[ExI] Language models are like mirrors

[Gordon Swobe]

On Sun, Apr 2, 2023 at 6:32 PM spike jones via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

What we are seeing with ChatGPT appears to me to be Eliza plus fifty years,
> with access to the internet for a database.  Its grammar is perfect,
> vocabulary is erudite, it sure sounds like it knows what it is doing.
> Until… you ask it about stuff you really really know a lot about.  Then…
> the spell is broken.  It appears to me that ChatGPT is really good at
> looking up stuff on the internet, then forming an answer with perfect
> grammar and advanced vocabulary.
>

Yes. As I keep saying (following others smarter than me who have said the
same) these language models are sophists, in the derogatory sense of that
word. Some are more sophistic than others, (e.g., LaMDA more so than
ChatGPT4, I would say) but they're all sophists. They are gifted in the
ways of language and can be very persuasive, but care little to nothing
about the truth value of anything they say.  And this is to be expected as
they quite literally do not know the meanings of the words they generate.

-gts
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230402/e2d1657a/attachment.htm>