The threats are BS, but fyi models have repeatedly shown capability to produce n... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

adastra22 15 days ago | parent | context | favorite | on: Our position on open-weights models

The threats are BS, but fyi models have repeatedly shown capability to produce novel things that are NOT in their training set.

dolebirchwood 15 days ago [–]

Sure, based on predicate knowledge.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact