The Bewilderness

The Age of Convergence

The New York Times examines the coming superbrain, and details the various schools of thought on the Singularity or artificial consciousness. Most of those fall around two axes: a hostile consciousness that is inimical or indifferent to humanity, or a friendly AI. Researchers who take the Singularity seriously work to set up the ground conditions for the latter.

This most recent flurry of media attention stems from the enthusiastic activities of Kurzweil, who has complemented book with film. This is also coupled with another product of the Terminator franchise, a popkult outgrowth of the anxiety which Vernor Vinge filled in later..

“Kurzweil will probably die, along with the rest of us not too long before the ‘great dawn,’ ” said Gary Bradski, a Silicon Valley roboticist. “Life’s not fair.”

We could argue that the debates on the singularity just pitch too far ahead. The next twenty years are a product of the age of convergence where information and research flow through the white water rapids at an ever faster pace. The formulation of such convergence will have the strongest impact on what type of collective or super-intelligence emerges:

To get any idea what the future might be like, you need to look at all the “currents” of research that are going on at the same time, and think about what might happen when they converge. These convergences can be very beneficial, or very dangerous. The CAT scan, a vital lifesaving tool of modern medicine, is a result of one such convergence (of X-ray imaging, medicine and high-speed computing). The 9/11 World Trade Center attack was made possible by another (of aviation and large-scale building).

We live in an age of convergence and divergence. So far, we have little inkling of where the divergences will appear....

Spectate the Transcendant Man

Clive Davis, in the Spectator, provides a short summary of Kurzweil's idea, in the 'rapture of the nerds' meme, and embeds a video of the neighbourhood dentist opining. Kurzweil has that dental approach of explaining well complex matters that may affect you, and without any sorrow on his part, potentially cause you pain...

Just a pity that there are no public comments, as this would have allowed us to measure the acceptability of the meme.

Transcendant Man

A film has been made about Ray Kurzweil and his ideas, with superhuman nomenclature, the "Transcendant Man". The film was screened at the Tribeca festival.

The film was designed to showcase the ideas in a sympathetic but not uncritical light:

Ptolemy says the title of the documentary, “Transcendent Man” is a play on words — it is meant to describe the human-machine constructs Kurzweil envisions, as well as his efforts to live out his own views. For instance, Kurzweil takes more than 100 nutritional supplements each day, as he hopes to live long enough to benefit from artificial life extension technologies.

Kurzweil, who travels around the world, is an accomplished lecturer and his audiences are largely self-selecting enthusiasts.

The film is available on torrent.

Long life expectancy, the closest continuer and the memory problem

I attended the Transhumanist meeting over the weekend where Dr Richard Faragher gave an enlightening talk on the possibilities of gerontology over the next five years. This is the only timescale that Faragher would entertain as realistic, rejecting some of the wilder claims of SENS. One can see why there may be some antipathy on the part of the laboratory scientist to the self-styled visionary, though this may also be the standard divide between experimentalist and theorist.

My concern  with Faragher's vision was long-lived individuals acting as a vessel with tenants. The neural structure cannot store enough memories to retain a linear psychology down the centuries. Therefore, we need to resolve by storing memories in an artificial substrate and allowing ourselves to tap them as experiences. Overcoming the problem of qualia is one difficulty here. Only then could we say that the tenant in the vessel is the closest continuer of the previous incarnation.


.

Post human potential and AIs

J Storrs Hall looks more closely at bootstrapping self-improving and self-actualising AIs on the Foresight Institute website. With the better awareness of how knowledge increases within scientific communities, a model of self-improvement as the outcome of a network of scientists, is applied to the concept of a bootstrapping AI. Hall concludes that such an AI may appear to have all its thoughts and ontologies derived from human precursors and that, if the scientific community had enough time and resources, would be able to catch up with any AI.

And that means that when we do build a self-improving AI, it will look a lot like the scientific community — but run faster. To state the opposite, you’d have to argue that there was a form of intelligence that was possible (a) to run on a computer, (b) as a program initially written by humans, that (c) could not be simulated by any conceiveable arrangement of intelligent humans running however fast. (I trust I don’t have to point out that the scientific community can simulate a Turing Machine …)

The line of descent is characterised by Hall as memetic, and the outcome is that all AIs descended from humaniform civilisations would have a specific commonality that could be a signature or style.

The scientific community in some sense isn’t composed of people, but of ideas. The humans under different circumstances could just as well have been knights errant, or seafaring traders, or hunters and gatherers. AIs will be composed entirely of human ideas, and everything they ever think of will be traceable back to us in direct line of memetic descent.

Is this one of those brave and striking insights that allow us to reconceptualise bootstrapping AIs or is it a baseline that proves less important than Hall's essay concludes. I suspect the latter, as the knowledge we produce is not restricted to science and man's iterations are not confined to theoretical correlations with observed reality. A more fruitful comparison may be with the mathematical community where abstraction is given full play and the number of potential iterations widens by orders of magnitude. The level of complexity and self-improvement, the lack of a linear line of descent, and potential infinitude of outcomes, should result in possible AIs where a mathematical community of humans could never catch up with the iteration. This is the threshold where the AI ceases to be fully human and becomes something else: posthuman, post human potential, or whatever term fits.  

