Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 30 Jan 2025 08:47:32 +0000
From:      David Chisnall <theraven@freebsd.org>
To:        paige@paige.bio
Cc:        hackers@freebsd.org
Subject:   Re: Provisions to the contribution guidelines for using LLM generated code
Message-ID:  <4922BB4E-1361-4AE9-A40D-D75E4875033D@freebsd.org>
In-Reply-To: <49B92974-E37A-4786-A456-E258D5A1D35E@paige.bio>
References:  <49B92974-E37A-4786-A456-E258D5A1D35E@paige.bio>

next in thread | previous in thread | raw e-mail | index | archive | help
I am not a lawyer. If you want legal advice, you should talk to a lawyer. As=
 a not-a-lawyer, my opinion is:

Copyright law, in general, does not in any way describe how copying occurs. I=
f you photocopy a book, or if someone reads it to you and you write it down,=
 that=E2=80=99s equally copyright infringement or fair use based on the resu=
lt: the mechanism does not matter.

If you take a load of existing exFAT implementations, apply a lossy compress=
ion algorithm to them (neural network training) and then decompress, is the o=
utput a derived work of the input? That will depend on a load of tests that a=
 court can apply to judge similarity and so on. In general, a good legal rul=
e of thumb is that judges are not idiots (ignoring the Texas West District).=
 If you use an obfuscated process to hide your illegal action then they will=
 regard it as both illegal and wilful (and be annoyed with you), which is no=
t a good place to be.

Is your exFAT implementation a new creative work or a derived work of someth=
ing else? Does it infringe Microsoft=E2=80=99s exFAT patents? I don=E2=80=99=
t know and going to court is probably the only way of getting a definitive a=
nswer. Please don=E2=80=99t expose the FreeBSD project to that legal risk, d=
efending it would cost more than the annual budget of the Foundation.

