logoalt Hacker News

alex43578today at 8:41 AM0 repliesview on HN

LLMs are already way too prepared for "Yes and..." improv, given GPT's ridiculous need to click-bait the end of every conversation.