Re: Amazing Visual Perception, even for ASCII Input [GPT 5.3] (Re: Only, three decades in software engineering?)

Liste des GroupesRevenir à s math 
Sujet : Re: Amazing Visual Perception, even for ASCII Input [GPT 5.3] (Re: Only, three decades in software engineering?)
De : ross.a.finlayson (at) *nospam* gmail.com (Ross Finlayson)
Groupes : sci.logic sci.math
Date : 08. Mar 2026, 18:59:29
Autres entêtes
Message-ID : <v2SdnS1V_KKAJDD0nZ2dnZfqn_adnZ2d@giganews.com>
References : 1 2 3 4 5 6 7 8 9 10
User-Agent : Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.6.0
On 03/08/2026 10:21 AM, Tristan Wibberley wrote:
On 08/03/2026 09:37, Mild Shock wrote:
Hi,
>
The amazing things is ChatGPT, and possibly
other LLMs don't think linearly. I literally
posted the code like this:
>
040C AF
040D 11 0000
0410 CD3300
0413 21 0000
0416 36 00
0418 E5
0419 D1
041A 13
041B 01 0904
041E ED B0
0420 3E 02
0422 32 0000
>
XOR A
LD DE,0000h
CALL 0033h
LD HL,0000h
LD (HL),00h
PUSH HL
POP DE
INC DE
LD BC,0409h
LDIR
LD A,02h
LD (0000h),A
>
Two consecutive columns. And it moved
the columns side by side all on its own.
I noticed that many times,
>
I'd expect that to be a very rudimentary technique. Now, if it
recognises that the bottom half and top half of the first column have
been swapped for each other I'd be marginally more impressed but imagine
if 1000 fake usenets have had this conversation... of course it's been
tried before and supported correctly today.
>
Can it produce instructions for housecats? I was watching videos of cats
with buttons and a cat announced its next two actions with:
>
Verb A
Verb B
Noun A
Noun B
>
where its owner kept trying to communicate with Verb A Noun A Verb B Noun B.
>
A usual notion of "templating" and "typing" (type theory's)
atop assembler as a sort of macro-assembler is a rather underserved
idea of the language as about a "1GL machine-wise, 4GL expressitivity".
There are usual ideas about the arithmetic logic unit, then as
with regards to the movement of data, register transfer logic
and the like, then as well about SIMD/SWAR and SIMD/SIMT,
or vector super-scalar units and subdividing those,
or vector super-scalar units and wide parallel simple units like GPUs,
about that those make for alike the, "pseudo-operations",
that are common across architectures.
The notion of templating and typing encompasses a lot
of expressitivity, concisely.
The metaprogramming is a usual account of the derived
after schema, while "generative programming" the idea
has been around since the 1960's.

