This makes no sense LLMs are stochastic – non deterministic You might get very-similar-to-the-original-prompt results But never *exact* original prompts And *definitely* not any other constraints loaded into the inference engine outside the initial prompt https://t.co/yNS2EadFdO

This makes no sense

LLMs are stochastic – non deterministic

You might get very-similar-to-the-original-prompt results

But never *exact* original prompts

And *definitely* not any other constraints loaded into the inference engine outside the initial prompt https://t.co/yNS2EadFdO

— Mɐɹɹǝu Wʎǝɹs (@warrenmyers)
Aug 11, 2026

from Twitter https://twitter.com/warrenmyers

August 11, 2026 at 12:54AM
via IFTTT