Re: AGENTS.md and CLAUDE.md Are Concessions to Proprietary Services, Not Defenses Against Them

On 4 August 2026 08:58:54 GMT+03:00, Jean Louis <bugs@gnu=2Esupport> wrote:
> Eli, you say I am "confused" and that article about Service as a Softwar=
e Substitute has nothing to do with this discussion: Who Does That Server R=
eally Serve? - GNU Project - Free Software Foundation: https://www=2Egnu=2E=
org/philosophy/who-does-that-server-really-serve=2Ehtml
>=20
> Maybe you miss to read what I actually wrote due to rejecting me persona=
lly?
>=20
> I talk about strategic concession while you wish to do tactical defense=
=2E
>=20
> You said that you are using "features of LLM agents against them" to dis=
courage LLM generated code, and yet what you are going to achieve is making=
 Emacs compatible with proprietary service=2E
>=20
> You are not using "features of the LLM" as AGENTS=2Emd and CLAUDE=2Emd a=
re not features of the LLM, they are features of some pieces of software th=
at use LLMs, not features of LLMs=2E

This just reiterates what you already wrote, so it is not useful=2E

> I have 450 LLMs on my computer as I run them and have experience with th=
em=2E Do you? Do you think that all of the development is done exclusively =
over those few software packages that use AGENTS=2Emd or CLAUDE=2Emd? It is=
 not=2E

A partial solution is still an improvement, especially if it affects the m=
ost popular and frequent user choices=2E

> LLM as model is different from software running model!
>=20
> Software running model gives prompts to follow AGENTS=2Emd -- not the LL=
M=2E
>=20
> Even if LLM would be trained on AGENTS=2Emd while majority of LLMs in th=
e world ARE NOT TRAINED ON THAT
>=20
> The AGENTS=2Emd and CLAUDE=2Emd are inventions by proprietary company An=
thropic and made originally and specifically for their proprietary software=
=2E

I don't see why we should care about their motivation=2E  We use this for =
our purposes, not theirs=2E

> The file itself is concession=2E As those files are exclusive for those =
few software packages using the guide from Anthropic company=2E

Using a file by some name cannot be a concession to anything=2E  This is a=
 ridiculous claim, which basically makes your entire argument invalid=2E

> And those files are not related to LLMs as much, please do your research=
, LLMs are many, millions of them, and those files do not block LLM or guid=
e them, as there is no universal standard for it=2E

We are aware that not all agents will pay attention, but some popular ones=
 will=2E  And we can add more files for those other agents=2E  Again, a par=
tial solution is still an improvement=2E

> Those files are related to software that uses LLMs, and that software gi=
ves guidances prompts, and user who wish to do anything with Emacs can give=
 prompt to ignore those files and to override file's instructions=2E

We are trying to help our contributors avoid good-faith mistakes=2E  It is=
 clear that a malevolent person can overcome these measures, but that is tr=
ue regardless of LLM use, and we can do very little about it, if at all=2E =
 This is not our aim here=2E

> I understand very well what you are trying to do, but due to extensive u=
se of LLMs on my side I see that the only effect you will achieve is bow to=
 proprietary Anthropic=2E

How can preventing code generation by LLM be seriously considered "bowing"=
 to them?  That's absurd=2E

> You wish to prevent LLM-generated code, what you can do is maybe put som=
e legal notices=2E

We plan to do that as well=2E  This is just one of the measures we are con=
sidering=2E

> What works is review process, legal notices, git hooks, CONTRIBUTE polic=
y, talking more to developers to gain general understanding=2E

And we are discussing addition of those as well=2E  These =2Emd files are =
not replacements for the other measures, they are add-ons=2E  And we haven'=
t even decided yet whether we will add such files, we are still discussing=
=2E


Why did you send 3 identical copies of the same message?  Please don't do =
that, as it only increases the noise level here=2E