[ExI] LLM's cannot be concious

[BillK]

On Sat, 18 Mar 2023 at 12:59, spike jones via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
 > I think you are right gts.  I studied a bunch of experiments where
chatbots were talking to each other, looking for any indication that
there was any invention of ideas, the way two humans invent ideas when
they talk to each other.  The two-bot discussions never blossomed into
anything interesting that I have found, but rather devolved into the
one-line fluff that often comes with a forced discussion with a dull
unimaginative person at a party you don’t have the option of leaving.
>
> These chatbots are not yet ready to train each other and cause the singularity.  But there might be an upside to that: humans continue to survive as a species.  That’s a good thing methinks.  I like us.
>
> spike
> _______________________________________________


GPT-4 is surprisingly good at explaining jokes
OpenAI's latest language model can grasp visual jokes, too.
By Stephen Johnson      March 18, 2023

<https://www.freethink.com/robots-ai/gpt-4-jokes>
Quote:
But what really sets apart GPT-4 from ChatGPT is that it’s multimodal,
meaning it can “understand” text and images as inputs.
That includes visual jokes.
-----------
Explain why someone might laugh at this Anthony Jeselnik joke:
“Whenever I meet a pretty girl, the first thing I look for is
intelligence; because if she doesn’t have that, then she’s mine.”
-----------

And GPT-4 explained the humour very well. You might even think it
understood the joke.  ;)


BillK