Re: What do LLMs mean for GnuPG?

"Robert J. Hansen via Gnupg-users" <[email protected]> Mon, 30 Mar 2026 06:58:17 -0400

Newsgroups	gmane.comp.encryption.gpg.user
Message-ID	<[email protected]>

> I am a user of tools like Cursor,- and my personal opinion is that LMM 
> is not perfect. But for those who cannot program because  of 
> neurological conditions, it is a valuable tool.

As a hacker who deals with mental illness, I am massively in favor of 
creating a community that is welcoming to people with psychological, 
psychiatric, and/or neurological troubles. But I draw the line at 
thinking we should lower our professional standards to accommodate these 
conditions.

I do not believe LLMs should be authoring security-sensitive code, ever.

> If the programming follows programming standards like  PEP 8, |rustfmt|, 
> clippy etc..

None of these verify quality code. The version of pwgen that Claude.ai 
created passed the normal clippy checks.

> tested against Wycheproof test vectors, RFC 5639, and BSI specifications 

Having implemented more cryptographic algorithms in my life than I ever 
want to think about, "it passes the test vectors" does not create in me 
very much faith in the overall quality of the implementation.

Circa 2008 at USENIX/EVT we gave a round of applause to a team of nerds 
who had implemented AES in Java, and given rigorous Floyd-Hoare proofs 
of correctness. It took them two and a half years of work.

I once had to implement a highly reliable Galois counter mode. I yearned 
for the sweet release of death.

> But,- who can write 100% perfect code ?

It's hard but it's been done. The provably-correct AES implementation 
comes to mind, as does the RSRE VIPER processor. The IBM System 360 
minicomputers that controlled the Space Shuttle never suffered a 
life-threatening bug, ever: even as _Challenger_ and _Columbia_ came 
apart the HAL/S software stack continued functioning correctly.

https://en.wikipedia.org/wiki/VIPER_microprocessor
https://en.wikipedia.org/wiki/HAL/S

_______________________________________________
Gnupg-users mailing list
[email protected]
https://lists.gnupg.org/mailman/listinfo/gnupg-users

OpenPGP_signature.asc (application/pgp-signature, 236 B)