Non-Constructive Logic: Any AI method that approximates AIXI will lack the human capability to reason about non-computable entities.
On Sat, Oct 18, 2008 at 8:20 PM, Steve Richfield <[EMAIL PROTECTED]> wrote: > Everyone, > > Ben has made a really wonderful offer here: > > On 10/18/08, Ben Goertzel <[EMAIL PROTECTED]> wrote: >> >> I will make one more attempt to address your objections. **If and only >> if** you will be so kind as to summarize them in a compact form in a single >> email. If you give me a numbered list of your objections against my >> approach to AGI and other similar approaches, in which each objection is >> summarized in a few dozen words at most, then I will respond by summarizing >> my reaction to each of your objections. > > > I am not the expert in many/most of these and quickly bow to the talents of > other here. Please help me assemble this list by responding to this posting > with short summary descriptions of your objections (as I think everyone > already knows the details), and please AVOID cross-discussions on this > thread. Then, I can combine these items into the organized list that Ben has > offered to respond to, and which others here may wish to respond to. > > To get this ball started, I have posted below a single item, more to > illustrate formatting than to advance a particular viewpoint. Please DO NOT > COMMENT on this item, at least not on this thread, unless of course, you can > state this more clearly or based on better facts. > > Sample: > > Algorithmic Difficulty: Evolution favors the simplest workable solutions to > challenges. There are ~200 different types of neurons in the human brain, > which would seem to be an indication of the number of different types of > manipulations needed to "think". The absence of any one of these would seem > to delegate a system to sub-human performance. This would seem to preclude > any sort of simplistic algorithm to exceed what neurons now do. > > Of course, this list will only be a starting point, as the responses will > doubtless start other discussions, uncover other feasibility challenges, > etc. Ultimately, valid challenges will constrain future approaches, and > invalid challenges will be uncovered as such. At least with this list and > Ben's responses, it will finally be possible for other participants to > assess for themselves whether various approaches are likely to work. With > luck, this might even draw some lurkers here into the thick of coding to > deftly avoid these challenges and finally succeed in making a working AGI. > > Thanks for your help in pulling this together. > > Steve Richfield > > ________________________________ > agi | Archives | Modify Your Subscription ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=117534816-b15a34 Powered by Listbox: http://www.listbox.com
