From Newsgroup: comp.lang.c++

AI is great but you have to be able to call on everything you learnt prior
to the arrival of LLMs to leverage the stochastic parrots properly (i.e. fixing their errors and flat out hallucinations): this is why pure vibe
coding by people with no coding experience remains bullsh*t. #compsci #AI

/Flibble
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On Sat, 2026-02-21 at 21:47 +0000, HAL 9000 wrote:

AI is great but you have to be able to call on everything you learnt prior to the arrival of LLMs to leverage the stochastic parrots properly (i.e. fixing their errors and flat out hallucinations): this is why pure vibe coding by people with no coding experience remains bullsh*t. #compsci #AI

/Flibble

https://www.hindustantimes.com/world-news/microsoft-ai-ceo-says-most-white-collar-jobs-to-be-replaced-with-ai-in-12-months-models-coding-better-than-humans-101770906229347.html
Microsoft AI CEO Mustafa Suleyman said that today there are AI models that can code better than vast
majority of human coders, (AFP/File Photo)
Don't complain, just show how you can beat AI, losers. https://edition.cnn.com/2025/07/13/business/video/gps0713-ai-workforce-nvidia-productivity
... the fundamental thing is, do we have left more ideas...
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On 21/02/2026 21:47, HAL 9000 wrote:

AI is great but you have to be able to call on everything you learnt prior to the arrival of LLMs to leverage the stochastic parrots properly (i.e. fixing their errors and flat out hallucinations): this is why pure vibe coding by people with no coding experience remains bullsh*t. #compsci #AI

Vibe coding is just requirements engineering practice on a model of a
segment of human coders. Vibe coders create policies for a software construction gig economy (in effect, the defence industry, because who
else would need software construction prompts applicable to gig coding).
--
Tristan Wibberley

The message body is Copyright (C) 2026 Tristan Wibberley except
citations and quotations noted. All Rights Reserved except that you may,
of course, cite it academically giving credit to me, distribute it
verbatim as part of a usenet system or its archives, and use it to
promote my greatness and general superiority without misrepresentation
of my opinions other than my opinion of my greatness and general
superiority which you _may_ misrepresent. You definitely MAY NOT train
any production AI system with it but you may train experimental AI that
will only be used for evaluation of the AI methods it implements.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On 2/22/2026 5:47 AM, HAL 9000 wrote:

AI is great but you have to be able to call on everything you learnt prior
to the arrival of LLMs to leverage the stochastic parrots properly (i.e. fixing their errors and flat out hallucinations): this is why pure vibe coding by people with no coding experience remains bullsh*t. #compsci #AI

Why do you trust source codes found by A.I.?

A.I. merely helps programmers find canned source codes. Opening the can
of worms, doing quality control and auditing security should still be
done by human.

BTW, I think you were talking about:

THE 2028 GLOBAL INTELLIGENCE CRISIS
<https://www.citriniresearch.com/p/2028gic>

How AI agents could destroy the economy | TechCrunch <https://techcrunch.com/2026/02/23/how-ai-agents-could-destroy-the-economy/>
--
@~@ Simplicity is Beauty! Remain silent! Drink, Blink, Stretch!
/ v \ May the Force and farces be with you! Live long and prosper!!
/( _ )\ https://sites.google.com/site/changmw/
^ ^ https://github.com/changmw/changmw
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On 25/02/2026 11:21, Mr. Man-wai Chang wrote:

On 2/22/2026 5:47 AM, HAL 9000 wrote:

AI is great but you have to be able to call on everything you learnt
prior
to the arrival of LLMs to leverage the stochastic parrots properly (i.e.
fixing their errors and flat out hallucinations): this is why pure vibe
coding by people with no coding experience remains bullsh*t. #compsci #AI

Why do you trust source codes found by A.I.?

A.I. merely helps programmers find canned source codes. Opening the can
of worms, doing quality control and auditing security should still be
done by human.

helps, helped, or will help?

mere LLMs don't even help find canned source codes, expert systems do that.

