Microsoft stemt in met eisen Europese Commissie

Door Margreth Verhulst, 22 October 2007

BRUSSEL - Microsoft heeft ingestemd met de in 2004 door de Europese Commissie gestelde eis dat het informatie over zijn producten beschikbaar stelt aan andere softwarefabrikanten. Lees verder: http://www.nu.nl/news/1283244/50/Microsoft_stemt_in_met_eisen_Europese_Commissie.html

Via web 2.0 we are nearing super intelligence

Door Gerrit-Jan Wielinga en Bob Overbeeke, 22 October 2007

In September 2007, the director of the Rathenau Institute, an organisation that advises the Dutch parliament on science and technology, called for a political and public discussion concerning the fast development of technologies and its far-reaching long-term impact. The impact appears too great to be easily summarized for the viewers of the Dutch interview programme, Buitenhof.

The scientists at the Singularity Congress, which took place in San Francisco a weekend before, had a much more willing public when it came to the subject of new technologies and superhuman intelligence. Without a hint of taboo, they discussed the linking of people and machines, replacing body parts, uploading the human brain and saturating the universe with super intelligence. Both in Buitenhof and San Francisco a social discussion was proposed.

Developments on the runway
We underestimate the acceleration of technology according to Eliezer Yudkowsky, co-founder of the Singularity Institute. Think, as an example, of the technological developments of a Boeing 747 during its take off. Science is now perhaps in the first 10 seconds of the take off, on its way to equal the intelligence of an insect and then that of a mouse. At the end of the runway, thousands of meters further, but in a relatively short time, science is approaching the intelligence of a fool. Just before we take off at the end of the runway the inevitable happens: we have the intelligence of Einstein. The short time between the creation of the fools and Einstein’s intelligence is insufficient for a far reaching and detailed social discussion. The discussion may as well start now because even the creation of mouse intelligence can have considerable consequences for humanity.

Singularity has several schools of thought. The scientist and writer Ray Kurzweil believes in the law of ‘accelerating change’. According to Kurzweil, our intuitive path of the future is linear but scientific development is exponential. An iPod in 2020 is expected to be able to store every movie ever made. Technologies, moreover, reinforce each other. Research workers from IBM, for example, have successfully switched molecules ‘on’ and ‘off’ for storing information. Nanotechnology can likewise ensure indefinite storage of computer data. Nanotechnology can also accelerate genetechnology and medical science with the use of mini robots and new materials.

Vernor Vince saw it differently. As a science fiction writer in the 70’s he found it difficult to describe characters that were more intelligent than himself. He knew that super-human intelligence would certainly differ from the flying cars in science fiction movies. But how different, he could not suspect. His and his movement’s conclusion was: to be able to predict anything concerning super intelligence we ourselves must be super-intelligent. We look at the future like a snail listens to Mozart. We cannot look beyond the horizon.

Behind the horizon glimmers an explosion of intelligence, statistician I.J. Good foresaw in the sixties. Our spirit, that develops technology to improve the spirit, profits from a ‘positive feedback circle’ that will rapidly create super intelligence, says Good’s movement. The possibilities to improve human intelligence are enormous. Neurones with 1,000,000 times the current speed give the brain a thought capacity of a year in 31 seconds. And speed is only one of the dimensions to be improved. Good thinks “the first ultra-intelligent machine is the last invention that man need ever make”.

Evolutionary design
Besides the three schools of thought of ‘accelerating change’, ‘event horizon’ and ‘intelligence explosion’, several ideas exist concerning the start of super intelligence. It could come into existence by means of human design or in an evolutionary manner. Design and evolution differ considerably. Design assumes control, evolution is not about control. Design deals with simple problems, evolution with complex ones. Design learns, evolution learns from itself. But can we can speak of design by a human being, whereas that human being himself has originated through evolution? If we raise and educate an artificial baby with millions on the internet, is that design or evolution? Probably people will not develop super intelligence but only they can create the correct conditions, Windell Wallach of the Yale Center for Bioethics thinks. Opinions among scientists remain divided. They meet each other on Longbets, a website for making long period bets.

Integration of intelligence modules and technologies will become important. People will understand the process, however, not the final result. Like scientists understand their own models and formulas but frequently don’t understand their results. The programmer of a chess computer, for example, knows it will defeat humans but the programmer himself is no chess champion.

The next step
There is still no consensus concerning the milestones along the runway of making super-human intelligence, Singularity. The Foresight Nanotech Institute realises this handicap and therefore it is creating a roadmap. The current developments on the internet can only afterwards be drawn on the timeline of progress. Now that everyone can find any piece of knowledge on the internet, web 2.0 services and soon the semantic web will give more insight to this knowledge. Is web 2.0, and soon the semantic web, the middle step in the staircase of knowledge, insight and wisdom, as was claimed at the Oreilly web 2.0 conference in 2006 in San Diego? Peter Norvig of Google Research thinks that web 2.0 services certainly can ensure an acceleration towards Singularity. Web 2.0 services massively collect knowledge and collective intelligence of website visitors. This enormous quantity of data could be necessary to reach the threshold that will enable some computer models to function and learn. According to Norvig, Google is in for a change; in its early years the company wanted to create a mirror of the web, but later it came to understand it is part of the force of massive interaction and co-creation which is on the rise on the internet. That is where the company aims for now.

According to ethical futurist Jamai Cascio, the mirror aspiration of Google can be reused. It can turn out very useful to create an intelligent mirrored world beside virtual intelligence. Thus people can possibly predict all consequences of their actions in the mirrored world. Access to this world will appear very important, because the difference in intelligence between the real and the mirrored world will be enormous. Perhaps internet provider XS4ALL (access for all) will have a second youth after Singularity, with a new mission: Artificial General Intelligence access for all. Everyone that wants to lay a wager about that is welcome at www.longbets.com.

This article is the second of four articles concerning Singularity. The first part covers the Singularity conference in San Francisco, September 2007. Part three and four will focus in on morality and consciousness, and the threats and challenges of artificial intelligence.

Videolinks:
Singularity Institute for Artificial Intelligence
Machines with Minds
Human version 2.0

Links:
Ambient intelligence in the health industry (pdf available in Dutch), Rathenau Institute
Ray Kurzweil
Accelerating change
Vernor Vinge
I.J. Good
Foresight Nanotech Institute
Web 2.0
Semantic web
Longbets
The First Conference on Artificial General Intelligence

Een bijdrage van Gerrit-Jan Wielinga en Bob Overbeeke

Gerrit Jan Wielinga is schrijver en free-lance journalist. Bob Overbeeke is business developer bij XS4ALL en schrijft voor www.netr.nl.