Date Sujet#  Auteur
5 Mar 26 * The proper way to use LLMs to aid primary research into foundations59olcott
5 Mar 26 +- Re: The proper way to use LLMs to aid primary research into foundations1Mild Shock
5 Mar 26 +* Re: The proper way to use LLMs to aid primary research into foundations6Ross Finlayson
5 Mar 26 i+- Re: The proper way to use LLMs to aid primary research into foundations1Ross Finlayson
5 Mar 26 i`* Re: The proper way to use LLMs to aid primary research into foundations4Tristan Wibberley
6 Mar 26 i `* Re: The proper way to use LLMs to aid primary research into foundations3Ross Finlayson
7 Mar 26 i  `* Re: The proper way to use LLMs to aid primary research into foundations2Tristan Wibberley
7 Mar 26 i   `- Re: The proper way to use LLMs to aid primary research into foundations1Ross Finlayson
6 Mar 26 +* Re: The proper way to use LLMs to aid primary research into foundations38Mikko
6 Mar 26 i+* Re: The proper way to use LLMs to aid primary research into foundations30olcott
6 Mar 26 ii+* Re: The proper way to use LLMs to aid primary research into foundations3olcott
6 Mar 26 iii`* Re: The proper way to use LLMs to aid primary research into foundations2Ross Finlayson
6 Mar 26 iii `- Re: The proper way to use LLMs to aid primary research into foundations1olcott
7 Mar 26 ii+* Re: The proper way to use LLMs to aid primary research into foundations4Tristan Wibberley
7 Mar 26 iii+* Re: The proper way to use LLMs to aid primary research into foundations2olcott
7 Mar 26 iiii`- Re: The proper way to use LLMs to aid primary research into foundations1Ross Finlayson
7 Mar 26 iii`- Re: The proper way to use LLMs to aid primary research into foundations1Ross Finlayson
7 Mar 26 ii`* Re: The proper way to use LLMs to aid primary research into foundations22Mikko
7 Mar 26 ii +* Re: The proper way to use LLMs to aid primary research into foundations13olcott
7 Mar 26 ii i+* Re: The proper way to use LLMs to aid primary research into foundations2Tristan Wibberley
7 Mar 26 ii ii`- Re: The proper way to use LLMs to aid primary research into foundations1olcott
8 Mar 26 ii i`* Re: The proper way to use LLMs to aid primary research into foundations10Mikko
8 Mar 26 ii i `* Re: The proper way to use LLMs to aid primary research into foundations9olcott
9 Mar 26 ii i  `* Re: The proper way to use LLMs to aid primary research into foundations8Mikko
9 Mar13:34 ii i   +* Re: The proper way to use LLMs to aid primary research into foundations2olcott
10 Mar09:33 ii i   i`- Re: The proper way to use LLMs to aid primary research into foundations1Mikko
9 Mar13:42 ii i   `* Re: The proper way to use LLMs to aid primary research into foundations5olcott
10 Mar13:03 ii i    `* Re: The proper way to use LLMs to aid primary research into foundations4Tristan Wibberley
10 Mar14:45 ii i     `* Re: The proper way to use LLMs to aid primary research into foundations3olcott
10 Mar17:43 ii i      +- Re: The proper way to use LLMs to aid primary research into foundations1Ross Finlayson
10 Mar18:41 ii i      `- Re: The proper way to use LLMs to aid primary research into foundations1olcott
7 Mar 26 ii `* Re: The proper way to use LLMs to aid primary research into foundations8Tristan Wibberley
7 Mar 26 ii  `* Re: The proper way to use LLMs to aid primary research into foundations7olcott
9 Mar10:43 ii   `* Re: The proper way to use LLMs to aid primary research into foundations6Tristan Wibberley
9 Mar13:48 ii    +* Re: The proper way to use LLMs to aid primary research into foundations2olcott
10 Mar14:07 ii    i`- Re: The proper way to use LLMs to aid primary research into foundations1Tristan Wibberley
9 Mar20:34 ii    +* Re: The proper way to use LLMs to aid primary research into foundations2Ross Finlayson
10 Mar06:15 ii    i`- Re: The proper way to use LLMs to aid primary research into foundations1Ross Finlayson
10 Mar12:57 ii    `- Re: The proper way to use LLMs to aid primary research into foundations1Tristan Wibberley
6 Mar 26 i`* Re: The proper way to use LLMs to aid primary research into foundations7Ross Finlayson
6 Mar 26 i `* Re: The proper way to use LLMs to aid primary research into foundations6Ross Finlayson
8 Mar 26 i  +* Re: The proper way to use LLMs to aid primary research into foundations2Tristan Wibberley
8 Mar 26 i  i`- Re: The proper way to use LLMs to aid primary research into foundations1Ross Finlayson
8 Mar 26 i  `* Re: The proper way to use LLMs to aid primary research into foundations3olcott
8 Mar 26 i   `* Re: The proper way to use LLMs to aid primary research into foundations2Ross Finlayson
8 Mar 26 i    `- Re: The proper way to use LLMs to aid primary research into foundations1olcott
7 Mar 26 `* Don Knuth on the superb capability of Claude Opus 4.613olcott
7 Mar 26  `* Re: Don Knuth on the superb capability of Claude Opus 4.612Ross Finlayson
7 Mar 26   `* Re: Don Knuth on the superb capability of Claude Opus 4.611Ross Finlayson
7 Mar 26    `* Terrence Tao: On Jan 4, ChatGPT was able to produce a proof Erdos #728 (Re: Don Knuth on the superb capability of Claude Opus 4.6)10Mild Shock
7 Mar 26     `* Aristotle: The Era of Vibe Proving is Here (Re: Terrence Tao: On Jan 4, ChatGPT was able to produce a proof Erdos #728)9Mild Shock
8 Mar 26      +* Only, three decades in software engineering? (Re: Aristotle: The Era of Vibe Proving is Here)6Mild Shock
8 Mar 26      i`* Re: Only, three decades in software engineering? (Re: Aristotle: The Era of Vibe Proving is Here)5Mild Shock
8 Mar 26      i `* Amazing Visual Perception, even for ASCII Input [GPT 5.3] (Re: Only, three decades in software engineering?)4Mild Shock
8 Mar 26      i  `* Re: Amazing Visual Perception, even for ASCII Input [GPT 5.3] (Re: Only, three decades in software engineering?)3Tristan Wibberley
8 Mar 26      i   `* Re: Amazing Visual Perception, even for ASCII Input [GPT 5.3] (Re: Only, three decades in software engineering?)2Ross Finlayson
8 Mar 26      i    `- Visual Perception helps reading Matrices, Fractions, etc.. (Was: Amazing Visual Perception, even for ASCII Input)1Mild Shock
9 Mar11:24      `* Biobrain cannot digest "Artificial" in AI (Re: Aristotle: The Era of Vibe Proving is Here)2Mild Shock
10 Mar13:09       `- Re: Biobrain cannot digest "Artificial" in AI (Re: Aristotle: The Era of Vibe Proving is Here)1Tristan Wibberley

Haut de la page

Les messages affichés proviennent d'usenet.

NewsPortal