There is a tremendous opportunity to provide an obvious application and correct everything that is wrong with the ebook platform, including that it is a separate platform.

The multiplicity of use cases for mobile text reading perhaps dwarfs music and video consumption: commuting, waiting time, airplanes, at the gym, on the beach, in the bathtub, relaxed reading at home; content types also abound: news, learning-oriented material for school or work, books, magazines, blogs, Internet browsing, etc. Privacy of viewing also need not be underestimated with personal eyewear; though work meetings may need to wait for a less obtrusive headset.

Text reading has long been available on mobile devices but is still primitive and prohibitory in most cases, for example, the iPod's limited screen width and manual scrolling. At least two kinds of functionality are needed: the capability of the text reader application and the ease of loading/refreshing content.

First, the text reader application must have appropriately high resolution, font size, number of lines of text displayed and auto-scroll speed such that the reader can adjust and feel natural with the content and the default is not too far from current written norms. Subsequent application generations can allow for annotation, hypertext and other additions.

Second, for wide adoption, one must be able to obtain content with as little friction as possible. Free content, html (including books), Word, PDF, txt documents, etc. should be easily readable and importable into the text reader management application and RSS/RSS-like subscriptions should accommodate updating content like current blogreaders and iTunes with podcasts. Protected content should be as above and with an easy rights purchasing mechanism.

Full-function text reading should be a feature of every mobile platform ASAP.

Thursday, January 05, 2006

The changing nature of human intelligence...

One way the definition of an intelligent human is subtly shifting is from one who knows information (not data (facts) but information (knowledge derived from study, experience, or instruction)) to one who can access information (knowledge derived by others). As the profusion of data has led to directories, search and ratings as tools to manage the data deluge, people are serving as meta-tools for interpreting and synthesizing information. Accessing information is more important than knowing it directly. Evolutionarily, the fittest are already demonstrating superior capabilities in accessing information and the nature of job requirements is changing to reflect this as well.

Previously information was not available (and not all information is yet or may ever be available) and so there was value in knowing/storing directly-derived information in the brain, but the way humans interact with information is changing. We no longer need to know many things, just how to access that information, for example how to spell (a ubiquitous low level utility function) or how to remember facts (a fairly easy online search).

The question is whether anything is lost in the higher levels of human cognition when the data and information levels are externalized? Is thought processing more limited with fewer facts on board? Probably not in the direct functional sense because important data and information related to our core thoughts of the moment are likely stored automatically just as a function of thought processing, but in the possibility sense, other potentially related areas cannot be easily drawn upon or pattern-recognized without a greater interconnection to data and information.

This fact and information storage issue is presumably only an interim problem before there are direct interfaces from data and information stores to the brain (the seamless in-brain InfoChip or other prosthesis) and the assumption is that higher level human processing can then be more efficient. Better (e.g.; more rational and fact-supported) arguments could then be created more efficiently. The human thought process doesn't necessarily change, it is just improved by having fact-injected thought and the real benefit is in the expansion possibility - to automatically sift through vast additional information reserves to pattern match and learn.

For example, now when considering a problem, we often ask, to what other situations is this similar? We generally only have the capability to look very narrowly in our own experience, some limited experience of others in the same field and maybe talk to a few people who understand the situation. In the future, being able to search across the whole possibility space of human thought and experience to find similar situations and generate universal laws and meta-insights could provide some of the multi-disciplinary energy that has been lacking in our approach to siloed science area problems in almost every field, particularly in Artificial Intelligence and computing, physics and biotech.

Monday, January 02, 2006

The compromise of collective intelligence?

It is interesting to look at the contemporary life extension and collaborative intelligence chatter through the example of octospider species as described by Arthur C. Clarke and Gentry Lee in Rama Revealed (1995, the third book in the Rama trilogy).

The octospiders, a more advanced species than humans, ingest a substance to prevent the onset of puberty. This stifles aggression, sexual desire and ambition and triggers an orderly society working effectively toward common goals. Another result is that lifespans are doubled.

Stifled aggression and a neutralized sex drive (e.g.; auto-controlled emotion) sound potentially helpful, but is anything lost in the stifling of ambition? To gauge a loss from stifling the positive aspects of ambition (e.g.; motivation), it would be necessary to know the meta-goal or primary motivator of the octospiders (e.g.; the surrogate to happiness as the general highest goal of humans), but this is not made clear.

Three of the most interesting issues pointed up by the octospiders example are:

1. What is the role and value of the individual? What is the value of individual intelligence? Studies show that a group of people pooling their efforts towards a goal does better than any one individual. But individual intelligences seem to be a pre-condition for the collective intelligence. With the advent of collective intelligence, does individual intelligence become less strived for and less important? A spark of creativity seems to be a unique and paramount quality of the human individual which should be preserved and extended no matter what the form of future intelligence.

2. The usual debate concerning the role of emotions and drives in intelligence; e.g.; can intelligence occur without emotions and sensory drives (narrowly, is intelligence possible in a non-human substrate?) At first blush and non-anthropomorphically, the answer is probably yes. At the broader level, the question would be what do emotion and drive provide to intelligence and the answer is feedback and maybe something else; the next question is how else could that feedback or other quality be provided by a non-emotion mechanism.

3. Is sacrifice necessary for collective intelligence? The naive promise of collective intelligence is that it is more not trade-offs (like the privacy security continuum). The octospiders example frames collective intelligence as a trade-off between individual expression and peaceful collectively engaged existence. Since the octospiders' meta-goal is not clear, collective intelligence may be a compromise for that species. For the human species, with its current premium on individuality (even in non-Western cultures), any form of collective intelligence which means giving up something would be perceived as a compromise whose benefits would be closely questioned.

Would the pace of progress slow without individualistic drives? It seems that so much of human progress to date has been a function of males seeking status; to be effective, other models for progress will probably dramatically restructure most facets of the complexities and dynamics of human society.