It's quite obvious to me that at one point humans will take AI so far
that they will end up inventing ourselves. That will be an amusing
day.

##
Advertising

On Mon, Sep 22, 2008 at 6:48 PM, <[EMAIL PROTECTED]> wrote:
>
> Let the algorithm that represents the brain of a typical new-born baby
> be denoted as B1.
>
> Now surely we can agree that the brain of a new-born baby does not
> have sophisticated Bayesian machinary built into it? Yes, there must
> be *some* intrinsic built-in reasoning structure, but everything we
> know suggests that the intrinsic reasoning mechanisms of the human
> brain must be quite weak and simple.
>
> Let the algorithm which represents the brain of the baby B1 which grew
> up into a 20-year old with a PhD in Bayesian math be denoted as B2.
>
> Now somehow, the algorithm B1 was able to 'optimze' its original
> reasoning mechanisms by a smooth transformation into B2. (assume there
> was 'brain surgery', no 'hand coding').
>
> The environment! you may shout. The baby got all its information from
> human culture (Reading math books, learning from math professors), you
> might try to argue, that's how B1 (baby) was able to transform into B2
> (PhD in Bayes)
>
> But this cant be correct. Since, humans existed long before Bayesian
> math was developed. Every single Bayesian technique had to be
> developed by a human in the past, without being told. So in theory,
> B1 could have grown into B2 entirely on its own, without being told
> anything by anyone about Bayesian math.
>
> The conclusion:
>
> *There exists a very simple algorithm which is only a very weak
> approximation to PhD Bayesian reasoning, which is perfectly capable of
> recursive self-improvement to the PhD level! No hand coding of
> advanced Bayesian math is needed.
>
> Or to simply rephrase:
>
> Humans could reason before they discovered Bayes.
>
>
>
>
> >
>
--
noon silky
http://www.themonkeynet.com/armada/
--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---