I think the negative reactions to the original post are so strong because
there is no demonstrated value. Most of us can write a quick description
of this listserv, MARC, and open source. And if we really needed word
counts we can do it with a 10 line nltk script. If you're using this kind
of approach on a corpus you don't have expertise with, you need an expert
to verify it. The verification tax with LLM use is widely discussed.
Glossing over the massive environmental, socio-economic, intellectual,
legal and moral problems it creates or exacerbates as if they couldn't
possibly contribute to the hesitation is extremely concerning at best.
(And say goodbye to serendipitous discovery.)
As to the original question of why people are hesitant to use these tools,
the Leiden Declaration is a relevant read. It is from a group of
mathematicians, but many of their expressed concerns, (convincing but
incorrect output, undermining the traditional systems of attribution,
incentivizing the use of AI for its own sake, the increasing role corporate
decision-making is playing in our discipline,) easily translate to
Librarianship.
https://leidendeclaration.ai/
David
On Thu, Jul 9, 2026 at 1:39 PM Tim Spalding <
000001d841ece815-dmarc-request_at_lists.clir.org> wrote:
> *"But coming in to assert that 'LLM technology is especially useful' as if
> we don't hear it from every angle, every single day, is unnecessary and
> demeaning."*
>
> Eric, who has been involved with Code4Lib from its start in 2003, did not
> hop in to make a statement and leave. He did the things you asked for. He
> described his project, named the models he used. Laid out his workflow. He
> gave costs, details of the data, quoted his actual outputs, added a
> footnote. You asked that people "unveil new tools, talk about processes"
> and he did exactly that. Very few posts here go into the detail he did.
>
> After talking about what he did, how he did it, and how it fared, yes, he
> asked a question on many of our minds: Why are libraries so often hostile
> to LLMs? This "irritating time waster" shall "broadcast his insecurity" by
> suggesting that the vociferousness of the replies demonstrate the relevance
> of his question.
>
> Tim Spalding
> Founder, LibraryThing
>
--
R. David Beales - rdbeales_at_gmail.com - 732-299-0390
Earth, Sol System, Orion-Cygnus Arm of the Milky Way Galaxy, Laniakea
Supercluster
Received on Thu Jul 09 2026 - 13:43:59 EDT