[14:24:48] tizianop: hey [14:25:35] tizianop: I checked out the README file, but it doesn't have too much info. You mention the point F in your email. Is it in README? [14:38:30] bmansurov: ooops, wrong file, sorry. Check you email [14:38:37] thanks [17:35:31] hey miriam_. how was your panel in BBC yesterday? :) [17:36:37] aloha [17:36:44] aloga leila [17:36:55] aloha leila [17:37:59] leila: so it was nice. There was one guy from UCL's NLP team. And some 'tech and politics' expert from wired magazine [17:39:32] we discussed about the future of machine learning (we were asked for '2018 predictions'). We also had a very interesting meta-discussion on the meaning of algorithmic explainability and the use of adversarial machine learning for that. [17:40:41] miriam_: sounds like it was a good visit, at least for keeping connections, creating new ones, and thinking with others on these topics. :) [17:41:28] leila: yes - going for a coffee next week with the head of data science (she rocks!) to discuss more about overlapping areas between us and them. [17:41:45] that's great. :) [17:43:04] leila: yes \o/ [21:46:14] halfak: does Technology have 1 Goal we are all defining programs under? remember how Audiences has 1? or is it that for Technology each team defines 1+ goals? :D [21:46:25] (I swear this is not a meta question;) [21:47:04] leila, I don't think we have anything like that. [21:49:09] halfak: interesting. In retrospect, I think it would have been a nice exercise for the whole Tech to try to have one Goal, it would help us arrive at a more shared understanding on what we as a whole do in Tech. [21:51:36] leila, IMO these "one goal" exercises are just for presentation purposes. We could still have one goal. But it would be silly to have so many teams focus on one thing -- especially when we're planning for a multi-year trajectory where building up various infrastructures will be key to long-term success. [21:52:03] Even with Product, I expect that they'll still do what they think is important regardless of "one goal"-ness. [21:52:19] But it certainly does help to describe how what they want to do might contribute to a common cause [21:54:30] halfak: I think the importance of such exercise is beyond communication to the outside of world. If done right, arriving to that one goal means the whole department needs to sit together and think through all the different things they're doing and see if there are common patterns/directions, things they should consider not doing, things that they should consider adding. That's a super useful exercise in terms of alignment. [21:55:19] It's hard to imagine how a systemic intervention my team is working on could share a goal with the concerns of the SRE or security teams. [21:55:32] halfak: one things which is I guess a barrier for Tech doing that kind of exercise is that we need to know the Audiences outcomes, outputs, etc. /before/ we can have that conversation within Tech, as what is decided in Audiences will have impact on SRE for example. [21:55:36] Unless we choose something really abstract like "Evolving our infrastructure" [21:56:04] leila, while it will impact, I think a lot of our work gets ahead of Audience/Product. [21:56:24] Yeah, I mean a goal at the department level will be a bit more high level for sure, and then the outputs will define the products and services that feed into that goal. [21:56:51] leila, hard to imagine how that would work. [21:57:00] halfak: I agree, but because part of the work can't get discussed without having Audience plans, we can't really sit as a team and discuss until Audience plans are in. [21:57:25] on the contrary, we know a lot of what tech needs at the WMF and product can take advantage of that next cycle. [21:58:17] halfak: Ok, that can be the case, too (at least in research some of that is definitely true), then we need to have many smaller iterations (something Katie brought up at All-hands). [21:58:29] Smaller iterations? [21:59:58] halfak: yeah. (her comment was in the context of 3-5 years plan, though). Basically, each department creates a first draft of AP, then we talk with each other, and then we iterate and update the AP. [22:01:54] leila, IMO, the AP as a point of planning what we *do* is an anti-pattern. Many projects take more or less than a year. We should constantly be adjusting our notions of what we want to do. The AP is a chance to ask "How well are we resourcing what we want to do?" [22:02:17] So in that respect, I agree on quick iterations, but it's very waterfall like to plan out details at AP time. [22:04:39] halfak: yeah, I'm with you. I'm waiting for the 3-5 year planning cycle, cuz that would help address the issue you've pointed out. [22:05:05] We ought to be working on it now. It's silly not to. [22:05:13] * leila tries to walk away from IRC and write down the Goal, Outcome, Output, Resources, Target. :D [22:05:16] All of my plans are 3-5 year plans ;) [22:05:27] halfak: DO NOT SAY THAT. [22:05:32] :P [22:05:32] Thinking we can write up what we'll actually do in 3-5 years is crazy pants. [22:05:35] Any any point in time [22:06:00] Working out critical directions and talking about how we'll find resources for them along the way, now that I can get behind. [22:06:02] halfak: yeah, I mean in our case it's like that. pretty much no major effort ends in 1 year. [22:06:10] Right. [22:19:49] leila, Maybe we can start by talking about what 3-5 years looks like from our perspectives and seeing if a common goal makes sense. [22:21:58] halfak: The AP is a chance to ask "How well are we resourcing what we want to do?" is EXACTLY what I think too [22:22:21] :D! [22:22:26] * halfak high fives [22:37:52] halfak: I think that's a good exercise.