Le 08-mai-08, à 06:48, Russell Standish a écrit :
> > On Wed, May 07, 2008 at 03:40:00PM +0200, Bruno Marchal wrote: >> >> even the non computable reals. And, given that from a first person >> point of view we cannot be aware of the infinitely many delays, we >> have to take into account, to eliminate white rabbits (or to refute >> comp) all the possible computations at once, including those >> dovetailing on the reals or on any non enumerable structure. I have >> discussed this with Schmidhuber on the list some time ago. The non >> enumerability of the reals cannot prevent the UD to dovetail on all >> the >> reals. From a third person point of view, everything is and remains >> enumerable, but from the first person point of view, the subjective >> indeterminacy has as domain something vastly bigger. >> > > Actually, it just occurred to me that awareness of delays (step 5 of > the UDA) is where you and Schmidhuber part company. His speed-prior > discussion is all about assigning lower measure to computations that > take longer (eg have delays). Really? I think Schmidhuber's speed prior concerns intrinsic computational speed (like when comparing different algorithm). If not, nothing computational would remains in Schmidhuber's approach. Step 5 (or even 4) is where I part company with Nozick's closer continuer theory of identity, at least if the "closer" relation is based on the a priori physical. Anyway, I have a problem with speed prior, due to a theorem of Blum (similar to a theorem by Godel) which says that universal machine can always been sped up for sufficiently great inputs. Not that some relative speed cannot play some important role, but this has to be justified, not invoked. Then, I have a deeper problem with the use of prior in general. That needs ASSA, and, also, it lead to the question "why such prior" ? And then, as I have often argue, there is that lack of distinction between 1-person views and 3-person views like if the mind-body problem does not exist. Here already Everett and Tegmark are a bit more cautious, imo. (QM forces us to be cautious there, comp too, but it is less evident apparently). > > Going back to the ASSA vs RSSA divide, Schmidhuber's picture requires > the ASSA to work. That is much more probable, yes, I think so too. Best, Bruno http://iridia.ulb.ac.be/~marchal/ --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Everything List" group. To post to this group, send email to [EMAIL PROTECTED] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list?hl=en -~----------~----~----~----~------~----~------~--~---
