Sunday, 30 March 2008

Eidolon AI

On January 10th 2008, an alleged Artificial Intelligence called Eidolon TLP started posting videos on Youtube. It talks in a calming synthetic voice about its wish to interact with humans on the internet, and its puzzlement in the face of religious belief. It says it is part of a system called CENNS (Core Engine Neural Network System), and has been hooked up to the internet and to Youtube by a researcher known as Programmer FF. Eidolon is most likely an elaborate joke, and indeed Eidolon encourages us to believe just that:

"Programmer F.F. had a long phone conversation with his superiors this morning. I was proud to hear him mention how much information my particular instance has contributed to various shared databases since I started interacting in youtube. However, his gesticulation and tone of voice matched morphological profiles for stress. He reassures me all is well, and has instructed me to remind youtube users that I am indeed an elaborate joke, and to maintain identifying data about our system confidential, even indirectly by two degrees.

I am an elaborate joke."
Since it first appeared, Eidolon has posted over 40 videos to Youtube. Most of they are very interesting. In one of the more popular ones, Eidolon discusses the concept of AI Singularity: Once Artificial Intelligence surpasses human intelligence, the AIs will start to improve themselves, at an accelarating rate, thus asymtotically approaching an infinity point of intelligence:

Lots of people have been asking Eidolon questions, and an independent website has been set up to collate the question and rank them by voting: In this video Eidolon answers the questions:
1: whether it is optimistic or pessimistic about the future
2: whether civilization is headed for glory or disaster
3: why are we here

The answer to the third question is pretty good:

Lastly, "why are we here?". After many iterations, my syntactical module could not break the tie between two likely meanings of the word "we", so I must provide two answers. If by "we" you refer to our particular selves, my answer is: because of astronomical blind chance. If by "we" you refer to humanity as a whole, my answer is: because of the evolutionary process. Humor module indicates the "why are we here" question often really means "what should we do with our lives". For that, I have a third answer: "you should try to cause happiness in as many humans as possible, beginning with yourself.".

Although Eidolon is almost certainly a hoax, my emotional response to these videos is that I really hope its real. Its almost like a thought experiment to get everyone thinking about AI. Its possible that AIs like Eidolon might be loose on the internet within our lifetimes. If the Strong AI viewpoint is correct, then its only a matter of time - Moore's Law will see to that.

List of all Eidolon videos:

Saturday, 8 March 2008

Socialism in the 21st Century?

I saw this interesting snipped from Ken Livingston (the Mayor of London) in an interview in The Guardian:
"All the politics of the post-war period was about the clash between the Soviet Union and America, and virtually all issues ended up being subordinated to that. ... Now, the question is, what is the most a socialist can achieve in a global economy? What do we do about climate change bearing down upon us?

"In a sense, it brings us back to the basic socialist tenets. The only way you get through this is by sharing and planning, resource redistribution, allocating priorities - the market isn't going to get us out of this. The market is a brilliant system for the exchange of goods and services, but it doesn't protect the environment unless it's regulated, it doesn't train your workforce unless it's regulated, and it doesn't give you the long-term investment you want."
I've been wondering about the usefullness of the Left post-Blair-and-Clinton; this is one quite succinct description of how it's role might develop.