XPost: comp.theory, sci.logic, sci.math
From: 643-408-1753@kylheku.com
On 2025-11-12, olcott wrote:
> On 11/12/2025 12:39 PM, Kaz Kylheku wrote:
>> On 2025-11-12, olcott wrote:
>>> The huge advantages of LLM systems is that they do not
>>> begin their review on the basis that [Olcott is wrong]
>>> is an axiom. No humans have ever been able to do this
>>> in thousands of reviews across dozens of forums.
>>
>> Your principal modus operandi is that you reject any piece of evidence
>> which contradicts your set view in any matter.
>
> Not at all. Not ever.
Even now, in this post.
>> Here you are also doing it again: completely overlooking all the times
>> someone has agreed with you in some point, and declaring that that
>> everyon has a /personal/ bias against you.
>
> People on these forums have only agreed with
> me at most 1% of the time and the only case
1% is much larger than zero.
> besides Ben the agreement was on trivialities.
Grasping trivialities is the extent of your skill, so that's
all you get.
For me, it's a triviality that your H's are deciding the D's
wrongly, but at that point we go over your head.
> You are still trying to get away with the utter
> nonsense that once correct non-termination
> behavior criteria have been correctly met
your criteria cannot decide the fact that the NON-DIAGONAL
test case void DDD(void) { HHH(DDD); return; } terminates.
The 1 return value is correct, and HHH(DDD) /can/ return it.
DDD will not "behave opposite"; it will terminate as the 1 says.
> that you can try again and get a different
> result. I would really like to think that
> you are not a damned liar, yet no alternative
> seems reasonably plausible.
Right; no alternative even seems plausible, like that:
- Turing was right;
- Church was right;
- Goedel was right;
- C. A. R. Hore was right;
- P. Olcott was wrong;
....
Nah, cannot be! Not even /plausible/ seeming.
>> Anyway, that /is/ the right approach toward anything. Anything anyone
>> says should be suspected of having a factual or logical flaw, until
>> proven otherwise.
>>
>
> Not when the basis of proof requires them
> to actually pay close attention when they
> are utterly unwilling to do this because they
> are so sure that I must be wrong.
You have a small body of claims that you have repeated for years.
The claims have been examined in excruciating detail by all
your interlocutors.
You have /never/ responded to any criticism. You will not analyze
any rebuttal, look at any code, nothing.
Just casual dismissals, accusations of lying, stupidity, ...
>> Here is an idea for you: maybe try being right 95% of the time, for a
>> while. Say two weeks, or a month. Instead of your usual, 95%+ wrong.
>
> I have been completely right on the essence of
> what I have been saying for 22 years.
L'essence; that's French for "gas".
>> There is a human bias at play here and I will explain it to you:
>> people are more motivated to respond when you are wrong.
>
> OK some honesty, that it refreshing.
>
>> If you're 80% wrong, the 0.8 fraction of your remarks that is
>> wrong will get more engagement than the 0.2 that are right.
>> Thus it might be that, among those of your remarks which fetch
>> engagement, the fraction which are wrong might be amplified to
>> something much higher, like 0.97.
>>
>> This is why social networking algorithms are rigged to spread
>> rage bait: to drum up engagement.
>>
>> It's amazing you don't have the maturity to know all this on your own;
>> that it has to be explained to a grown up.
>>
>
> It is a verified fact that I have been continually
Verified by what third party? Oh, you mean Claude AI and Chat GPT?
Your casual dismissals and personal attacks do not verify anything.
Get someone with serious academic credentials to "validate"
your shit, then talk.
> Now I have LLM systems that show the complete details
> of exactly how and why I am correct. If they were
> simply "yes men" they could not possibly do this.
The commercially available LLM systems provided by Anthropic,
OpenAI and others all have system prompts telling them not to
antagonize the user.
An extremely common complaint (probably in the top five)
is that they are "sycophantic".
>> Most of us here struggle not to say an incorrect thing. in a comp.*
>> newsgroup or elsewhere, yet here you practically made a sport out of it;
>> you say some wrong shit more times in a day than an NBA player takes a
>> shot at the hoop.
>>
>
> I say things that do not conform to conventional
> wisdom and people here don't even understand the
> reasoning behind conventional wisdom.
>
> When I point out the error in this reasoning people
> here are utterly helpless. The most they can do is
> say that I must be wrong entirely on the basis
> that I contradict conventional wisdom.
You do not contradict conventional wisdom; you contradict
air-tight logic.
Conventional wisdom is shit like "eating carrots makes you
see sharper".
Incomputability of halting is entirely unconventional; it took
someone very smart to recognize the problem and answer it.
--
TXR Programming Language: http://nongnu.org/txr
Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
Mastodon: @Kazinator@mstdn.ca
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
