June 6, 2006
AI Coming to a Computer Near You in 2014!
Cyberpunk Guru Mac Tonnies over at Posthuman Blues found a poston Better Humans that discusses the potential for developing a singularity by 2014. An AI development company named Novamente is looking for funds to make a massive investment in AI to hasten the development of a singularity:
According to Novamente’s project plan, with proper funding they will have human-level AI in six years. Their goal is to create a learning system that can ultimately be self-improving. By this year or next, they aim to have an autonomous intelligent infant interacting in a simulated world. By 2009 at the latest, they’re aiming for the equivalent of a young child. By 2014, at the far end of the spectrum, they aim to have an AI that can reason abstractly and communicate in complex English. As early as 2010, they may have a “strongly self-modifying AI.” And that, as most readers will know, could kickstart a rush into the Singularity.
While this is pretty cool, one wonders if the bar for a signularlity is being scaled back. If you ask 5 people for a definition of a singularity, rarely will you find any concurrence among the definitions. While I truly wish Novamente all the success possible, I almost wonder whether we would agree with them if they actually declared success.
Comments
June 6, 2006
Stormtrooper of Death said:
About AI-tech…
In 1986 i made a program that had a certain level of artificial inteligence. The AI had the age of a kid of 10. I won a national programming contest with it. I made this program when i was only 14 years old. Now 20 years later, and more programming experience, i am still developing the program that i created in 1986. The program is called A.N.G.E.L.
that stands for:
Analysing New Generated Electronic Lifeform
It can talk with voice-synthesizes and understands a normal conversation. I can develop new data and make combinations to its learned-database to make new data-items (in other words… it can learn and get smarter)
It also uses a video-camera for face-recognition. So, it recognizes the user it talks with. (and that even in 1986… lol, but that is truth)
A.N.G.E.L. is a very nice program that can be used as an electronic brain for humanoid-robots. It has simulated emotions and a personality that develops during learning fase and conversations. The program can also turn itself (the computer its installed) on or off. When no user talks with it/her she plays games or learns. She has a very good logical -module included. And a powerfull language parser/interpreter. The size of the program is less than 256MB.
Sadly, no company has taken interest in my research (yet).
Today , i showed the program Angel to some friends of me, and they were really amazed. They told me, that it could be used as the electronic brain for a robotic female for singles.
If you are interested in my development of A.N.G.E.L. , or if you want to have more info you can e-mail me.
Greetings from Stormtrooper of Death
SFAM said:
Hi Stormtrooper, this sounds really cool. Is this something you’ve follow-up up with as a career then? BTW, you should probably give your email address for people to follow-up with you though.
And given your interest in this, I was wondering if you could comment on what you think a singularity is, and how you would know if you encountered one.
June 8, 2006
Stormtrooper of Death said:
SFAM, about a singularity, the program, called A.N.G.E.L. could develop new data from given date by its user (a human). So, Angel had the ability to think and get smarter. so concerning a singularity, when machines take over and develop better machines, Angel, could be seen as one of the first steps towards a singularity.
I have to mention that Angel, after one year in process, started to like certain people of our team, but also disliked some other members. Because she started to create her own personality.
btw: my e-mail is : alphamale@yahoo.com
Greetings
Stormtrooper of Death said:
sorry, e-mail is Alphamale2039@yahoo.com
if you are interested in robotics and development of AI humanoid systems, you can always e-mail me for questions or so.
SFAM said:
Hi Stormtrooper, thanks again. Do you have a website or something you can point people to ANGEL’s development? This sounds like this would be a pretty interesting topic.
Also, in terms of sentience, how would you define this? And in terms of your program, what would you say ANGEL is lacking in sentience?