ZacCurryMar 6
AI Ideas:
Dual-Stack AI Tokenization system in which kanji radicals, word components, grammatical declension and conjugation, (in the case of audible speech, models for stress/timing/accent/emphasis), and particles/prepositions are processed in their own “flow” order,
Show threadZacCurryMar 6
followed by a comparative etymological analysis to determine how words act upon each other, followed by that traditionally-used tokenization in order to assist in cross-idiomatic processing and for simpler & enhanced dialectal/stylistic preprocessing & recognition; more-easily serialize tokenization of semantic intent; occasional intent clarification inquiries to personalize for individual end-user
Show threadZacCurry

Basically, non-blackbox interpretive AI seems a lot more useful than generative AI from a “let’s not destroy the world” standpoint

#AI #generativeAI #interpretiveAI #tokenization #blackbox #nonblackbox #savesocial #SaveTheUS #kanji #grammar #speech #speechprocessing #languages #language #LLM #dialectrecognition #SaveTheWorld #mediapreservation

Mar 6 at 7:26pmWeb