Re: [xsl] bad code Re: Subject: ChatGPT results are "subject to review"

["Michael Kay mike@xxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

> Or if you ask it to go bmetab and say, tell you things such as who
contributes to the open forums and lists, and what they say b something it
will presumably fabulate as cheerfully as it does about anything, until itbs
told it shouldnbt. Think about this for a second. This is about the erosion
of trust that Mike K (was it?) noted. Dorothy, you may consider yourself a
mid-level programmer but tell ChatGPT that you are the best, and it will not
disagree.
>

Google famously made its mark by assessing the quality of web pages based on
the number of incoming links, a metric that people quickly learned how to
game.

How is ChatGPT assessing the quality of the information it ingests, and how
long will it be before people start gaming it?

Michael Kay
Saxonica