Leap forward the cognitive enhancers!

There is a call for 'cognitive enhancing' pills to be available over the counter if they are seen to be safe. One candidate, designed to treat the problems of Alzheimer's disease, may be marketed in a weaker form as a lifestyle pill.

If the time ever came that companies demanded parmocological performance for greater productivity, you may literally have to work on a treadmill.

Americans today burn 700 calories less per day than they did during the 1960's. Treadmill desks offer a unique solution compared with fad dieting. They offer employees the chance to burn significant amounts of calories during the day, enhance their health and at the same time enhance cognitive function.

Singularity Summit 1

The Singularity Summit took place this weekend at the same time as the Libertarian Conference. Both complement each other, if like me, you are a libertarian and entertain the possibility of a Singularity. Next Big Future has an interesting summary of discussions: including Intel's plan to swamp the world with digital radios and progress in programmable matter, where they use local rules to combine and form a wrench.

One of the bigger debates that appears to have dominated the conference was whether the state should step in when robots proved to be smarter and more eficient than carbon-based lifeforms, leading to increased unemployment. James Hughes attacks those who argue that free individuals are more likely to adjust to this environment by calling for a minimal wage. Calling his opponents libertopians, Hughes invokes a current era of capitalism to keelhaul the future, though Marshall Brain's particularly vision is unlikely to be realised. I must admit I thought we were beyond one size fits all solutions to what may be a patchwork of local economic effects. So an unrealistic scenario resolved by a contemporary social democratc solution. Hmmm.

The Singularity's 'darwinian default'?

Extrobritannia ran a discussion on the Singularity this Saturday at Birkbeck college, with the usual admixture of strong passion and IT focus. Part of the discussion that threw out a line of enquiry for me was the role that Darwinian forces (the pressures of natural selection) would play upon any artificial consciousness. The original response from other participants suggested that the consciousness would be designed and that Darwinian forces would only come into play if this formed part of the programming.


This focus may tell us about the possibilities of constructing an artificial consciousness without illuminating the forces that may come into play when this consciousness moves into bootstrap phase or starts to design its successors, redesign itself, or reproduce and use of these terms could be interchangeable. Whatever terms we wish to use, the forms of artificial consciousness that come into being can be described as living organisms and will, in various strange and arcane ways, duplicate the phenomena of living organisms: they will self-organise and they will self-produce, possibly reproduce. Could we view a postsingularity environment, with these new forms of consciousness, from an ecological point of view? Would this provide insights that economic, social and political interpretations do not bring to bear?

When I raised this point, my initial view was a 'Darwinian default', with the bootstrapping AI conforming to the agenda of natural selection. Whilst this is a possibility and potentially, a bad and simplistic scenario, there is the potential for new systems of information exchange that are rooted in our understanding of evolutionary change. Our theory of mind, enriched by evolutionary psychology, will form the initial template for communication and exchange.

Singularity encroaches upon the main stream

We are currently still at the media stage: where explorations of the concept of the Singularity form part of a speech by a Chief Technology Officer, hi-tech guru or dynamic professor. The latest example was Justin Rattner, CTO of Intel, whilst talking at the Intel Developer Forum in San Francisco. These displays are useful in so far as they will cite examples of what technologies will play a role in the merging of mind and machine. Always a moving target:

In his closing keynote speech Rattner said that Ray Kurzweil’s concept of ‘the Singularity’, a point when human and artificial intelligence merges to create something bigger than itself, could be just 40 years away.

Rattner described some emerging technologies that sound like they come out of science fiction movies, including shape shifting, programmable matter, neural interfaces that allow applications to be controlled by the human mind, and advanced robots that seem almost human.

With the authority of Intel behind his position, the Singularity gains more mainstream support every year. More and more articles like the interview with Vinge in the NYT will appear and we can expect greater interest in the forthcoming Singularity Summit.

 

The Alzheimer fightback begins

The alzheimer fightback begins. Promising news on the drugs front with a promising new treatment called rember that is currently in trials and shows promise in reversing the effects of cognitive decline. There are two drawbacks to this treatment, neither of which are based upon the drug. One is the length of time that it takes to conduct clinical trials and bring a drug to market; five years is too long. The second is the National Institute of Clinical Excellence:

The drugs are expected to cost the same as current treatments for the illness such as Aricept, which are £2.50 a day.

However, the National Institute for Clinical Excellence (Nice) the Government’s drugs watchdog, ruled that Aricept, which has been shown to improve the memory and day-to-day life of those in the late stages of the disease, was too expensive for widespread use in Britain.

Terry Pratchett, the best selling author who has been diagnosed with Alzheimer’s, disclosed earlier this year that he was being forced to pay for the drug himself.

Drugs that alleviate and prevent dementia pay for themselves in the costs of long term care averted. Whatever system that NICE uses to cost a drug, it is clear that this does not take such matters into account. If the open market prevailed this would not be such a costly error. Yet, whilst the government impedes the entry of new treatments onto the marketplace through regulation and controls access to healthcare, the policy of NICE is a legitimate concern.