[Hplusroadmap] [SL4] Re: Paper: Artificial Intelligence will Kill our Grandchildren

[Dan Bolser]

2008/6/25 Bryan Bishop <kanzure at gmail.com>:
> On Wednesday 25 June 2008, Dan Bolser wrote:
>> 2008/6/25 Bryan Bishop <kanzure at gmail.com>:
>> > On Friday 13 June 2008, Anthony Berglas wrote:
>> >> http://berglas.org/Articles/AIKillGrandchildren/AIKillGrandchildre
>> >>n.h
>>
>> [snip]
>>
>> >> As worms have evolved into apes, and apes to man, the evolution of
>> >> man to an AI is just a natural process and something that could be
>> >> celebrated rather than avoided.  Certainly it would probably only
>> >> be a matter of a few centuries before modern man destroys the
>> >> earth, whereas an artificial intelligence may be able to survive
>> >> for millenia.
>> >
>> > Holy shit man, you don't understand evolution. Particularly the
>> > part about programming and "man evolving into AI". What most people
>> > consider AI to be is something about programming. I truly doubt
>> > that a directed intelligent process, like programming, is a
>> > naturally occuring evolutionary process.
>>
>> What? In that case please define what is naturally occurring and what
>> is 'artificial'. Or at least outline how you make the distinction?
>
> You talking to me, Dan? :) I just figure that saying that 'the evolution
> of man into ai' illustrates a lack of understanding of evolution.

Perhaps, but its the concept of evolution that interests me here most.


> Please click here to evolve.

Done. ;-)


>> In my opinion, all of human activity is a part of the ongoing
>> biological process that was initiated on the planet approximately 4bn
>
> No argument there, we're still ticking.
>
>> years ago. The planet was the result of an ongoing biological process
>> that was initiated about 10bn years ago. Saying something like
>> 'evolution is what animals do' is reducing the scope of your
>> imagination. Don't cars evolve? Hasn't wheat evolved during the
>
> I'm not sure if it's valid to say that something evolves when it's not
> self-replicating. You're just talking about the selection of vehicles
> and the design elements that go into the engineering of the cars,
> rather than the cars mutating on their own (i.e., their own errors of
> polymerase, their own self-replicational mechanisms ...).

Yeah, that is the crux of my argument. What is self replication
really? Can bacteria self rep. in space? Of course not, they need the
substrate of their environment to perform the trick of self rep. They
are a nice little encapsulated bundle of self rep., carrying all the
information and tools they need to do the trick within a membrane...
within a given environmental context. Now... Can a virus self rep?
Sure, given the right host, a virus can make copies of itself
ad-nausiem (pun intended ;-). However, somehow, the viral self rep. is
a little bit of a stretch... they don't quite 'encapsulate' self rep.
like a bacteria does. They don't carry all of the tools they need
neatly bound up in a membrane. They require to hijack other 'host'
tools to get the job done. Can a bacteria self rep. in space?

Did you know that your mitochondria split and divide independently of
cellular replication? Their 'life cycle' is highly regulated within
the whole cellular environment, but they self rep. within this
context. Some mitochondrial diseases are attributed to mutations that
come to selectively dominate the 'population' of the mitochondria in
your cells. (Unlike somatic mutations that lead to genetic disease,
mitochondrial mutations and diseases have a 'dosage effect', because
some fraction of your mitochondria can be normal / mutant).

So what is self replication really? Would wheat be so successful on
planet earth if it wasn't so tasty to humans? Would it out compete the
weeds that we fend off with herbicides? Would a Chihuahua exist if it
were not for the human cultural environment that supports it?

So, with thinking like this, it is not hard to make a jump and say
that cars are biological objects. No, they do not neatly encapsulate
the machinery needed to self rep. within some 'membrane' or other.
They can only self rep. in the wider context of human society. Just
like a bacteria needs something other than the vacuum of space.
Remember, as you pointed out previously, evolution does not happen
'intentionally' - it just happens.

In this way all self. rep. is essentially viral. I see the universe as
a bootstrapping virus, with the bio-layer as the most interesting
(information processing intensive) substrate.

After all this, however, I have to say that I don't think self rep. is
the key to evolution. I think its about self stability, but that is
something else again.


Bacteria self rep. within a membrane. Memes self rep within our
brains. Cars also self rep., based on memes and human intellegence,
but without such a clearly defined 'membrane'. I don't see any
'fundamental' difference - of course there are differences that arise
by convention.


>> history of human existence? Isn't GM just the latest phase of that
>> ongoing co-evolutionary process? Is it 'artificial' that dogs
>
> Wheat/GM is bio.
>
>> recognize human emotion better than chimps? Aren't radios self
>
> Dogs are bio. I think a common argument is that domesticated animals are
> somewhat selected too, and it's the reason why your kitten will talk to
> you and usually not run away from home and such.
>
>> reproducing concepts / entities in the context of human society?
>
> That's memes, not radios. ;-) Neural selectionism won a Nobel Prize.

Its radios too. What is the difference? Aren't the patterns in your
brain as physically real as radios?


>> This thread discusses Darwin and evolution as if they were the same
>> thing - I don't think this is the case. Just as the movement of the
>> planets in the sky was described by the ancients, then Euller, then
>> Kepler, then Newton, then Einstein (and a lot of others that I have
>> skipped over), the observable phenomenon of evolution can have more
>> than one explaination / interpretation. Given these broader contexts
>> within which to understand evolution (neo-darwinism for example),
>> phenomenon like 'cultural evolution' are very easy to understand as
>> an extension of the core idea of 'evolution' in general.
>
> Certainly.
>
>> Anyway, sorry for butting in on a couple of threads that I haven't
>> been following - prolly you discussed all the points that I made
>> above ad-nausiem elsewhere. I'll look forward to you pointing out all
>> the 100s of documents I should have read before opening my mouth to
>> type.
>
> Opening my mouth to type would be awesome.

;-)

BTW, have you seen season 2 of Ghost in the shell? Its nothing
conceptually new to this list, but a lot of fun.


Dan.


> - Bryan
> ________________________________________
> http://heybryan.org/
> _______________________________________________
> Hplusroadmap mailing list
> Hplusroadmap at heybryan.org
> http://heybryan.org/cgi-bin/mailman/listinfo/hplusroadmap
>



-- 
hello