On 3/19/26 11:02 PM, Stuart Stevenson wrote:
I should have included AI (Actual Incompetence) also.
Right on! There is a reason why people familiar with these trained transformer programs call them large language models (LLM) and do not use the terms "artificial" or "intelligence" in the description of these _programs_. Computers are always following a predetermined program and come always to the same conclusion from the same inputs. If they suddenly fail to follow that program, then you should throw away that computer and buy a new one that works.
These LLM programs are very large statistical systems and behave like huge Markov chains (https://en.wikipedia.org/wiki/Markov_chain). And while we're there, these Markov chains were already used as bullshit generators a long time ago (https://en.wikipedia.org/wiki/Parody_generator).
-- Greetings Bertho (disclaimers are disclaimed) _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
