Re: OT: Vaccine safety

Liste des GroupesRevenir à ol advocacy 
Sujet : Re: OT: Vaccine safety
De : nospam (at) *nospam* needed.invalid (Paul)
Groupes : comp.os.linux.advocacy alt.comp.os.windows-11
Date : 26. May 2025, 18:31:02
Autres entêtes
Organisation : A noiseless patient Spider
Message-ID : <10128gl$243hd$1@dont-email.me>
References : 1 2 3 4 5 6 7 8 9 10
User-Agent : Ratcatcher/2.0.0.25 (Windows/20130802)
On Mon, 5/26/2025 7:18 AM, Carlos E. R. wrote:
On 2025-05-26 05:54, rbowman wrote:
On Sun, 25 May 2025 14:42:43 -0400, Paul wrote:
>
The AI can't write code, in computer languages for which insufficient
"history" exists to train them. The AI hung on me (not even the safety
timer went off), when I tried that.
>
https://www.bleepingcomputer.com/news/security/ai-hallucinated-code-
dependencies-become-new-supply-chain-risk/
>
Synopsis: AIs hallucinate package names and helpful blackhats create
packages to match, filled with malware.
 
:-o  !
 

Each article about AI, always lets slip some detail
about the "issues" they have.

There are different models that load sequentially. Some
of them have the parameters turned up, and it is partially
because the model is no good at what it is doing. These
can be placeholder modules, until they find a better solution.

Whatever counts the letter "R" in Strawberry, is not a math
module, it's some other module that hallucinates as part
of providing a solution. There are also math modules that
are trained on math problems and math files, that can produce
properly formatted math output, but you don't get those
for free particularly. And being isolated as payware, you don't
get to learn anything about the veracity.

Anyone who gets code slop from an AI, is checking it. Unlike
lawyers, you can't just submit that to the source tree
straight from the AI, as that will never work (never pass
all the test benches).

They can't write code in languages where insufficient training is
present. The AI doesn't know this. It doesn't know that the APL
knowledge it has, is too low to write code. The user has to
recognize the behaviors at the output.

When I asked for a printf in Raku, it knew
that was "say Hello World". I figured that was evidence
it knew something. But a request to write a short program,
did not produce output. And without symptoms or a log or any
information to go on, it's really hard to say whether the
problem even "got out of the gate". Maybe it didn't try and
the Strategy Planner cut it off.

   Paul

Date Sujet#  Auteur
24 May 25 * Re: OT: The government KNEW the "safe" vaccine would cause myocarditis11T
24 May 25 `* Re: OT: The government KNEW the "safe" vaccine would cause myocarditis10T
24 May 25  +- Re: OT: The government KNEW the "safe" vaccine would cause myocarditis1T
25 May 25  `* Re: OT: Vaccine safety8Philip Herlihy
25 May 25   `* Re: OT: Vaccine safety7rbowman
25 May 25    `* Re: OT: Vaccine safety6Paul
26 May 25     `* Re: OT: Vaccine safety5rbowman
26 May 25      +- Re: OT: Vaccine safety1Paul
26 May 25      +* Re: OT: Vaccine safety2Carlos E. R.
26 May 25      i`- Re: OT: Vaccine safety1Paul
27 May 25      `- Re: OT: Vaccine safety1Chris

Haut de la page

Les messages affichés proviennent d'usenet.

NewsPortal