Re Jiri Jelineks below 10/2/2007 1:21 AM post: Interesting links.
I just spent about a half hour skimming them. I must admit I havent spent enough time to get my head around how one would make a powerful AGI using Hadoop or MapReduce, although it clearly could be helpful for certain parts of the job, like getting information for use as a basis for induction or inference. If you have any more detailed thoughts on the subject I would be interested in hearing them. Also, the pricing listed on the first Amazon link seemed to indicate the charge per instance was 10 cents/hour, does that mean you could use 1K machines, with a total nominal 1.7 Topp/sec, 1.75 TByte RAM, and 160 TBytes of hard drive for one hour for just $100? Or, that you could use that much hardware for one year, 24/7, for just $876,000? Of course the interconnect, which is very important for AGI, is slow, 250Mbits/sec or just 1/40 that of a 10Tbit infiniband networked system, but still the pricing is impressive. It provides a valuable potential resource for at least some types of AGI research. I only read enough of the Google PDF about MapReduce to understand what MapReduce was and the major types of things it could be used for. What that reading made me think of was that it represented the type of sub-human computation that human-level AGIs will be able to execture and/or command and interface with millions of times faster than humans. If it had access to the Googleplex -- once a hierarchy of MapReduce software objects had been created it would be able to generate and specify task appropriate MapReduces more rapidly than we can generate NL sentences. This again emphasizes one of my key points, that an AGI with the hardware to be human-level at mental tasks we humans currently do much better than machines will be able to do and interface with things that computers already do much faster than humans thousands or millions of times faster than we can, meaning their overall capability for many tasks will be thousands of times ours. So human-level AGI wil easily be made superhuman for many tasks, greatly increasing their commercial value. Edward W. Porter Porter & Associates 24 String Bridge S12 Exeter, NH 03833 (617) 494-1722 Fax (617) 494-1822 [EMAIL PROTECTED] -----Original Message----- From: Jiri Jelinek [mailto:[EMAIL PROTECTED] Sent: Tuesday, October 02, 2007 1:21 AM To: [email protected] Subject: Re: [agi] The Future of Computing, According to Intel -- Massively multicore processors will enable smarter computers that can infer our activities Talking about processing power... A friend just sent me an email with links some of you may find interesting: ----------- cut -------------- Building or gaining access to computing resources with enough power to complete jobs on large data sets usually costs a lot of money. Amazon Web Services (AWS) allows users to create and run (nearly) unlimited numbers of virtual servers for a per-minute fee. That means you could potentially have a server farm of dozens of machines, all of which run for only a few minutes, and be charged only for the time that you need them to finish your computation job. First, look at AWS ec2: http://www.amazon.com/b/ref=sc_fe_l_2/104-0929857-7317547?ie=UTF8&node=201 590011&no=342430011&me=A36L942TSJ2AJA Then, look at what MapReduce is: http://labs.google.com/papers/mapreduce.html Then, look at the open-source MapReduce framework, Hadoop: http://lucene.apache.org/hadoop/ Then, look at how to use Hadoop on ec2: http://developer.amazonwebservices.com/connect/entry.jspa?externalID=873&c ategoryID=112 I believe that using AWS with Hadoop would be a very useful and cost-efficient way to develop and test powerful AGI algorithms. ----------- cut -------------- Regards, Jiri Jelinek On 10/1/07, Edward W. Porter <[EMAIL PROTECTED]> wrote: > > > > Check out the following article entitled: The Future of Computing, > According to Intel -- Massively multicore processors will enable > smarter computers that can infer our activities. > > http://www.technologyreview.com/printer_friendly_article.aspx?id=19432 > > Not only is the type of hardware needed for AGI coming fast, but one > of the world's biggest, fastest, smartest computer technology > companies is focusing on developing software using massively parallel > hardware that is directly related to AGI. > > It's all going to start happening very fast. The race is on. > > Edward W. Porter > Porter & Associates > 24 String Bridge S12 > Exeter, NH 03833 > (617) 494-1722 > Fax (617) 494-1822 > [EMAIL PROTECTED] ________________________________ > This list is sponsored by AGIRI: http://www.agiri.org/email To > unsubscribe or change your options, please go to: > http://v2.listbox.com/member/?&; ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?&; ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244&id_secret=48891836-8abbbd
