[ExI] Why create something if AI can do better in a few minutes?

Adrian Tymes atymes at gmail.com
Wed May 22 21:33:03 UTC 2024

On Wed, May 22, 2024 at 2:04 PM BillK via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> But any thoughtful person can
> conclude that it’s only a matter of years or months before AI performs
> perfectly.

And they would  be incorrect to do so.  That it takes decades to get from
"good initial capacity" to "perfect" can be evidenced by how far from
perfect autocorrect still is, and the length of time that good-enough
autocorrect has been in wide use.

> Why should a student study photography when before graduation, genAI
> tools will be able to produce perfect pictures on any subject in
> seconds?

So as to be able to use the tools better than those who don't know the
basics of what those tools are trying to emulate.  The comparison is not
"unaugmented (but trained or not trained) human" versus "hypothetical
perfect future genAI".  The comparison is either "untrained human with
basically current genAI" versus "trained human with basically current
genAI" or "untrained human with hypothetical perfect future genAI" versus
"trained human with hypothetical perfect future genAI".  The trained human
wins out in either situation.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20240522/3caeb722/attachment.htm>

More information about the extropy-chat mailing list