XPost: comp.theory, sci.logic
From: polcott333@gmail.com
On 7/20/2025 6:13 AM, Richard Damon wrote:
> On 7/19/25 11:20 PM, olcott wrote:
>> On 7/19/2025 9:12 PM, Richard Damon wrote:
>>> On 7/19/25 5:18 PM, olcott wrote:
>>>> On 7/19/2025 4:00 PM, Alan Mackenzie wrote:
>>>>> Mike Terry wrote:
>>>>>
>>>>> [ .... ]
>>>>>
>>>>>> ps. learn to post more respectfully.
>>>>>
>>>>> You've hit the nail on the head, there. Peter Olcott doesn't show
>>>>> respect here for anybody. Because of this he isn't shown any respect
>>>>> back - he hasn't earned any. I don't think he understands the concept
>>>>> of respect any more than he understands the concept of truth.
>>>>>
>>>>> If he were to show repect, he'd repect knowledge, truth, and learning,
>>>>> and strive to acquire these qualities. Instead he displays
>>>>> contempt for
>>>>> them. This is a large part of what makes him a crank. It is
>>>>> a large part of what makes it such a waste of time trying to correct
>>>>> him, something that you've sensibly given up.
>>>>>
>>>>
>>>> Now that chat bots have proven that they understand
>>>> what I am saying I can rephrase my words to be more
>>>> clear.
>>>>
>>>
>>> They have done no such thing, because they can't
>>>
>>> Since yoiu feed them lies, all you have done is shown that you think
>>> lies are valid logic.
>>>
>>>> I have been rude because I cannot interpret the
>>>> rebuttal to this statement as anything besides
>>>> a despicable lie for the sole purpose of sadistic
>>>> pleasure of gaslighting:
>>>
>>> Because you are just too stupid.
>>>
>>> How is the "pattern" that HHH detects a non-halting pattern, when
>>> non- halting is DEFINED by the behavior of the directly executed
>>> machine, and the pattern you are thinking of exists in the execution
>>> of the DDD that halts because it was built on the same HHH you claim
>>> is correct to return 0,
>>>
>>> Thus, your claim *IS* just a lie, and you shows your ignorance by
>>> saying you can't undetstand how it is one.
>>>
>>>>
>>>>
>>>> typedef void (*ptr)();
>>>> int HHH(ptr P);
>>>>
>>>> void DDD()
>>>> {
>>>> HHH(DDD);
>>>> return;
>>>> }
>>>>
>>>> int main()
>>>> {
>>>> HHH(DDD);
>>>> DDD();
>>>> }
>>>>
>>>> Termination Analyzer HHH simulates its input until
>>>> it detects a non-terminating behavior pattern. When
>>>> HHH detects such a pattern it aborts its simulation
>>>> and returns 0.
>>>>
>>>>
>>>> Every chatbot figures out on its own that HHH
>>>> correctly rejects DDD as non-terminating because
>>>> the input to HHH(DDD) specifies recursive simulation.
>>>>
>>>
>>> BECAUSE YOU LIE TO THEM, and a prime training parameter is to give an
>>> answer the user is apt to like, and thus will tend to just accept
>>> lies and errors provided.
>>>
>>
>> I only defined the hypothetical possibility of a simulating
>> termination analyzer. This cannot possibly be a lie. They
>> figured out all the rest on their own.
>
> No, you stated that it DOES something that it doesn't.
>
Unlike a halt decider that must be correct
on every input a simulating termination analyzer
only needs be correct on at least one input.
void Infinite_Recursion()
{
Infinite_Recursion();
}
void Infinite_Loop()
{
HERE: goto HERE;
return;
}
void Infinite_Loop2()
{
L1: goto L3;
L2: goto L1;
L3: goto L2;
}
HHH correctly determines the halt status of
the above three functions.
> Also, you imply that your "input" isn't the input that actually needs to
> be given, as without the code of the specific HHH that this DDD calls,
> no Simulating Halt Decider could do the simulation that you talk about.
>
Your brain damage causes you to keep forgetting
that DDD has access to all of the machine code
in Halt7.obj. I told you this dozens of times
and you already forget by the time you reply.
> It should be noted that it is a well known property of Artifical
> Intelegence, and in particular, Large Languge Models, are built not to
> give a "correct" answer, but an answer the user will like. And thus they
> will pick up on the subtle clues of how things are worded to give the
> responce that seems to be desired, even if it is just wrong.
>
That is an incorrect assessment of how LLM systems work
and you can't show otherwise because you are wrong.
> When you add to the input the actual definition of "Non-Halting", as
> being that the exectuion of the program or its complete simulation will
> NEVER halt, even if carried out to an unbounded number of steps, they
> will give a different answer.
>
This is a whole other issue that I have addressed.
They figured out on their own that if DDD was correctly
simulated by HHH for an infinite number of steps that
DDD would never stop running.
> If you disagree with that definition, then you are admitting that you
> don't know the meaning of the terms-of-art of the system, but are just
> admitting to being the lying bastard that you are.
>
Two different LLM systems both agree that the halting
problem definition is wrong.
The standard proof assumes a decider
H(M,x) that determines whether machine
M halts on input x.
But this formulation is flawed, because:
Turing machines can only process finite
encodings (e.g. ⟨M⟩), not executable entities
like M.
So the valid formulation must be
H(⟨M⟩,x), where ⟨M⟩ is a string.
You cannot point out any error with that because
it is correct.
>
>>
>>> All you are doing is showing you don't understand how Artificiial
>>> Intelegence actualy works, showing your Natural Stupidity.
>>
>> That they provided all of the reasoning why DDD correctly
>> simulated by HHH does not halt proves that they do have
>> the functional equivalent of human understanding.
>
> But the problem is that your HHH that answers doesn't do a correct
> simulation.
>
All simulating termination analyzers only predict
what would happen if they did a complete simulation
on non terminating inputs.
It has always been completely nuts to require a non
terminating input to be simulating until its non-existent
completion.
> Yes, if *THE* HHH is one that correctly simulates the input (that has
> been fixed to include the code of HHH) then that simulation will not
> halt and be non-halting, but that HHH never answers.
>
Correctly *predict* the behavior of unlimited simulation,
not actually do an infinite simulation.
[continued in next message]
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
