| 20 Mar 2023 |
 Alyssa Ross | i too am a linus fan | 21:13:48 |
 infinisil | In any case, feedback appreciated, there are unfortunately some downsides to the repo approach | 21:13:55 |
| infinisil | Just posted https://discourse.nixos.org/t/looking-for-somebody-to-take-meeting-notes-for-the-nixpkgs-architecture-team/26548 | 22:39:43 |
| 21 Mar 2023 |
 Growpotkin | I added another goofy example to the working group draft just to highlight that "we are just recruiting folks to work on this, the interface itself isn't settled, you don't need to know anything about drv-parts or have a panic attack if you aren't a fan of drv-parts" ( not that anyone ought to have an issue with drv-parts ). Made a few other minor tweaks to phrasing of some stuff; but it looks ready to post in my opinion. | 00:26:58 |
|  @teutat3s:pub.solar joined the room. | 07:55:22 |
 profpatsch | linus is the best <3 | 09:00:30 |
| 22 Mar 2023 |
 davidak | FYI: I tested to create meeting notes from an audio recording using OpenAI Whisper and ChatGPT.
https://discourse.nixos.org/t/2023-03-20-nixpkgs-architecture-team-meeting-33/26547/2?u=davidak | 02:27:21 |
|  snowytrees joined the room. | 03:35:14 |
| snowytrees | Im interested in joining the wg for package modules. I am looking to get involved with more formal nix teams. I have experience with the module system, and hacking around normal derivations to make them configurable. Two questions. How long do WG tend to last for? And how regular are the “regular meetings”, weekly? | 03:49:51 |
| infinisil | snowytrees: Thanks for the interest! I don't think there's an answer to how long WG's tend to last, it really depends on many things and I don't have data on previous WG's. Regular meetings I think should be either weekly or bi-weekly, depending on how often they are needed. The main work should happen between meetings though | 08:21:50 |
| infinisil | Of course this is flexible, whatever works best for the people in the WG to achieve progress | 08:24:12 |
| Alyssa Ross | davidak: does this run offline, or does it involve sending our recordings to OpenAI? | 09:26:44 |
 raitobezarius | In reply to @qyliss:fairydust.space
davidak: does this run offline, or does it involve sending our recordings to OpenAI? Whisper run offline | 09:34:43 |
| raitobezarius | The weights are public; training set no | 09:35:03 |
| Alyssa Ross | And the summarising? | 09:35:22 |
| raitobezarius | GPT-4 is not public | 09:35:24 |
| raitobezarius | It involve sending the transcript data to OpenAI | 09:35:35 |
| Alyssa Ross | davidak: then please don't do that again without asking the participants | 09:35:50 |
| raitobezarius | (and requires beta access /whatever ATM) | 09:35:54 |
| Alyssa Ross | I don't want to be training data for OpenAI | 09:36:02 |
| Alyssa Ross | we have other community members that feel similarly strongly | 09:36:09 |
 tea | In reply to @qyliss:fairydust.space
I don't want to be training data for OpenAI they claim they don't do that | 11:05:30 |
| tea | not sure that this is the right place to ask, but why do we mark abandoned PRS/issues as stale in nixpkgs repo, and not close them? | 11:06:23 |
 K900 | Do you mean the "stale" label? | 11:08:11 |
| K900 | They're not closed because "a robot closed a PR I care about" is generally not a nice experience | 11:08:31 |
| K900 | And having the number lower isn't worth alienating potential contributors | 11:08:40 |
 @piegames:matrix.org | Also, "this has not been touched in a while" turns out to be a pretty bad metric for deciding to close something | 11:12:38 |
| @piegames:matrix.org | Feel free to use the stale label as an invitation to triage issues and close those we don't need anymore | 11:13:00 |
| tea | In reply to @k900:0upti.me
Do you mean the "stale" label? yes | 11:14:32 |
 Wanja Hentze | stalebotting is awful, I think we should be doing it even less | 11:17:23 |
