[ExI] GPT-4 on its inability to solve the symbol grounding problem

Jason Resch jasonresch at gmail.com
Sun Apr 9 16:14:51 UTC 2023

On Sun, Apr 9, 2023, 10:54 AM Gordon Swobe <gordon.swobe at gmail.com> wrote:

> On Sat, Apr 8, 2023 at 10:13 PM Jason Resch via extropy-chat <
> extropy-chat at lists.extropy.org> wrote:
> > You keep making comments suggesting it's an absurd belief but you don't
> say why or how it is absurd. Please explain what about my view is absurd.
> Yes I do think it absurd to say that smart doorbells and cars have
> consciousness.

Smart doorbell systems able to detect the presence of a person in proximity
to a door and alter behavior accordingly have some primitive sensory
capacity. One cannot sense without consciousness.

An adaptive cruise control can detect the proximity of a car up ahead and
take action by slowing down. The car then has some primitive sensory
capacity, abd again I would ask you to explain how anything can sense
without sensing, feel without feeling, know without knowing, etc.

For that matter, I find it absurd to say that all living things are
> conscious,

I don't believe I said all living things are conscious. But I think those
that can detect and respond to changing variables in their environment have
at least some miniscule state of awareness.

Let's take it from another angle: do you think human consciousness suddenly
appeared out of nothing along the evolutionary tree, where one unconscious
parent had a child who was conscious, or do you think consciousness started
off very simple and changed and expanded gradually over the generations?
Which view makes more sense to you?

let alone any non-living inanimate objects. What about the lowly virus?
> Assuming even that we can call the virus alive (debatable, imo), it has
> nothing remotely resembling what we have as sense organs and nervous
> systems, and for that reason I find it a giant logical leap and thus absurd
> to say it or anything like it has consciousness.
> But let us say I am wrong and that your philosophical framework is
> accurate. Let us say there is nothing absurd about the claim that my
> automobile is conscious. In that case, as I was explaining about my
> supposedly conscious doorbell, the claim that “X is conscious” becomes
> trivial and uninteresting.

Who says consciousness has to be deeply mysterious and interesting?

Human consciousness might be interesting, but consciousness need not be.

The word “consciousness” loses most or all its meaning.

'Life' is still a useful and meaningful word, even though it's class of
objects is immense. Likewise with 'Physic object' or 'Mathematical object'.
Many things, perhaps even most things could fit into those categories.
Consciousness might be like that.

Many things if not all things in the world have this kind of consciousness
> and it becomes trivially true and uninteresting that GPT-4 is conscious.

Right. As I said before the interesting question is not if, but in what way
is GPT-4 consciousness.

If you have another way to understand and describe consciousness and what
triggers the dividing line between consciousness and unconsciousness I
invite you to share it.


>  "etics work is by something callesmd "cognitive unbinding": separate
> brain regions are unable to communicate with one another as signaling is
> confused, but each small region continues to operate independently. The
> mind is fragmented and each region becomes an isolated island unto itself.
>> We all know what the word means. Yes that does not mean his entire
>>> brain is dead, but he is unconscious.
>>> > For example, if smelling salts can still awaken him, then the part of
>>> his brain...
>>> When he awakens, he is no longer unconscious.
>> What part of his brain is aware enough to know to wake up fully after
>> administering smelling salts?
>>> > If you define consciousness in terms of human consciousness, then only
>>> humans are conscious, by definition.
>>> That is the only kind of consciousness with which we have any
>>> familiarity. I think it is reasonable to infer something similar in other
>>> people and in other higher mammals, as their anatomies and nervous systems
>>> and lives and behaviors are so similar to ours, but then things start to
>>> get sketchy as we go down the food chain. In the effort to justify the
>>> belief that even software
>> Using the phrase "only software" suggests to me you are not familiar with
>> the implications of the Church-Turing thesis. This thesis says software of
>> the right type can replicate the behaviors of *any* computable process. The
>> right software could mimic and replicate all the goings-on of the whole
>> milky way galaxy or observable universe.
>> It is a bit like saying a musician is so good that their music could not
>> be recorded on a CD, when we know CDs can capture the music of any possible
>> musician. Software is to behavior as CDs are to music. All you need is the
>> right CD (or software) to replicate any music (or behavior). This is the
>> magic of computers. We don't need to change or upgrade the hardware to
>> install new applications. One computer is enough to run any possible
>> software (that ever has been, or ever will be, written).
>>  can be conscious, people find themselves saying all sorts of silly
>>> things, for example that doorbells and cars are conscious. Their arguments
>>> lose by reductio ad absurdum except on ExI, where anything goes. :-)
>> Explain how it is a reductio ad absurdum.
>> I have shown the converse, denying their awareness, leads to a logical
>> contradiction.
>> What you call absurd is not a logical contradiction, just something
>> strange to your intuition. When choosing between something unfamiliar and
>> counter intuitive, vs. something logically inconsistent, go with the
>> unfamiliar and counter intuitive as the more likely to be true.
>> Jason
>>> _______________________________________________
>> 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/20230409/1d23537e/attachment.htm>

More information about the extropy-chat mailing list