Video summary
The meeting focused on preparing for the upcoming All Members Meeting scheduled for October 28th and aligning these efforts with the Technical Advisory Committee (TAC) reporting process. The primary goal is to consolidate information from various working groups into a cohesive presentation that can be used by both the All Members Meeting and the TAC, thereby avoiding duplicate work and ensuring consistent sources are utilized. To achieve this, the organizers proposed a new approach where chairs of working groups would be responsible for gathering structured data regarding their group's highlights, deliverables, community adoption, progress, and overall health. This shift aims to move away from the current method where specific individuals manually compile notes from every meeting, a process that often fails to accurately reflect the output of each group and creates last-minute scrambling.
To facilitate this data collection, the team discussed utilizing an AI pipeline or specific skills to automate the synthesis of information from various sources like Confluence wiki pages and GitHub repositories. While an automated system is considered a midterm goal for creating a more robust knowledge base that cross-references projects across the organization, the immediate proposal allows chairs flexibility in how they generate the report. Chairs can either manually fill out structured fields or use an AI tool to ingest their group's documentation and synthesize it into the required format. The organizers emphasized that this task should not take more than an hour with AI assistance and is viewed as a core responsibility for chairs rather than an additional burden, ensuring that working groups actively demonstrate their status and activity levels to potential members.
The discussion also highlighted the broader challenge of maintaining organized reporting within the Linux Foundation, noting that messy information has been a consistent issue across different projects. One participant compared the effort required to clean up data and structure reports to the unglamorous but necessary work of weeding a garden; while it is not exciting or flashy, it is essential for the ecosystem to function properly without falling apart. The consensus was that while the current state of reporting may seem disorganized, addressing these foundational issues is critical for long-term success. Consequently, the proposal will be presented at the steering committee meeting with the understanding that chairs are not required to use a specific AI agent but can choose their preferred method as long as the information is structured and submitted effectively.
Read the full video transcript
Hello.
>> Hello, Sashan.
Yeah,
I left the meeting. I I don't know who
Drummond is interested in this meeting,
but he's away, of course. Um, but I did
need to
uh work here uh get some work done here
because the upcoming TAC reporting. Yay.
So, I'm going to share my screen
to talk about um
with Tomas. We had a meeting to talk
about
um
the There we go. to talk about the
upcoming all members meeting that's
October 28th
and we are now trying to have we're
we're soliciting our chairs for
information about their working groups
and uh hello Lynn um I'm just talking
about uh the up
>> yeah so the all members meeting is
coming up October 28th at the same time
we are going to prepare our TAC
reporting the technical advisory
committee because then if everybody's
collecting information for the all
members meeting then we can double it up
with the TAC and use the same sources in
order to come up with the TAC report.
[sighs and gasps]
So, I had a conversation with Tomz this
morning because he can't make it to the
steering committee meeting this
afternoon.
And Tomz and I talked about the approach
for how to get um how to get members to
contribute to the all members meeting
and contribute information to the TAC.
Now, uh, of course, it would be nice to
have an AI pipeline that we could
automatically use for generating this
information, but we don't have that.
Don't have that yet.
So,
instead then, um, so I asked this group
what they think. We were going to ask
chairs and stuff to pull together
information to put it in the slide deck
for the presentation of the all members
meeting.
And then to either request that they
provide
um where their resource where their work
is being documented
or then potentially to also share this
skill to say this is what we are looking
for and if you as a chair can run this
skill or something similar to it
with your working group's content events
and give us the reporting afterwards and
it'll also help you uh um figure out
what you want to report in the
in the all members meeting.
So that's what I was going to that's
what I proposed to this group for
feedback and then proposed to the
steering committee for feedback and then
include in the email that Thomas is
going to send to the chair's mailing
list. [snorts]
This sounds simple enough for block is
the place where this will get blocked.
>> Sorry, it sounds simple enough for what?
>> I meant that this this plan is simple
enough that you ask the chairs for a
structured set of information. provided
and that gets rolled into some sort of a
slide that gets presented at the all
members and the tab.
>> Yeah.
>> But I suppose that even though it sounds
simple on the surface, there is probably
some blockers in there and I wanted to
figure out the project,
>> right? I could also ask in in the email
that gets sent out, I could also ask for
feedback on the process
and how we could improve it. Because
right now, I mean, [laughter] our chairs
have had it kind of easy because it's
been Drummond and I going through
everybody's meeting notes and trying to
assemble it all and and that's not
necessarily reflective of what each
group produces and it would be better if
each group had their say in their stuff.
>> Okay.
>> Yes. Sorry, then go ahead.
I was just I was just wondering if uh
can't AI somehow
bring these things all together and put
it into
>> Yeah. So that's the skill. This is the
AI prompt.
>> Ah, okay. Gotcha. I I see it.
>> Yeah. So it would be like if you did do
this on your on whatever AI account you
own, put in the stuff and run this
prompt.
And uh uh
yeah, run it as a prompt or add it as a
skill. But yeah, just run it as a prompt
is just as well. And uh
yeah, I I mean when I do it, I give it
the files and then I give it the skill
file and I say use this skill and
generate for me my annual review report.
Okay.
Thank
>> Yes. Again, I mean, so let me be a bit
cruel over here because uh I I think as
a if someone is a chair of a working
group, this should be part of the
responsibility. This is not something
that uh an additional stuff anyone is
asking them to do. So
>> yeah,
>> with that with that out of the way, I
think that now that you have a set of
structured fields against which you
require them to provide information.
In many cases, the chairs shouldn't be
actually having to refer back to
consuants and all of that because they
likely are more familiar and can put in
the information without having to go
through all the wiki pages and the
GitHub issues and all of that. Yep. Yep.
Yep. Okay. So, the the offer then what I
could do is say
um
Right.
Right. You you can enter the data
manually.
>> What are your highlights?
What are your deliverables?
What are is your community in adoption?
What is your progress?
How is your project health? Fill that
out
instead manually or use an AI.
>> I mean, if they if they want to download
all the Confluence wiki pages, feed it
to an LLM that they have, point the LLM
to the GitHub repo of their working
group and say, synthesize all of this
into this structure and so that I can
send it back to Carly. Yeah,
>> that's fine. Nobody's objecting to that.
But all I'm saying is that
>> this doesn't seem to be like something
that would take an entire day of work.
This should likely take no no more than
an hour of work if you're using an AI
and probably two if you're not using an
AI.
>> Yeah.
Yep.
Okay. So the the proposal will be this
is a regular expectation of the working
group chairs.
Uh um we are asking for a TAC report.
You can either generate the information
manually if you would like. You could
use the skill as well like you can also
you can also pull in the skill and then
say ask me questions so that I can
manually give you the information
and and then synthesize it into the
report. I mean like
>> people are smart. They can figure it out
>> and and
>> I mean
>> all I'm trying to say is that what you
and Thomas are trying to do by using a
LFDTwide AI system to be able to make
this available to everyone. I think that
is a midterm goal. In the near term,
there are enough simple smart ways to
continue to do the reporting without
having you and Drummond do last minute
scrambles, which I'm sure you you want
to avoid.
>> Yeah. Yeah. And you're right, that's a
uh the the the LFDT pipeline that we're
talking about is uh um would be a a lot
more than just generating a report
because then it might look to say, oh, h
how do these how do the different where
where are the pro where are the projects
talking about the same things? Where
could you know that kind of stuff? Yeah,
I I think the LFDD pipeline that you are
discussing has the potential of
providing enough cross referenced cross-
linked information that becomes
knowledge whereas the current state is
mostly information that gets broadcasted
like I want to give my audience a status
of where this working group is so that
if you are interested in joining trust
over IP and working on this topic you
know that we are not a dead working
group we actually a very active one
exploring very interesting topics.
>> Yeah. Yeah. Okay.
Okay. So I will come with that proposal
uh talk about it with mentioned to Tomas
just that in the email that to not
require
that they use an AI agent. just however
you get the information to me and if you
wish to use a an AI skill here's one
that has been drafted.
>> Okay.
>> I mean I think that is way more generous
and sufficient at this point if somebody
also gets a AI skills to use at that
point a lot of their how am I going to
do it question is also resolve
developments right
>> you have a skills to start off with.
Yep. Good. Okay. And I will bring this
up at the
steering committee meeting tonight.
Okay. Any other comments
because otherwise I'm going to close the
meeting because got
>> I just wanted to say that I I watched
the TAC meetings that I published on the
YouTube channel.
>> Yeah. Yeah.
>> And I realized that
slightly messy information or slightly
messy reporting seems to be a hallmark
across LF. It's not something that is
only a symptom that we see in trust
overall. I'm not going to say that it
makes me happy, but I'm I'm going to say
that it just seems to be the way things
are at this point. [laughter]
>> Yeah. because it it's it's work to
garden and it's not it's not sexy or
lashy.
>> Yeah. But if you have if you have a
garden, weeding is always not the
glamorous bit of gardening.
>> Yeah. [laughter]
[gasps] Yeah. But you got to do it
or it falls apart. Okay, then. Uh that's
everything. Thank you so much. Thank you
so much.
>> Hey, bye.