XPost: comp.theory, sci.math, comp.ai.philosophy   
   From: polcott333@gmail.com   
      
   On 12/26/2025 6:59 AM, Richard Damon wrote:   
   > On 12/25/25 11:51 PM, olcott wrote:   
   >> On 12/25/2025 10:32 PM, Richard Damon wrote:   
   >>> On 12/25/25 10:37 PM, olcott wrote:   
   >>>> On 12/25/2025 9:17 PM, Richard Damon wrote:   
   >>>>> On 12/25/25 10:12 PM, olcott wrote:   
   >>>>>> Three different LLMs have been totally convinced   
   >>>>>> a total of 50 times, you just don't understand.   
   >>>>>   
   >>>>> LLM LIE, so are not reliable sources.   
   >>>>>   
   >>>>   
   >>>> *Anyone that disagrees with this is not telling the truth*   
   >>>> "Any result that cannot be derived as a pure function   
   >>>>   of finite strings is uncomputable."   
   >>>>   
   >>>   
   >>> But Halting *IS* a "pure function of finite strings"   
   >>>   
   >>> And it is uncomputable   
   >>>   
   >>   
   >> Not exactly. Usually ⟨M⟩ simulated by H == UTM(⟨M⟩)   
   >> Sometimes ⟨M⟩ simulated by H != UTM(⟨M⟩)   
   >   
   > Only if H doesn't CORRECTLY simulate (M).   
   >   
      
   Correctly simulated is defined by the semantics   
   of C applied to the finite string input for   
   the N steps until H sees the repeating pattern.   
      
   I know that you are not stupid. I know that you   
   can pay attention to all of those words.   
      
   When I state a true fact and you understand that   
   it is a true fact yet deny it anyway what could   
   explain this denial?   
      
   > All you are doing is proving you don't understand the meaning of   
   > "Correct", which is part of the source of your pathology that makes you   
   > a pathological lair.   
   >   
   > Please try to explain, preferably with a concrete example, how H can   
   > CORRECTLY simulate a step in (M) that CORRECTLY describes the algorithm   
   > of M and get a result different from the actual step done by M?   
   >   
   > Remember that (M) is supposed to be a complete description fully showing   
   > ALL the steps in M with enough detail to recreate it, and does not refer   
   > to anything not in that description, thus for P, it includes an encoding   
   > of the actual algorithm of H, and not just a "reference" to say do what   
   > H does.   
   >   
   >>   
   >>> Maybe you don't know what those words mean.   
   >>>   
   >>>> When the LLMs   
   >>>> (a) apply correct semantic entailment to   
   >>>> (b) standard definitions   
   >>>> any conclusions so derived are infallible by definition.   
   >>>   
   >>> How do they do that? I guess you don't know how a LLM works.   
   >>   
   >> https://iep.utm.edu/val-snd/   
   >   
   >   
   > And since LLMs don't follow those rules of logic, or even work just from   
   > correct statements, their "reasoning" is neither "valid" or "sound".   
   >   
   > If you actually look at what LLMs are, they are effectively just large   
   > Markof chains built to generate reasonable sounding continuations from   
   > your prompt, and their data source was everything said in the training   
   > corpus, both correct and erroneus statements, trained by the criteria of   
   > "does it sound reasonable", with explicit instructions NOT to judge on   
   > factual correctness of non-obvious matters.   
   >   
   > And, since you have just followed their lead, neither does your logic.   
   >   
   > In fact, since you never learned the actual meaning of the words you try   
   > to use, you never had that correct basis to work from, so your own logic   
   > has never been valid or sound when talking of the field.   
      
      
   --   
   Copyright 2025 Olcott

   
      
   My 28 year goal has been to make 
   
   "true on the basis of meaning expressed in language"
   
   reliably computable.

   
      
   This required establishing a new foundation
   
      
   --- SoupGate-DOS v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)