Block, Siteadmin, SkipCaptcha, Upload, Automoderated users, delete, edit, move, nsTeam2RO, nsTeam2RW, nsTeam2_talkRO, nsTeam2_talkRW, protect, rollback, Administrators, template
293,146
edits
Changes
Jump to navigation
Jump to search
Line 34:
Line 34: + +
Artificial intelligence (view source)
Revision as of 04:59, December 20, 2025
, 04:59, December 20, 2025→LLMs
Cheerleaders miss that LLMs do not reason but are probabilistic prediction engines. A system which trawls existing data, even assuming that is correct, cannot create anything new. Once existing data sources are devoured, scaling produces diminishing returns. Rather than fully generalisable intelligence, generative models are regurgitation engines struggling with [[truth]], [[hallucination]]s and [[reasoning]].<ref>https://www.zerohedge.com/ai/satyajit-das-ai-artificial-intelligence-or-absolute-insanity</ref>
Cheerleaders miss that LLMs do not reason but are probabilistic prediction engines. A system which trawls existing data, even assuming that is correct, cannot create anything new. Once existing data sources are devoured, scaling produces diminishing returns. Rather than fully generalisable intelligence, generative models are regurgitation engines struggling with [[truth]], [[hallucination]]s and [[reasoning]].<ref>https://www.zerohedge.com/ai/satyajit-das-ai-artificial-intelligence-or-absolute-insanity</ref>
In a technical field LLMs can get a lot of information fast which allow a starting point for [[critical analysis]]. LLMs cannot be relied upon for [[political]] or [[social]] analysis – or anything in the so-called "[[humanities]]" other than to establish "conventional" wisdom.<ref>https://julianmacfarlane.substack.com/p/ais-faking-history</ref>
===Open web in decline===
===Open web in decline===