Oh and in terms of buying the latest hardware, my suggestion is to save your money. I spent like $60 on a hardrive for a desktop computer, back when I wanted to have 2 desktops as kinda like a backup server, though now I just have online hosting for my website, anyways so I decided to sell my desktop, but I only managed to get $50 for the whole thing, even though it was amd64 with 3Ghz.
If you genuinely need the computing power, it may be cheaper to use a Beowulf cluster out of cheap used machines, good for practicing parallel computation. I got this laptop for free, since my dad decided to upgrade his to a newer model. You may remember Kasparov vs Blue Gene, which basically brute forced it's way to winning. However years later, there were more efficient chess programs such as Deep Fritz, that could also defeat the world's chess masters with a mere 3Ghz and 512MB RAM. Point being, I think efficiency is important in the long term. Especially considering there are many relatively slow embedded processors. neuro-nets are a brute-force approach, having algorithms that are task-specific is efficient. On Mon, Jan 28, 2013 at 12:34 AM, Logan Streondj <[email protected]> wrote: > hmmm, so it recognizes number 0-9 and can do a little bit of simple > arithmetic, as well as write the numbers. > I guess that's impressive for a neural net. Though with optical character > recognition, and motor controls for writing characters can accomplish the > same thing, likely with a lot less computer resources. > > Of course, considering it takes several hours to simulate a second, it's > not going to be a threat any times soon. > Though in general, I don't consider people making models of birds much > closer to building a flying machine. > > I used to use a desktop with perhaps as much as 4GB ram, and 3 or 4 Ghz, > and still have it available. but it's kinda noisy, and I don't really need > it, so I just use a laptop with 2GB RAM and 2Ghz, with Lubuntu, it runs > lightning fast. > > Since my AGI project is coded in assembly, I don't really need nearly this > much computer power for the R&D. Though likely it'll eventually be useful, > when at the operating system stage and above. > > > > On Fri, Jan 18, 2013 at 3:01 PM, just camel <[email protected]> wrote: > >> Chris Eliasmith on Singularity 1 on 1: We Have Not Yet Learned What The >> Brain Has To Teach Us! >> >> http://www.youtube.com/watch?**v=ZjDp8EnfDvo<http://www.youtube.com/watch?v=ZjDp8EnfDvo> >> >> >> On 12/02/2012 04:15 PM, Alan Grimes wrote: >> >>> 'ello. >>> >>> I'm going to have to issue a singularity alert for a package called >>> Spaun. >>> >>> >> >> ------------------------------**------------- >> AGI >> Archives: >> https://www.listbox.com/**member/archive/303/=now<https://www.listbox.com/member/archive/303/=now> >> RSS Feed: https://www.listbox.com/**member/archive/rss/303/** >> 5037279-a88c7a6d<https://www.listbox.com/member/archive/rss/303/5037279-a88c7a6d> >> Modify Your Subscription: https://www.listbox.com/** >> member/?&id_**secret=5037279-8beb0005<https://www.listbox.com/member/?&;> >> Powered by Listbox: http://www.listbox.com >> > > ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