Automated theorem provers can translate requirements direct to correct
(and provably correct) programs (by the Curry-Howard correspondence).
Auditing security should be done by these.

Expert Systems and Automated Theorem Provers are AI too.

LLMs could assist in generating conjectures for the automated theorem
provers but then human language could become formalised and not natural
and then we're in a double-plus-dystopia. LLMs should be very useful to
create semi-formal statements that requirements encoders and customers
mutually understand, but perhaps not by linguistic prompts but by
providing faster navigation through the space of meaning - and faster
mapping direct from the space of meaning to formal expressions of
requirements.
--
Tristan Wibberley

The message body is Copyright (C) 2026 Tristan Wibberley except
citations and quotations noted. All Rights Reserved except that you may,
of course, cite it academically giving credit to me, distribute it
verbatim as part of a usenet system or its archives, and use it to
promote my greatness and general superiority without misrepresentation
of my opinions other than my opinion of my greatness and general
superiority which you _may_ misrepresent. You definitely MAY NOT train
any production AI system with it but you may train experimental AI that
will only be used for evaluation of the AI methods it implements.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On 2/27/2026 6:39 PM, Tristan Wibberley wrote:

helps, helped, or will help?

mere LLMs don't even help find canned source codes, expert systems do that.

Automated theorem provers can translate requirements direct to correct
(and provably correct) programs (by the Curry-Howard correspondence). Auditing security should be done by these.

Expert Systems and Automated Theorem Provers are AI too.

Would you, as a human, blindly trust A.I. buttons called "security" and "programming"? You still need to examine and audit the source codes of
A.I. You still need to patrol & sample generated codes to gurantee that
they were not tampered with. Persons doing these jobs still need to know programming and read source codes.
--
@~@ Simplicity is Beauty! Remain silent! Drink, Blink, Stretch!
/ v \ May the Force and farces be with you! Live long and prosper!!
/( _ )\ https://sites.google.com/site/changmw/
^ ^ https://github.com/changmw/changmw
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On 2/24/2026 3:51 AM, wij wrote:

Microsoft AI CEO Mustafa Suleyman said that today there are AI models that can code better than vast
majority of human coders, (AFP/File Photo)

Don't complain, just show how you can beat AI, losers.

Prove that A.I. is really coding not copying or stealing some human's
work! :)
--
@~@ Simplicity is Beauty! Remain silent! Drink, Blink, Stretch!
/ v \ May the Force and farces be with you! Live long and prosper!!
/( _ )\ https://sites.google.com/site/changmw/
^ ^ https://github.com/changmw/changmw
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On 2/25/2026 2:07 AM, Tristan Wibberley wrote:

Vibe coding is just requirements engineering practice on a model of a
segment of human coders. Vibe coders create policies for a software construction gig economy (in effect, the defence industry, because who
else would need software construction prompts applicable to gig coding).\

What made vibe more trustworhty than non-vibe? Because vibe has "weapons
of mass destruction"? Because vibe is authorized to kill humans? :)
--
@~@ Simplicity is Beauty! Remain silent! Drink, Blink, Stretch!
/ v \ May the Force and farces be with you! Live long and prosper!!
/( _ )\ https://sites.google.com/site/changmw/
^ ^ https://github.com/changmw/changmw
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.lang.c++

On 2/22/2026 5:47 AM, HAL 9000 wrote:

AI is great but you have to be able to call on everything you learnt prior
to the arrival of LLMs to leverage the stochastic parrots properly (i.e. fixing their errors and flat out hallucinations): this is why pure vibe coding by people with no coding experience remains bullsh*t. #compsci #AI

Ultimately, Who is/are to be fired or even sued if A.I. went wrong? You?
You? You?

Who wanna take total responsibilty?

Not me! Because I dunno how to trust A.I. and its programmers! :)
--
@~@ Simplicity is Beauty! Remain silent! Drink, Blink, Stretch!
/ v \ May the Force and farces be with you! Live long and prosper!!
/( _ )\ https://sites.google.com/site/changmw/
^ ^ https://github.com/changmw/changmw
--- Synchronet 3.21d-Linux NewsLink 1.2