Re: DD emulated by HHH cannot possibly terminate normally --- x86 code

On 3/1/25 6:47 PM, olcott wrote:

On 3/1/2025 2:22 AM, Mikko wrote:

On 2025-03-01 00:47:58 +0000, olcott said:
>

_DD()
[00002133] 55         push ebp      ; housekeeping
[00002134] 8bec       mov ebp,esp   ; housekeeping
[00002136] 51         push ecx      ; make space for local
[00002137] 6833210000 push 00002133 ; push DD
[0000213c] e882f4ffff call 000015c3 ; call HHH(DD)
[00002141] 83c404     add esp,+04
[00002144] 8945fc     mov [ebp-04],eax
[00002147] 837dfc00   cmp dword [ebp-04],+00
[0000214b] 7402       jz 0000214f
[0000214d] ebfe       jmp 0000214d
[0000214f] 8b45fc     mov eax,[ebp-04]
[00002152] 8be5       mov esp,ebp
[00002154] 5d         pop ebp
[00002155] c3         ret
Size in bytes:(0035) [00002155]
>
When we hypothesize that the code at machine address
0000213c is an x86 emulator then we know that DD
remains stuck in recursive emulation and cannot possibly
reach its own "ret" instruction and terminate normally.

>
The emulator itself is stuck and cannot return normally but it doesn't
know it cannot return normally. At some point it runs out of memory
and terminates normally or abnormally.
>

 Yes you are correct about this sub-step of two steps.
 

When we add the additional complexity that HHH also
aborts this sequence at some point then every level
of recursive emulation immediately stops. This does
not enable any DD to ever reach its "ret" instruction.

>
When we add an additional complexity we must note that there are other
additional complexities that could be added instead.
>

 Sure we could carefully examine every detail about the price
of tea in China.
 When we are answering the question that seems impossible for
anyone here to pay attention to even when repeated hundreds of times:
 Can the above DD correctly emulated by HHH possibly
reach its own "ret" instruction and terminate normally?

The problem is that if HHH correctly emulates DD, then it can not answer, and thus YOU are missing that you are making an assumption that you are not following.
And, the question you are asking is just a strawman, proving that you don't know what a halt decider is, as a halt decider does not determine if it can simulate the input to a final state, it determines if the program when run or actually completely correctly emulated would halt.
The fact that HHH's aborted emulation doesn't reach a final state means nothing, as for this case, the correct emulation of THIS PROGRAM (which calls the HHH that aborted its emulation) will reach the final state, and the HHH that does a complete emulation (and get stuck) was given a DIFFERENT version of DD, since it calls a different version of HHH, which means you had to change the code of the input program.
All you are doing is proving you don't understand what a program is, or that you don't care about lying about it.

 The answer is dead obviously "no" for everyone that is:
(a) Technically competent
   and
(b) Not deliberately deceptive.

But that HHH can not correctly emulate the input to that point, doesn't say that something else can't if HHH doesn't also do that correct emulation.
YOU are showing that it is YOU who is technically incompetent and being deliberately deceptive, by your instance of making assumptions that are in fact not true, like your HHH is actually doing a correct emulation.
Until you stop doing that, nothing you say has any value, and you are just proving that you are just a pathologically lying hypocrite.

 Once we get through this then people will understand
that my system would eliminate AI hallucinations.
 Until AI has a foundation of basic facts it will always
view everything as some kind of fiction.
 

AI, as currently implemented, isn't based on a foundation of facts, as the problem of trying to prove answers is just intractable and is too slow.
There are small systems that use that sort of AI, it just is limited to small systems. You just don't understand the nature of the problem.
And, it would still be impossible for the AI to answer about things whose answer is unprovable.