Two Mastodon posts argue that large language models do not possess intelligence but rather function as predictive filters. The authors contend that the "languages" LLMs appear to invent are merely amplified noise or feedback squeal, not genuine communication. They assert that humans also do not think in words, drawing an analogy to baseball player Sadaharu Oh, who achieved success by entering a 'mindless' state rather than thinking verbally. AI
IMPACT These posts challenge the notion of AI intelligence, suggesting models are sophisticated predictive filters rather than conscious entities.
RANK_REASON The cluster consists of opinion pieces from a social media platform discussing the nature of AI models.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →