<@jflory7:fedora.im>
14:02:28
!startmeeting Fedora Council
<@meetbot:fedora.im>
14:02:30
Meeting started at 2024-05-22 14:02:28 UTC
<@meetbot:fedora.im>
14:02:31
The Meeting name is 'Fedora Council'
<@jflory7:fedora.im>
14:02:40
!hi
<@dcantrell:fedora.im>
14:02:41
!hi
<@zodbot:fedora.im>
14:02:43
Justin W. Flory (jflory7) - he / him / his
<@zodbot:fedora.im>
14:02:44
David Cantrell (dcantrell) - he / him / his
<@ffmancera:fedora.im>
14:02:46
!hi
<@moralcode:fedora.im>
14:02:47
!hi
<@zodbot:fedora.im>
14:02:56
None (moralcode)
<@zodbot:fedora.im>
14:02:56
FAS Fernando F. Mancera (ffmancera) - he / him / his
<@bookwar:fedora.im>
14:02:58
!hi
<@zodbot:fedora.im>
14:03:01
Aleksandra Fedorova (bookwar) - she / her / hers
<@jonatoni:fedora.im>
14:03:03
!hi
<@zodbot:fedora.im>
14:03:07
Jona Azizaj (jonatoni) - she / her / hers
<@jflory7:fedora.im>
14:03:11
Adrian Edwards: zodbot has all kinds of commands and shortcuts. It integrates with FAS. Feel free to open a DM with Zodbot and explore some of the commands :)
<@blackwell:fedora.im>
14:03:23
!hi
<@zodbot:fedora.im>
14:03:26
Jason Blackwell (blackwell)
<@jflory7:fedora.im>
14:03:36
OK, I think Aoife was going to chair today but I am not 100% certain. If not, I can help drive today
<@jflory7:fedora.im>
14:05:22
!info Present: @jflory7, @dcantrell, @moralcode, @ffmancera, @bookwar, @jonatoni, @blackwell, @mattdm
<@mattdm:fedora.im>
14:05:35
hi everyone
<@mattdm:fedora.im>
14:05:42
apparently i wasn't in this room
<@t0xic0der:fedora.im>
14:06:01
Alright - I am here
<@amoloney:fedora.im>
14:06:05
!hi
<@zodbot:fedora.im>
14:06:08
Aoife Moloney (amoloney)
<@jflory7:fedora.im>
14:06:09
!topic Team news and community announcements
<@t0xic0der:fedora.im>
14:06:12
!hi
<@zodbot:fedora.im>
14:06:14
Akashdeep Dhar (t0xic0der) - he / him / his
<@jflory7:fedora.im>
14:06:21
!info +Present: @t0xic0der, @amoloney
<@jflory7:fedora.im>
14:06:30
Does anyone have news or announcements to share?
<@jflory7:fedora.im>
14:06:33
I definitely have a couple.
<@mattdm:fedora.im>
14:06:43
wow, fedocal is a mess. There are so many different 'fedora-meeting' locations.
<@jflory7:fedora.im>
14:07:21
!info The F40 Release Party is happening this weekend on Friday and Saturday. Join us for a weekend of celebration for the last release of Fedora! 🎉 Register now to get an invite to the Matrix room (which we are using for the first time over Hopin).
<@jflory7:fedora.im>
14:07:23
!link https://pretix.eu/fedora/f40-party/
<@jflory7:fedora.im>
14:07:42
I suspect it is an easy fix, but I am not sure if there is a UI or someone has to do some database-level hackery.
<@jflory7:fedora.im>
14:07:55
Adrian Edwards: Maybe something we could have you check out, depending on the level of access needed to fix.
<@jflory7:fedora.im>
14:08:07
Also, re: F40 party
<@jflory7:fedora.im>
14:08:10
Aoife Moloney++
<@jflory7:fedora.im>
14:08:13
Joseph Gayoso++
<@zodbot:fedora.im>
14:08:15
jflory7 has already given cookies to amoloney during the F40 timeframe
<@jflory7:fedora.im>
14:08:15
Neil Hanlon++
<@zodbot:fedora.im>
14:08:16
jflory7 has already given cookies to joseph during the F40 timeframe
<@zodbot:fedora.im>
14:08:19
jflory7 has already given cookies to neil during the F40 timeframe
<@moralcode:fedora.im>
14:08:22
am down for whatever! i just like fixing stuff
<@jflory7:fedora.im>
14:08:31
nekonya3++
<@zodbot:fedora.im>
14:08:36
jflory7 has already given cookies to nekonya3 during the F40 timeframe
<@jflory7:fedora.im>
14:08:39
ekidney++
<@nhanlon:beeper.com>
14:08:43
/me eats the breakfast cookies
<@zodbot:fedora.im>
14:08:45
jflory7 has already given cookies to ekidney during the F40 timeframe
<@jflory7:fedora.im>
14:09:01
All have been indispensable for getting this party off the ground and going to Matrix for the first time! Thank you all :)
<@mattdm:fedora.im>
14:09:17
yes!
<@mattdm:fedora.im>
14:09:26
did you get the stream recording stuff figured out?
<@jflory7:fedora.im>
14:09:27
Any other news and announcements, folks?
<@jflory7:fedora.im>
14:09:32
Anything cool happening in the Fedora-verse? :)
<@jflory7:fedora.im>
14:09:52
Yes. I am likely going to use Restream this time. It will be annoying because the logo will be on things, since I am not sure I want to pay for it yet. But if it works well, I just might.
<@mattdm:fedora.im>
14:10:18
how much would the one-time cost be?
<@jflory7:fedora.im>
14:10:19
That beats an OBS set-up on my own laptop which means everything is dependent on my home Internet connection :)
<@amoloney:fedora.im>
14:10:23
!info Fedora Week of Diversity is June 17th - 22nd and the organisers are looking for content
<@amoloney:fedora.im>
14:10:34
!link https://discussion.fedoraproject.org/t/join-us-in-crafting-a-video-for-fedora-week-of-diversity/116728
<@jflory7:fedora.im>
14:10:42
$190/year
<@moralcode:fedora.im>
14:10:57
would it be better to have something set up in the office or on the community infra?
<@amoloney:fedora.im>
14:11:05
!info Elections for Council, Mindshare, FESCo & EPEL are live until May 31st
<@amoloney:fedora.im>
14:11:12
!link go vote!! https://elections.fedoraproject.org/
<@jflory7:fedora.im>
14:11:29
Possibly. Let's discuss in #commops:fedoraproject.org after this meeting!
<@amoloney:fedora.im>
14:12:54
there seems to be a weird lag on my side for the msgs, Im seeing I posted duplicates and I didnt get any of Justins msgs
<@jflory7:fedora.im>
14:13:37
Might be if you just joined this room today, there is a lot of history in this room. Matrix can be… slow
<@t0xic0der:fedora.im>
14:13:41
Aoife Moloney: I experience the same but I like to think my ISP is to blame
<@t0xic0der:fedora.im>
14:13:49
But it might not be - Matrix ain't that heavy
<@jflory7:fedora.im>
14:14:10
It is working fine for me but I have been in this room for a while. New folks who just joined will likely take some time for their clients to "catch up"
<@t0xic0der:fedora.im>
14:14:14
Justin W. Flory (he/him): Interesting. I did join the room right now so that might be the case too.
<@jflory7:fedora.im>
14:14:16
Anyways…
<@jflory7:fedora.im>
14:14:19
Any other announcements?
<@t0xic0der:fedora.im>
14:14:23
Anyway...
<@jflory7:fedora.im>
14:14:28
And then Aoife Moloney, should I pass the gavel to you or want me to keep driving?
<@jflory7:fedora.im>
14:14:49
Going once…
<@jflory7:fedora.im>
14:15:05
Going twice…
<@jflory7:fedora.im>
14:15:20
Going thrice…
<@jflory7:fedora.im>
14:15:35
💥
<@jflory7:fedora.im>
14:15:47
!topic Ticket review
<@jflory7:fedora.im>
14:16:01
Aoife Moloney: Did we already have a list of prioritized tickets? I know we were in the process of optimizing this.
<@jflory7:fedora.im>
14:16:34
!link https://pagure.io/Fedora-Council/tickets/issues
<@jflory7:fedora.im>
14:16:47
I would love to talk about the bootc initiative
<@amoloney:fedora.im>
14:16:49
Id keep driving please Justin, my msgs are not reliable
<@amoloney:fedora.im>
14:16:50
we did
<@amoloney:fedora.im>
14:16:50
Fedora boot-c initiative - #492
Fedora AI policy - #486
<@jflory7:fedora.im>
14:17:06
Can do
<@jflory7:fedora.im>
14:17:19
Excellent. Then let's do these two. I suspect both of these will be enough to fill the hour.
<@jflory7:fedora.im>
14:17:36
And there are so many pretty, colorful labels now, yay! https://pagure.io/Fedora-Council/tickets/issues?tags=Next+Meeting&status=Open
<@jflory7:fedora.im>
14:17:58
!topic #492: New Initiative: Fedora bootc
<@amoloney:fedora.im>
14:18:00
oh yeah I got rid of the prioritization and added them as tags :)
<@jflory7:fedora.im>
14:18:04
!link https://pagure.io/Fedora-Council/tickets/issue/492
<@zodbot:fedora.im>
14:18:12
jflory7 has already given cookies to amoloney during the F40 timeframe
<@jflory7:fedora.im>
14:18:21
!link https://discussion.fedoraproject.org/t/fedora-council-tickets-ticket-492-new-initiative-fedora-bootc/116062
<@jflory7:fedora.im>
14:18:31
There has not been a lot of feedback so far…
<@jflory7:fedora.im>
14:18:36
Does anyone have questions?
<@jflory7:fedora.im>
14:18:47
Jason Brooks: Are you here? Any questions for Council about bootc initiative?
<@dcantrell:fedora.im>
14:19:23
I am on this project on the engineering side. We have this change proposal posted already: https://fedoraproject.org/wiki/Changes/DNFAndBootcInImageModeFedora
<@mattdm:fedora.im>
14:19:29
I need to sit down with Jason around logic models :)
<@jbrooks:matrix.org>
14:19:33
Hi, I'm here
<@jflory7:fedora.im>
14:19:37
!link https://fedoraproject.org/wiki/Changes/DNFAndBootcInImageModeFedora
<@jflory7:fedora.im>
14:20:01
OK. We have a lot of the right folks here to either move this ticket forward or ask for blocking feedback to approval.
<@jflory7:fedora.im>
14:20:07
I would really like us to process this now if we can!
<@mattdm:fedora.im>
14:20:23
I think the various goals are _good_ but they're not really fit to the format.
<@jflory7:fedora.im>
14:20:26
I have reviewed the proposal and I am happy with it
<@jflory7:fedora.im>
14:20:41
I think it engages a lot of our key stakeholders and the outcomes look good to me.
<@jflory7:fedora.im>
14:20:51
Logic model specifically, or as an Initiative?
<@jbrooks:matrix.org>
14:20:54
I tried to use another initiative as an example, but I'm happy for additional feedback
<@mattdm:fedora.im>
14:21:06
logic model specifically
<@mattdm:fedora.im>
14:21:23
some of the other initiatives aren't great examples because we haven't been strict with that
<@mattdm:fedora.im>
14:21:36
I don't want to penalize you unfairly for that, though :)
<@ffmancera:fedora.im>
14:21:54
I think all the initiatives got that feedback so probably we all need to sit together to discuss about logic models
<@mattdm:fedora.im>
14:22:23
Yes, exactly. And I promised I'd help but things have been crazy.
<@mattdm:fedora.im>
14:22:52
So I'm +1 to this, with the understanding that I _actually will_ meet with Jason and Sumantro to refactor.
<@amoloney:fedora.im>
14:23:04
I support the initiative, and we can coach/guide the initiative lead(s) for a better refined logic model
<@jflory7:fedora.im>
14:23:05
Are we ready to vote on this?
<@amoloney:fedora.im>
14:23:11
yep
<@jflory7:fedora.im>
14:23:18
Is there any other feedback or discussion on the outcomes specifically?
<@ffmancera:fedora.im>
14:23:28
yes from my side
<@jflory7:fedora.im>
14:23:28
Beyond the logic model, I want to make sure we are confident in this as an Initiative.
<@jflory7:fedora.im>
14:23:51
Okay, then let's do it. Old-school please! +1/+0/-1 votes in the chat :)
<@jflory7:fedora.im>
14:23:52
+1
<@dcantrell:fedora.im>
14:24:01
+1
<@ffmancera:fedora.im>
14:24:05
+1
<@t0xic0der:fedora.im>
14:24:06
+1
<@mattdm:fedora.im>
14:24:12
+1
<@jflory7:fedora.im>
14:24:30
!group members council
<@jonatoni:fedora.im>
14:24:30
+1
<@zodbot:fedora.im>
14:24:35
Members of council: Aoife Moloney, bt0dotninja, David Cantrell, FAS Fernando F. Mancera, Justin W. Flory, Jona Azizaj, Matthew Miller, Robert Wright, smeragoel, Sumantro Mukherjee, Akashdeep Dhar
<@amoloney:fedora.im>
14:24:41
I want to see this in Fedora, and aside from supporting it I would like to make sure that Jason Brooks feels confident that if this is approved (which it likely will be) that he and others can deliver on it
<@mattdm:fedora.im>
14:24:43
This is a direct, concrete effort to advance one of our Strategy 2028 goals.
<@jbrooks:matrix.org>
14:25:35
Aoife Moloney: I do feel confident, the WG reps I have listed on the wiki were quick to get involved, and they've been really responsive and running w/ it so far
<@amoloney:fedora.im>
14:25:37
and if there are specific things we can help you with @jbrooks if you are less confident about certain things :)
<@jbrooks:matrix.org>
14:26:13
I want to make sure we nail the infra aspects, that we do this the Right Fedora Way
<@amoloney:fedora.im>
14:26:36
Love that
<@amoloney:fedora.im>
14:26:50
Im a +1
<@jflory7:fedora.im>
14:27:38
!agreed +7/+0/-0 for bootc Initiative, four absent members.
<@jflory7:fedora.im>
14:27:55
Well, actually
<@jflory7:fedora.im>
14:28:03
!info +1 from Sumantro in the ticket, making it +8
<@jflory7:fedora.im>
14:28:20
Oh! Well, speaking of :)
<@jflory7:fedora.im>
14:28:22
Hi Sumantro Mukherjee
<@sumantrom:fedora.im>
14:28:23
+1 fro me
<@sumantrom:fedora.im>
14:28:28
hey all
<@sumantrom:fedora.im>
14:28:35
hey Justin W. Flory (he/him)
<@t0xic0der:fedora.im>
14:29:09
Hi Sumantro Mukherjee ! 👋
<@mattdm:fedora.im>
14:29:29
The ticket has been open for 18 days, so anyone who wanted to register a -1 vote could have. With comfortably more than 3 +1s, this passes.
<@jflory7:fedora.im>
14:29:33
The ticket has been open for 18 days.
<@jflory7:fedora.im>
14:29:43
!agreed The Initiative is officially approved!
<@jflory7:fedora.im>
14:29:55
!info Welcome Jason Brooks as the newest member of the Fedora Council. 🎉
<@jbrooks:matrix.org>
14:30:12
Awesome, thanks 🎉
<@mattdm:fedora.im>
14:30:18
If one of the council members who isn't here had other things they were attending to and has concerns later, we can address them. But I'm not particularly worried in this case.
<@jflory7:fedora.im>
14:30:31
I am not worried either
<@jflory7:fedora.im>
14:30:39
Lots of strong support for this, and I recognize lots of names on the Wiki
<@jflory7:fedora.im>
14:30:48
I am very excited for this Initiative! 💪
<@jflory7:fedora.im>
14:30:58
OK, well, we do have another ticket, and likely a hot topic too.
<@mattdm:fedora.im>
14:30:59
!action mattdm schedule meeting with Jason and Sumantro to hack on the logic model
<@jflory7:fedora.im>
14:31:04
Anything else on the bootc Initiative?
<@sumantrom:fedora.im>
14:31:15
yess!! mattdm
<@jflory7:fedora.im>
14:31:15
Aoife Moloney: We can tag-team the Council onboarding for Jason?
<@amoloney:fedora.im>
14:31:36
Yep sounds good
<@jflory7:fedora.im>
14:31:50
!action @jflory7 @amoloney Collaborate on onboarding @jasonbrooks to the Fedora Council (FAS, mailing list, Pagure, docs, etc.)
<@jflory7:fedora.im>
14:32:07
!action @jasonbrooks Write a self-introduction to be incorporated into the Fedora Council docs later, as an Initiative lead
<@jflory7:fedora.im>
14:32:17
Actually, we need to do that with ffmancera, Smera, and Robert too :)
<@jflory7:fedora.im>
14:32:27
re: self-introductions
<@jflory7:fedora.im>
14:32:35
Going once…
<@jflory7:fedora.im>
14:32:55
Going twice…
<@ffmancera:fedora.im>
14:33:08
just a reminder, I have a PR pending updating the docs of council members
<@jflory7:fedora.im>
14:33:15
Going thrice…
<@jflory7:fedora.im>
14:33:19
Oh!
<@jflory7:fedora.im>
14:33:30
!action @jflory7 Review @ffmancera's Council Docs PR
<@jflory7:fedora.im>
14:33:36
💥
<@jflory7:fedora.im>
14:33:49
!topic #486: Discuss and decide on a Fedora Council policy on the use of AI-based tooling for contributions
<@jflory7:fedora.im>
14:33:55
!link https://pagure.io/Fedora-Council/tickets/issue/486
<@jflory7:fedora.im>
14:34:14
Looks like most of this discussion is on the Pagure ticket, I forget if there is a corresponding Discussion topic
<@amoloney:fedora.im>
14:34:24
Ive just done that so one less action :) Thanks ffmancera too for the PR!
<@sumantrom:fedora.im>
14:34:26
i agree with dcantrell 's thoughts.
<@jflory7:fedora.im>
14:34:37
Happy for someone else to drive on this ticket too, if someone is ready to drive.
<@jflory7:fedora.im>
14:34:42
Thank you!
<@amoloney:fedora.im>
14:34:42
There should be...and if not we need one
<@jflory7:fedora.im>
14:35:09
Aha
<@jflory7:fedora.im>
14:35:10
!link https://discussion.fedoraproject.org/t/fedora-council-tickets-ticket-486-discuss-and-decide-on-a-fedora-council-policy-on-the-use-of-ai-based-tooling-for-contributions/108051
<@jflory7:fedora.im>
14:35:15
But no comments
<@jflory7:fedora.im>
14:35:18
Everything is on Pagure
<@dcantrell:fedora.im>
14:35:38
on this same topic I posted to council-private
<@jflory7:fedora.im>
14:35:53
Do we have a proposal?
<@jflory7:fedora.im>
14:36:06
I have seen a lot of ideas on what we should or might do, but I am not sure if it has been unified into a proposal
<@t0xic0der:fedora.im>
14:36:17
Justin W. Flory (he/him): I was tasked to work on a draft
<@jflory7:fedora.im>
14:36:22
I am not sure I am in favor of being so strongly discouraging to AI-based contributions. I am not sure I agree with the Gentoo approach.
<@jflory7:fedora.im>
14:36:35
Akashdeep Dhar: Ah. Any progress to report? I should probably assign this ticket to you then
<@t0xic0der:fedora.im>
14:36:59
On the basis of hte definition of contributors and contribution from mattdm
<@t0xic0der:fedora.im>
14:37:11
I have not gotten to it just yet
Once I am through with the release party, I plan to start working on it from the next week onwards
<@t0xic0der:fedora.im>
14:37:12
That is if it is still sought
<@dcantrell:fedora.im>
14:37:18
there are ongoing concerns that are still unanswered. the LLMs and their licenses are a giant "who knows?" right now
<@t0xic0der:fedora.im>
14:37:33
Justin W. Flory (he/him):
<@jflory7:fedora.im>
14:37:53
Post-Party is totally fine IMHO, especially given the party is in two days
<@dcantrell:fedora.im>
14:37:54
there are concerns around privacy and things like GDPR compliance
<@jflory7:fedora.im>
14:37:59
I would not expect a full draft before the party :)
<@jflory7:fedora.im>
14:38:18
From an engineering perspective, the inclusion in Fedora makes more sense. From a contribution POV, I am less confident.
<@jbrooks:matrix.org>
14:38:21
Have we spoken to our friends in RH Legal on this?
<@t0xic0der:fedora.im>
14:38:25
And I have not yet gotten to the Git Forge ARC investigation slide deck (God help me)
<@mattdm:fedora.im>
14:38:42
yes
<@jflory7:fedora.im>
14:38:52
Not officially. We might have spoken briefly with Amanda Newby but she recently departed Red Hat.
<@jflory7:fedora.im>
14:38:57
Oh, never mind?
<@jflory7:fedora.im>
14:39:05
I can't remember if I was there for that conversation or not
<@mattdm:fedora.im>
14:39:15
I spoke to Richard, too
<@amoloney:fedora.im>
14:39:25
Can we agree on some basics then please? I suggest as a starting point, can we agree that we *want* AI capabilities in Fedora first?
<@jflory7:fedora.im>
14:39:30
Deep breaths :) There is a lot going on right now.
<@ffmancera:fedora.im>
14:39:38
Even if we discouraged AI contribution, how can we know if something is AI generated? In Fedora during Outreachy mentorship contribution period we spotted some contributions that might be done with AI but it is difficult to know if it is really AI or not
<@jflory7:fedora.im>
14:40:00
This question makes me feel like we have different angles on this ticket even among our same group
<@jflory7:fedora.im>
14:40:14
The ticket, as I understood it, was about the acceptance of contributions that may have been derived from AI
<@jflory7:fedora.im>
14:40:19
But I see there is also a tooling perspective
<@amoloney:fedora.im>
14:40:24
I feel that too, thats why I want to find some common ground
<@jflory7:fedora.im>
14:40:25
Do we _use_ AI tools as aids in contributing?
<@mattdm:fedora.im>
14:40:31
As long as models are using real, open source licenses, legal isn't flagging any particular risk. It's a matter of policy.
<@mattdm:fedora.im>
14:41:06
I mean, some council members used ChatGPT to draft text in our Face to Face
<@jflory7:fedora.im>
14:41:13
Are we talking about including AI models in Fedora as software packages or some other format? Are we talking about the scrutiny applied towards contributions by our community that may have used AI tools? Are we talking about the use of AI tools in the contribution process of Fedora?
<@jflory7:fedora.im>
14:41:31
The first question feels out of scope for Council, and I would look to FESCo on leadership there
<@sumantrom:fedora.im>
14:41:32
mattdm: models can be (ie the base model) whatever goes into the training process, makes the world of difference
<@bookwar:fedora.im>
14:41:37
There is always Google Translate and the likes, which exist and used by contributors to help in communication. Any AI policy has to acknowledge that.
<@jflory7:fedora.im>
14:41:37
The second question is not a question I want us to solve
<@jflory7:fedora.im>
14:41:44
The third question, well, there is something interesting.
<@amoloney:fedora.im>
14:42:03
I agree, but we need to have a project policy first before we can ask them to weigh in
<@dcantrell:fedora.im>
14:42:13
how are the second and third questions different?
<@jflory7:fedora.im>
14:42:17
The second question is not a question I want us to solve (because it creates a lot of artificial work for us and we implicitly are telling people to start being adjudicators of whether someone is using AI or not – possible impact on CoC cases)
<@amoloney:fedora.im>
14:42:19
I agree, but we need to have a project policy first before we can ask them to weigh in - re:FESCo leadership
<@mattdm:fedora.im>
14:42:42
FESCo was already asked this and kicked the question to legal, who basically sent it back to _us_.
<@jflory7:fedora.im>
14:43:09
Argh 🤦
<@amoloney:fedora.im>
14:43:12
so lets pair it all the way back please! Do we want AI in Fedora? If yes, lets look at the use cases we need to cover when wordsmithing a policy
<@amoloney:fedora.im>
14:43:38
Use Case #1: AI applications in the OS
<@bookwar:fedora.im>
14:43:49
Proposal: do not write "AI policy", write "guidelines and recommendations on using AI in Fedora contributions".
<@dcantrell:fedora.im>
14:43:53
I'll give a specific thing I don't want. I don't want to see a hundred PRs to fix a CVE where the code was produced by a generative AI system.
<@amoloney:fedora.im>
14:43:54
Use Case #2: AI tooling for contribution aids
<@amoloney:fedora.im>
14:43:57
etc
<@amoloney:fedora.im>
14:44:09
Yes to this
<@jflory7:fedora.im>
14:44:18
The second question, to me, is about the scrutiny used towards accepting contributions that we don't necessarily know the source. For example, Outreachy applicants who may have submitted contributions that used AI tools to help them organize or present their contributions.
The third question is about tools. Do we use AI tools and models to help us be more effective at contributing to Fedora? The third question is less about the contributor/person, and more about the tool/software.
<@amoloney:fedora.im>
14:44:41
Yes to this - and this can be applied to those use cases we identify
<@dcantrell:fedora.im>
14:44:50
ok, fair
<@jflory7:fedora.im>
14:45:12
I like this as a starting place for anything we produce here. I think this is something that many people (everyone?) can agree on.
<@bookwar:fedora.im>
14:45:32
It is an emerging field and emerging technology. Writing strict formal rules for things which we do not fully understand or can not yet predict is hard. I would start with more generic concepts and guidelines, and then address issues as they happen when we see the cases of abuse of the AI in the way we want to prevent.
<@jflory7:fedora.im>
14:46:06
I am completely onboard with you there bookwar
<@mattdm:fedora.im>
14:46:09
The second question splits even further, I think:
a. _direct_ use (an outreachy contributor using chatgpt because they're not confident in English, plus _also_ those patches david is worried about)
b. _upstream projects that have used generative ai_
<@jflory7:fedora.im>
14:46:57
I honestly do not want us to answer the second question. Or at least, not right _now_. We might have to eventually. But I don't like it because I think it inherently is saying that we don't trust our contributor community. Or it encourages distrust instead of collaboration and unity.
<@jbrooks:matrix.org>
14:47:08
The output of gen AI isn't viewed as copyrightable currently. In order for it to be licensed as open source, it needs to be a also human edited, so, AI tools could be used in the process of producing a contribution, but in the end it needs to be a person's work, which is important anyway.
<@mattdm:fedora.im>
14:47:12
another thing we're seeing for real (and is an escalating problem): AI-assisted forum span.
<@jflory7:fedora.im>
14:47:14
We have enough polarization in our overall global state of affairs :P
<@mattdm:fedora.im>
14:47:44
Well, I think David's concern is squarely in this area.
<@mattdm:fedora.im>
14:48:31
We can avoid making strong dictates here, but I think this is the part we really need to address.
<@dcantrell:fedora.im>
14:48:41
I trust the contributors and community members, I don't trust the AI tools in question.
<@jflory7:fedora.im>
14:48:45
Is there anything I could do to persuade that answering the third question (tools/software using AI) sufficiently addresses the second question (contributions/contributors using AI)?
<@jbrooks:matrix.org>
14:49:25
We don't trust stackoverflow necessarily, but we trust community members to use it as they produce their contributions.
<@t0xic0der:fedora.im>
14:49:27
Totally okay with writing guidelines instead of policies here on my end
<@mattdm:fedora.im>
14:49:31
Can you give an example of what that answer might look like? Is it a list of approved tools?
<@jflory7:fedora.im>
14:49:50
Why this and not #3? Also, assume we all have limited time to push this forward and we have to prioritize what is the _most urgent_ question or angle to answer.
<@bookwar:fedora.im>
14:50:02
mattdm: I like this example because it is already happening, we know that this is happening, and we know the damage it creates. So it is easy to write a rule: "do not produce AI-assisted spam in forum or other communication channels(issues, trackers, etc)". It is not a complete AI policy, but rather a scoped rule to apply in specific area.
<@amoloney:fedora.im>
14:50:03
With 10 mins to go on this meeting, we need something actionable from this
<@jflory7:fedora.im>
14:50:04
++
<@jflory7:fedora.im>
14:50:24
Are we OK with things like https://log-detective.com/ ?
<@amoloney:fedora.im>
14:50:29
we had this 2 weeks ago and came out with not a lot and the tech is moving much faster than we are
<@jflory7:fedora.im>
14:50:44
I don't think we have to approve _every_ tool, that would be a lot of unnecessary bureaucracy
<@jbrooks:matrix.org>
14:50:47
Re AI-assisted spam, how about, don't post spam?
<@jflory7:fedora.im>
14:50:58
Thanks for the time check Aoife
<@dcantrell:fedora.im>
14:51:05
stackoverflow is a good example, a developer using it as a tool to understand something is different than relying on it as a black box and submitting the output as a PR
<@amoloney:fedora.im>
14:51:18
Were talking close to a month about this so...
<@bookwar:fedora.im>
14:51:22
I would even just add it to the Code of Conduct
<@jflory7:fedora.im>
14:51:23
I wonder if a Miro board exercise would be good for this
<@dcantrell:fedora.im>
14:51:27
we should have a policy that bans spam
<@t0xic0der:fedora.im>
14:51:29
How do we recognize a spam? What's the objective description of a spam?
<@jflory7:fedora.im>
14:51:31
Or something to organize our thoughts better
<@jflory7:fedora.im>
14:51:38
Each month, it feels like the scope gets wider 😅
<@jbrooks:matrix.org>
14:51:41
Just as you must understand what you read on stackoverflow, you must understand what comes out of the AI
<@amoloney:fedora.im>
14:52:01
if it helps progress these guidelines we are all agreeing we want, then its worth a shot!
<@jflory7:fedora.im>
14:52:07
+1
<@bookwar:fedora.im>
14:52:18
You recognize spam when you see it. There is no point in writing a formal definition for it.
<@moralcode:fedora.im>
14:52:36
"must be able to explain the reasoning behind aspects of your contribution/why you did what you did" soudns like how a CS prof would reasonably implement an AI policy in teh classroom and may be helpful here
<@jflory7:fedora.im>
14:52:37
I feel like we are still not on the same page about what we want to solve/address. I feel like if we get unified on that, our thoughts are more aligned than they might seem right now
<@jbrooks:matrix.org>
14:52:42
What I'm saying RE spam, is in the example of AI-generated garbage forum posts, the problem isn't the AI, it's the garbage posts, posted by a person who needs to be responsible for what they post, no matter how it was written.
<@dcantrell:fedora.im>
14:52:47
understanding what comes out of a generative AI system....that's what I'd like to read in a policy. we should consider them tools, not replacements for contributors
<@jflory7:fedora.im>
14:52:54
But the surface area is so vast, and ultimately, we are a small group of people with limited time to spend on this specific challenge right now. There are so many challenges!
<@jflory7:fedora.im>
14:52:59
We have to pick our battles
<@jflory7:fedora.im>
14:53:09
Rome was not won in a day, etc etc
<@mattdm:fedora.im>
14:53:29
Throwing another thing into the mix: https://simonwillison.net/2024/May/8/slop/
<@t0xic0der:fedora.im>
14:53:29
I disagree. It might be perceived exclusionary if we cannot objectively define what a spammy PR looks like. Some folks might be exposed to AI based creations and can decide for themselves but others might not. As such, with an evolving tech like that of AI - we need to tread carefully.
<@jflory7:fedora.im>
14:53:55
This is why bookwar's idea about calling it out in the CoC seems wise and the best possible way to address that angle. I really liked _this specific thing_ in the CoC
<@mattdm:fedora.im>
14:53:57
"Not all promotional content is spam, and not all AI-generated content is slop. But if it’s mindlessly generated and thrust upon someone who didn’t ask for it, slop is the perfect term for it."
<@bookwar:fedora.im>
14:53:59
Yes, but also it helps to call out certain typical behaviors on the code of conduct even when it is generally just variations on the generic "respect each other" rule
<@t0xic0der:fedora.im>
14:54:18
Agreed.
<@jflory7:fedora.im>
14:54:35
Aoife Moloney: Post-F40 Party, want to tag-team on a Miro board exercise with the Council?
<@jflory7:fedora.im>
14:54:44
To try and bring us to a unified point on this
<@amoloney:fedora.im>
14:54:45
To play devils advocate - right now we have *nothing* as a project on AI guidelines so this whole debate on Whats AI spam/spam, etc is a waste of time
<@amoloney:fedora.im>
14:54:52
we need something
<@amoloney:fedora.im>
14:54:57
we can build on a something
<@amoloney:fedora.im>
14:55:06
all we are getting right now are peoples concerns
<@t0xic0der:fedora.im>
14:55:12
Want me to hold off on the writing before we get through this
<@jflory7:fedora.im>
14:55:22
Probably yes!
<@jbrooks:matrix.org>
14:55:23
We should certainly require that contributions be authentic to the contributor, but I do think that an AI tool can be used to aid in creating authentic contributions
<@sumantrom:fedora.im>
14:55:24
There is a spam bot we use in Wordpress to collect spam comments.. it's not AI but when the plugin is suspicious , it marks it for moderation.
<@t0xic0der:fedora.im>
14:55:28
(Question mark in the end)
<@t0xic0der:fedora.im>
14:55:34
Roger that
<@mattdm:fedora.im>
14:55:42
I'd actually like you to _not_ hold off, because that'll give us a starting point
<@jflory7:fedora.im>
14:55:52
(I also suspect we are having some Matrix message delivery delays happening right now, that might be causing havoc in this conversation)
<@jflory7:fedora.im>
14:56:15
I am thinking Aoife Moloney and I could work on this together though
<@mattdm:fedora.im>
14:56:16
The same in Discourse (also akismet), but: generative-ai spam is particularly good at evading detection with this method.
<@jflory7:fedora.im>
14:56:21
I don't think we are all starting from the same place
<@t0xic0der:fedora.im>
14:56:29
You know what?
I will start off with a draft - nothing too polished or nothing too final
<@dcantrell:fedora.im>
14:56:45
I usually start from GO on the Monopoly board
<@mattdm:fedora.im>
14:56:45
It is also more work for _human_ moderators, because it can look relevant and often even talks about Fedora-specific things and references the subject of the post.
<@t0xic0der:fedora.im>
14:56:49
On the draft?
<@jflory7:fedora.im>
14:56:57
You can start with something, but don't invest _too_ much time into it? I don't want you to spend a day on this and then we decide we want to go in a different direction.
<@jflory7:fedora.im>
14:57:16
No, on a group exercise to figure out which problem/challenge on AI we want to solve, and why that problem/challenge
<@jflory7:fedora.im>
14:57:29
I think a policy will come easier once we figure _that_ part out
<@mattdm:fedora.im>
14:57:30
I think that's probably a sign that the Fedora community _in general_ won't be on the same place on this.
<@t0xic0der:fedora.im>
14:57:47
I don't mind spending a no-meeting Friday on it.
Or at least half of it. Should help us with finding stuff to further read from.
<@t0xic0der:fedora.im>
14:57:57
Just not this Friday though
<@t0xic0der:fedora.im>
14:58:04
Got it
<@jflory7:fedora.im>
14:58:06
I think that is why we need a unified Council to be advocates for what we think is right!
<@jflory7:fedora.im>
14:58:10
Right now, we are not all aligned
<@jflory7:fedora.im>
14:58:19
So, how can we expect the community to be aligned either? :)
<@dcantrell:fedora.im>
14:58:34
good point
<@jflory7:fedora.im>
14:58:37
I do think a bullet-point list would be helpful!
<@bookwar:fedora.im>
14:58:49
On the but slightly relevant topic, as it seems we don't get to Open Floor today :) As I announced in Council channel, I want to run a Survey again, this time in July. So I have a "Call For Questions". If you want to ask Fedora community about something as a part of the Survey, please send PRs to Council Docs https://pagure.io/Fedora-Council/council-docs/blob/main/f/council/modules/ROOT/pages/procedures/survey/questions.adoc before June 20.
<@jflory7:fedora.im>
14:58:50
References/real-world examples too
<@mattdm:fedora.im>
14:58:52
We can advocate, but we shouldn't set policy that doesn't represent community consensus.
<@jbrooks:matrix.org>
14:59:01
Maybe this already happened, but this sounds like a good async convo
<@bookwar:fedora.im>
14:59:13
This can be the AI question. If you come up with the wording for it.
<@mattdm:fedora.im>
14:59:34
I was just talking to Aoife Moloney about this yesterday :)
<@amoloney:fedora.im>
14:59:43
Yes this definitely is a good way to get real feedback from the community
<@jflory7:fedora.im>
14:59:54
I think this is causing more problems than we might realize during this meeting
<@jflory7:fedora.im>
15:00:14
I am going to pre-emptively take this on with Aoife Moloney for after the F40 party and put some time on our calendars to develop it together.
<@amoloney:fedora.im>
15:00:35
we were talking about having a dedicated survey for AI, but including one or two more general questions on the subject in the contributor survey
<@mattdm:fedora.im>
15:00:40
I think we should have 1 or 2 AI questions in the general survey, but also consider doing a more in-depth survey about AI specifically.
<@jflory7:fedora.im>
15:00:42
!action @jflory7 Set up time with @amoloney next week to design a Miro exercise on AI/ML questions in Fedora for the Council to go through together
<@mattdm:fedora.im>
15:00:44
yes that :)
<@jflory7:fedora.im>
15:01:21
!idea Run a exercise with the Fedora Council to help us get unified on which question/challenge on AI/ML is most important for Fedora to answer today, then focus on how we build/execute our answer to that question.
<@jflory7:fedora.im>
15:01:28
!topic Open floor
<@mattdm:fedora.im>
15:01:33
and, time! obviously a lot more to talk about here. Justin, thank you for taking that on.
<@jflory7:fedora.im>
15:01:38
!info @bookwar: On the but slightly relevant topic, as it seems we don't get to Open Floor today :) As I announced in Council channel, I want to run a Survey again, this time in July. So I have a "Call For Questions". If you want to ask Fedora community about something as a part of the Survey, please send PRs to Council Docs https://pagure.io/Fedora-Council/council-docs/blob/main/f/council/modules/ROOT/pages/procedures/survey/questions.adoc before June 20.
<@jflory7:fedora.im>
15:01:51
And with that, we are at time.
<@jflory7:fedora.im>
15:01:55
I am going to cut it here.
<@amoloney:fedora.im>
15:01:57
Thanks Justin! Appreciate the driving assist today!
<@amoloney:fedora.im>
15:02:03
Justin W. Flory (he/him): ++
<@jflory7:fedora.im>
15:02:06
Thanks folks for a packed meeting! And welcome to our newest Initiative! 🎉
<@zodbot:fedora.im>
15:02:07
amoloney gave a cookie to jflory7. They now have 358 cookies, 9 of which were obtained in the Fedora 40 release cycle
<@jflory7:fedora.im>
15:02:16
Jason Brooks++ Thanks for helping drive this one :)
<@jflory7:fedora.im>
15:02:20
Until next time 👋
<@jflory7:fedora.im>
15:02:22
!endmeeting
<@zodbot:fedora.im>
15:02:22
jflory7 has already given cookies to jasonbrooks during the F40 timeframe
<@zodbot:fedora.im>
15:02:22
jonatoni has already given cookies to jflory7 during the F40 timeframe