On Wed, Feb 26, 2014 at 2:45 PM, Ben Goertzel <[email protected]> wrote:

> I don't want to argue too much about this because it's just a wild
> speculative direction and not too related to AGI, but I'll respond to your
> immediate points...
>


You said that inference speedup is the critical bottleneck of AGI, so I'm
trying to speed up inference, and I found that P!=NP is an obstacle along
the way.


In the formalism I was (perhaps too crudely) describing, each program would
> solve exactly one problem...
>
> Anyway I don't claim to have a proof of anything, so it's probably not
> worth debating intuitions in depth...
>


If you look at all programs of length X, most of them will terminate
quickly, and most of them will not solve the problem.

If you restrict to programs that can solve SAT, it may be true that
exponential ones outnumber polynomial ones, but that distribution seems
irrelevant to the question since P=?NP just asks for the existence of 1
such algorithm.

Anyway, I'm not doubting your prowess in math, but you seem to count
something and not count something..... in your philosophy of *life*.



-------------------------------------------
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

Reply via email to