[ExI] ai class at stanford

[Samantha Atkins]

On Aug 22, 2011, at 7:32 AM, spike wrote:

>  
>  
> From: extropy-chat-bounces at lists.extropy.org [mailto:extropy-chat-bounces at lists.extropy.org] On Behalf Of Samantha Atkins
> …
>  
> …
> …
>  
>  
> >>…This is a smart enough group, I want to see some in-depth discussions, and perhaps some criticism of the concepts.  I am hoping this class is the catalyst to get some fizz going on ExI-chat… spike
>  
> >…Criticism of the concepts and especially techniques of narrow AI (which is what this is) before you have learned them does not seem worthwhile.  If they are covering as much as stated many of us will be scrambling to upgrade rusty skill sets.  – Samantha
>  
> Ja, that is a lot of how I am hoping to use this class.  In general it feels to me like the very broad field of AI is deeply divided into fractions that do not talk to each other.  Even if you focus down on AGI, it immediately divides itself into the singularity crowd, the game learning crowd, the neural net old timers, friendly AI sorts such as Eliezer, and plenty of others.  If you have ever been to Eliezer’s events, you find plenty of people with very similar interests who do not share views on basics and don’t seem to creating much synergy.

I have not seen quite so much seeming division in the narrow AI field and likely will not in this course form a rather unified highly rated text.  Reading the papers of the AGI conference also dose not give as much feel of division as mailing lists and impromptu discussions seem to.  Of course there are wide divisions in approach to AGI but that is bound to happen with any goal we don't have a really thorough generally accepted idea how to achieve.    This course is more about AI techniques as they have been proven of some value to date.

- s




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20110822/3b91a013/attachment.html>