Mark Bernstein
@eastgate.bsky.social
Designer of Tinderbox, information gardener. Research interests include literary hypertext, internet villainy, and intellectual history of computing. https://markBernstein.org/ https://www.eastgate.com/Tinderbox/
With respect, I disagree. Elementary textbooks don’t always exist; post-PhD, they often don’t. Google is lousy.
Training LLMs is resource-intensive, but using them is not.
Training LLMs is resource-intensive, but using them is not.
December 24, 2025 at 12:07 AM
With respect, I disagree. Elementary textbooks don’t always exist; post-PhD, they often don’t. Google is lousy.
Training LLMs is resource-intensive, but using them is not.
Training LLMs is resource-intensive, but using them is not.
Suppose a writer uses an AI to refresh their memory of (say) the second law of thermodynamics while working on a story. After this, they find they understand the subject more clearly.
If the story AI-generated? (John Varley IIRC had a very funny story about his commitment to unprocessed words)
If the story AI-generated? (John Varley IIRC had a very funny story about his commitment to unprocessed words)
December 17, 2025 at 10:04 PM
Suppose a writer uses an AI to refresh their memory of (say) the second law of thermodynamics while working on a story. After this, they find they understand the subject more clearly.
If the story AI-generated? (John Varley IIRC had a very funny story about his commitment to unprocessed words)
If the story AI-generated? (John Varley IIRC had a very funny story about his commitment to unprocessed words)