David

> On 30 Jan 2025, at 02:05, paige@paige.bio wrote:
>=20
> =EF=BB=BFHi there,
>=20
> As y=E2=80=99all have probably heard AI is the new big thing in town and p=
eople are at a bit of a loss for what it means. Despite the news about the s=
tock market sell off that came in the wake of the new DeepSeek thing, I=E2=80=
=99ve actually been playing around with this thing called Claude for the pas=
t couple of weeks and I=E2=80=99m still not really sure what to think of it.=
 I think it=E2=80=99s really cool to say the least, but I still have a lot o=
f questions myself.
>=20
> More specifically, I=E2=80=99m not really sure at what point does using so=
mething like Claude to create something like a native ExFAT filesystem becom=
e an issue of attribution;
>=20
> https://github.com/paigeadelethompson/exfat/tree/main/sys/fs/exfat
>=20
> it presumably created this based on the parameters in it=E2=80=99s model (=
presumably, it is not actually known how Anthropic=E2=80=99s models work bec=
ause as far as I know that information is proprietary.) I vaguely understand=
 how it is able to do this and to the best of my knowledge, it doesn=E2=80=99=
t plagiarize code but it does generate code based on facts that it can find i=
n it=E2=80=99s own model about ideas which are potentially subject to patent=
 restrictions. For what this is worth, I think that people are going to find=
 this to be incredibly valuable regardless of whether or not it produces an e=
xact desired result. What it doesn=E2=80=99t get right the first time is oft=
en the subject of something being really damn close.
>=20
> I=E2=80=99m really just dumbfounded by how much it actually can do that I h=
aven=E2=80=99t even tried to compile this code for this filesystem it create=
d; it didn=E2=80=99t take me more than an hour of saying =E2=80=9Cyes=E2=80=9D=
 following the initial "I'd like to make an ExFAT driver for FreeBSD in C ca=
n you give me the best starting point possible?=E2=80=9D To be honest I kind=
a had to fact check it a couple of times, it wanted to do things like implem=
enting extattrs which this filesystem patently doesn=E2=80=99t have. But as s=
oon as I asked it, it seemed to know exactly what I meant:
>=20
> "No, you're right - I apologize for adding unnecessary complexity. The ExFA=
T specification doesn't include support for extended attributes like other f=
ilesystems (e.g., UFS or ext4). The only attributes ExFAT supports are the b=
asic DOS/FAT attributes we already have defined=E2=80=9D
>=20
> And then it proceeded to make changes to remove the stubs and so forth (wh=
ich it may not have done right but I haven=E2=80=99t gotten that far yet.) I=
n fact, I don=E2=80=99t really feel like I can realistically move forward wi=
th this (because I=E2=80=99ll have to fork $20 to get more time out of it) b=
ut also I just don=E2=80=99t really know whether or not this is okay. Obviou=
sly I want to say yes, but I get the impression that some people might not b=
e okay with this, especially if what it creates is not well understood or vi=
olates copyright laws.
>=20
> "Under U.S. law, you cannot patent an idea, but you may be able to protect=
 your idea by bringing it to life.=E2=80=9D As far as I know the licensing f=
or ExFAT is a little bit of a gray area. It=E2=80=99s Microsoft=E2=80=99s pa=
tent, there=E2=80=99s a GPL implementation that exists but asides from that I=
 don=E2=80=99t know if it=E2=80=99s technically okay to make another impleme=
ntation that is licensed any other way. I assume so, but it=E2=80=99s not un=
imaginable that even simply ingesting an ExFAT filesystem could come with so=
me kind of stipulation.
>=20
> And I=E2=80=99m sure some people might even think =E2=80=9Cwhy would you, t=
here=E2=80=99s a FUSE implementation for this already=E2=80=9D and you know b=
ecause FUSE is FUSE and this is an implementation of ExFAT that uses VFS. Al=
so ExFAT/fuse does have problems but it works (sorta) in a pinch. I=E2=80=99=
d personally be more interested in improving something that is part of core =
FreeBSD than I would anything having to do with a port that I have to instal=
l in addition to the OS itself in order to use it.
>=20
> The reason why it matters; I just really like ExFAT. Virtually everything n=
ow has native support for it out of the box except for UEFI (they should, su=
rprised Microsoft hasn=E2=80=99t pushed the standard to adopt it given that .=
WIM files can certainly exceed 4.3GB on modern versions of Windows. It just m=
akes good sense to me to use it, even though it=E2=80=99s not a journaled fi=
lesystem. Using parchive is not lost on me, but I=E2=80=99ve seldom ever tru=
ly needed it even with ExFAT.
>=20
> Maybe I=E2=80=99m not even really trying to drive this to completion as mu=
ch as I just needed an example and am just wanting to understand are people a=
lready doing this? Is it possible that people have already done this and nob=
ody is really aware of it? I=E2=80=99d like to think if you can then you cer=
tainly should but where do you draw the line, and should there perhaps be co=
nventions for keeping track of code in FreeBSD that is produced by LLMs? May=
be there already is and I just haven=E2=80=99t found it yet but it wouldn=E2=
=80=99t come as any surprise if there weren=E2=80=99t given this is all stil=
l kind of novel. Either way I=E2=80=99m sure there are things much more subs=
tantial than ExFAT worth trying, but there should probably be something of a=
n understanding about what is and isn=E2=80=99t okay. I wonder if what we do=
n=E2=80=99t know about proprietary LLMs like Claude could potentially be an e=
asily overlooked problem that could have legal consequences later.
>=20
> In any case I=E2=80=99m sure people will figure it out, but if anybody was=
 looking for a cue to discuss this I mean.. it=E2=80=99d be really useful to=
 me if FreeBSD supported ExFAT out of the box (especially since I can=E2=80=99=
t get to my offline archive of the ports and it=E2=80=99s distfiles without i=
t.) The only available implementations at present are GPL=E2=80=94 so can we=
 just like=E2=80=A6 generate an implementation with Claude and license it BS=
D? I honestly wish that my friend hadn=E2=80=99t insisted on showing me this=
 kinda because I hoped to avoid something that I know is certainly going to h=
ave repercussions for the way things are currently done, but I can=E2=80=99t=
 unsee this and I feel like I=E2=80=99ve been =E2=80=9Cdoing it wrong=E2=80=9D=
 my whole life.
>=20
> -Paige
>=20



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4922BB4E-1361-4AE9-A40D-D75E4875033D>