On 2025-09-22 at 09:44:37 UTC-0400 (Mon, 22 Sep 2025 09:44:37 -0400)
Rich Bowen <[email protected]>
is rumored to have said:

> Bill’s reaction is, I presume, rooted in the hallucinations that various AI 
> assistants have all the time, and that’s a valid concern.

That's significant, but I'm more concerned with the issues of IP theft and 
covert biases in training input. An obviously wrong "hallucination" is nowhere 
near as bad as output that seems plausible but is wrong in subtle ways.

The proposed generic name and acronym for the type of tool is also a problem, 
as it seems like an intentional collision with the "Model Context Protocol."

And overall, I still see no improvement to be gained by creating a black box to 
hold documentation over simply having PMCs & contributors work openly and 
document their concepts and changes. The problem description implies chronic 
failure to spread, share, and record  knowledge on the part of major 
contributors. That problem should not be present in any ASF project that has 
been through the Incubator, and if it develops later can be resolved by 
documentation. Allowing a convoluted pile of machine-generated linear algebra 
to be the source of truth about a project is a fundamentally bad idea, if only 
because it acts as the inscrutable dead hand of former contributors who may no 
longer care about the project or who merely may have never considered modern 
circumstances which necessitate change. ASF projects "belong to" the current 
users and developers, not the ghosts of original authors.


-- 
Bill Cole

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to