[ExI] fun outsider's view on ai

Wed May 11 01:53:22 UTC 2016

I can strongly recommend this book:

http://www.amazon.com/Superintelligence-Dangers-Strategies-Nick-Bostrom/dp/0198739834/ref=sr_1_1?ie=UTF8&qid=1462931554&sr=8-1&keywords=super+intelligence

Jason

On Mon, May 9, 2016 at 3:46 PM, William Flynn Wallace <foozler83 at gmail.com>
wrote:

> I'm neither a Singularitarian nor an AItheist. I think human-level AI is
> inevitable, if President Trump doesn't manage to wipe out the human race
> first :-). But I don't buy the notion that super intelligence is akin to a
> superpower, and don't think it's necessary for an AI to have consciousness,
> human-like emotions, or the ability to set its own goals, and without those
> there is no need to fear them.  dave
>
> If you want an AI to be superintelligent, why reference the neuron,
> Spike?  Human brains are so fallible it's just silly.  A person super
> intelligent about one thing is totally at a loss about many other things.
> I think brains must be still evolving, because as they are, they are
> cobbled together among available equipment and have functioned well enough
> to get us to the present.  You don't have to be a psychologist to see the
> irrationality, the emotional involvement, the selfishness, of the output of
> human brains.  There are many functions of brains that we can do well
> without entirely.  Start with all the cognitive errors we already know
> about.
>
> OK, so what else can we do?  Every decision we make is wrapped up in
> emotions. That alone does not make them wrong or irrational, but often they
> are. Take them out and see what we get.  Of course they are already out of
> the AIs we have now.  So here is the question:  do we really want an AI to
> function like a human brain?  I say no.  We are looking for something
> better, right?
>
> Since by definition we are not yet posthumans, how would we even know that
> an AI decision was super intelligent?  I don't know enough about computer
> simulations to criticize them, but sooner or later we have to put an AI
> decision to experimental tests in the real world not knowing what will
> happen.
>
> In any case, I don't think that there is any magic in the neuron.  It's in
> the connections.  And let's not forget about the role of glial cells, about
> which we are just barely aware.  (see The Other Brain by Douglas Fields)
>  Oh yeah, and the role of the gut microbiome - also just barely aware of
> its functions.  Not even to mention all the endocrine glands and their
> impact on brain functions.  Raising and lowering hormones has profound
> effects on functioning of the brain.  Ditto food,  sunspots (?), humidity
> and temperature, chemicals in the dust we breathe, pheromones, and drugs
> (I take over 20 pills of various sorts,  Who or what could figure out the
> results of that?)  All told, an incredible number of variables, some of
> which we may not know about at present, all interacting with one another,
> our learning, and our genes.
>
> All told, we are many decades away from a good grasp of the brain, maybe
> 100 years.  A super smart AI will likely not function at all like a human
> brain.  No reason it should.  (boy am I going to get flak on this one)
>
> bill w
>
>
>
> On Mon, May 9, 2016 at 11:27 AM, Dave Sill <sparge at gmail.com> wrote:
>
>> On Mon, May 9, 2016 at 10:44 AM, spike <spike66 at att.net> wrote:
>>
>>>
>>>
>>> Nothing particularly profound or insightful in this AI article, but it
>>> is good clean fun:
>>>
>>>
>>>
>>>
>>> https://aeon.co/essays/true-ai-is-both-logically-possible-and-utterly-implausible?utm_source=Aeon+Newsletter&utm_campaign=6469cf0d50-Daily_Newsletter_9_May_20165_9_2016&utm_medium=email&utm_term=0_411a82e59d-6469cf0d50-68957125
>>>
>>
>> Yeah, not bad. Mostly on the mark, IMO, but he says a few things that are
>> just not rational.
>>
>> He reminds me a little of Roger Penrose’s take on the subject from a long
>>> time ago: he introduces two schools of thought, pokes fun at both while
>>> offering little or no evidence or support, then reveals he is pretty much a
>>> follower of one of the two: the Church of AI-theists.
>>>
>>
>> To be fair, he says both camps are wrong and the truth is probably
>> somewhere in between. And I agree.
>>
>>
>>> There are plenty of AI-theists, but nowhere have I ever seen a really
>>> good argument for why we can never simulate a neuron and a dendrite and
>>> synapses.  Once we understand them well enough, we can write a sim of one.
>>> We already have sims of complicated systems, such as aircraft, nuclear
>>> plants and such.  So why not a brain cell?   And if so, why not two, and
>>> why not a connectome and why can we not simulate a brain?  I have been
>>> pondering that question for over 2 decades and have still never found a
>>> good reason.  That puts me in Floridi-dismissed Church of the
>>> Singularitarian.
>>>
>>
>> Yeah, his "True AI is not logically impossible, but it is utterly
>> implausible" doesn't seem to be based on reality.
>>
>> I'm neither a Singularitarian nor an AItheist. I think human-level AI is
>> inevitable, if President Trump doesn't manage to wipe out the human race
>> first :-). But I don't buy the notion that super intelligence is akin to a
>> superpower, and don't think it's necessary for an AI to have consciousness,
>> human-like emotions, or the ability to set its own goals, and without those
>> there is no need to fear them.
>>
>> -Dave
>>
>> _______________________________________________
>> extropy-chat mailing list
>> extropy-chat at lists.extropy.org
>> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>>
>>
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20160510/be337d2d/attachment.html>