[2026-11-19] eagle: personal/projects/personal-os/personal-os-evidence/2026-11-19/follow-up-live-acceptance.md personal/projects/personal-os/personal-os-production-acceptance.md personal/projects/personal-os/personal-os-remediation-plan-2026-11-18.md work/zoom-transcripts/2026-10-16 All-Company Meeting 2026-10-16 21 31(GMT+6 00) klekRLtLSEuMv0PCCPCjtA/transcript.json work/zoom-transcripts/2026-10-16 All-Company Meeting 2026-10-16 21 31(GMT+6 00) klekRLtLSEuMv0PCCPCjtA/transcript.md work/zoom-transcripts/2026-11-03 Michal Smaga's Personal Meeting Room qQ2bzU5KTiCGluiBYA-ROA/transcript.json work/zoom-transcripts/2026-11-03 Michal Smaga's Personal Meeting Room qQ2bzU5KTiCGluiBYA-ROA/transcript.md work/zoom-transcripts/2026-11-17 Apple Weekly Team Meeting - Zoom ID GJhjFoDDRY6PkqEUdM-YFg/transcript.json work/zoom-transcripts/2026-11-17 Apple Weekly Team Meeting - Zoom ID GJhjFoDDRY6PkqEUdM-YFg/transcript.md work/zoom-transcripts/index.json
This commit is contained in:
1 parent
192aaee4be
commit
d6c22b3357
11 files changed
+462
-277
No files matched your search
+7
@@ -163,3 +163,10 @@
|
|||||||
- На предыдущем коде клиентский флаг ручного повтора изменил счётчик той же версии с 3 на 1 и повторно вызвал подготовку. В двух изолированных браузерных проходах текущий рабочий обработчик вернул HTTP 409 на четвёртый запрос, сохранил три попытки и отклонил конкурирующий запрос. При остатке четырёх минут запуск новой версии удержан. После создания рабочего резерва изменённый источник запустил отдельный черновик со штатным Qwen. Искусственными были три исходных тайм-аута, сами решения о допуске и новая генерация выполнялись штатными компонентами. Рабочий портал после загрузки кода отдаёт план и очередь; открытых карточек с исчерпанным лимитом в срезе нет.
|
- На предыдущем коде клиентский флаг ручного повтора изменил счётчик той же версии с 3 на 1 и повторно вызвал подготовку. В двух изолированных браузерных проходах текущий рабочий обработчик вернул HTTP 409 на четвёртый запрос, сохранил три попытки и отклонил конкурирующий запрос. При остатке четырёх минут запуск новой версии удержан. После создания рабочего резерва изменённый источник запустил отдельный черновик со штатным Qwen. Искусственными были три исходных тайм-аута, сами решения о допуске и новая генерация выполнялись штатными компонентами. Рабочий портал после загрузки кода отдаёт план и очередь; открытых карточек с исчерпанным лимитом в срезе нет.
|
||||||
- При чтении живых данных предыдущим интерфейсом из дневного уведомления открывались выпадающая шкала 1–5 и отдельный флажок «Облегчённый день». Текущая ссылка ведёт к единственной карточке самочувствия на «Сегодня». В двух размерах окна проверены шкала 1–10, пять вариантов настроения и отсутствие старых полей и кнопок ручного перерыва. На изолированном контуре при коррекции отправлены только фокус и причина. Энергия, настроение и журнал отметок сохранили прежние значения.
|
- При чтении живых данных предыдущим интерфейсом из дневного уведомления открывались выпадающая шкала 1–5 и отдельный флажок «Облегчённый день». Текущая ссылка ведёт к единственной карточке самочувствия на «Сегодня». В двух размерах окна проверены шкала 1–10, пять вариантов настроения и отсутствие старых полей и кнопок ручного перерыва. На изолированном контуре при коррекции отправлены только фокус и причина. Энергия, настроение и журнал отметок сохранили прежние значения.
|
||||||
- В Codex обновлено действующее расписание `personal-os` для этого же чата. Интервал 15 минут сохранён как в настройке, так и в записи приложения. Первый следующий запуск назначен на 18:26:55 по Бишкеку.
|
- В Codex обновлено действующее расписание `personal-os` для этого же чата. Интервал 15 минут сохранён как в настройке, так и в записи приложения. Первый следующий запуск назначен на 18:26:55 по Бишкеку.
|
||||||
|
|
||||||
|
## Вечерние действия и восстановление чтения Zoom
|
||||||
|
|
||||||
|
- В 18:27–18:34 рабочий план до 19:00 удерживал предложения по причине `discussion_zoom_incomplete`. Календарь Asana имел статус `fresh`. Квитанция Zoom от 18:18 показала 86 найденных заметок и остановку на 24-й после тайм-аута загрузки заголовка. Прежде «Сегодня» не называл этот источник и не имел отдельного вечернего блока. После правки в широком и мобильном окнах отображена конкретная причина без предложения неподтверждённых ответов.
|
||||||
|
- На изолированных данных через «Сегодня» открыт ответ в 18:45–18:50 с исходной репликой и проверенный черновик в 18:52–18:57. Проверены отказы для длинной задачи, ответа без источника, занятого окна, пересекающихся слотов и частичного календаря. В отдельном проходе браузер снял предложения в 19:00, пока штатный запрос плана всё ещё ожидал ответа.
|
||||||
|
- Штатный фоновый сбор Zoom в 18:35 завершил все три списка, нашёл 86 заметок и прочитал 31 заметку в пределах установленного периода. Обновлены четыре заметки, по остальным 27 изменения отсутствовали. Отсутствующих транскриптов в этой квитанции нет. На изолированном входе подтверждены продолжение обхода при тайм-ауте отдельной заметки, повтор и остановка при потере входа.
|
||||||
|
- Повтор в 18:38 также завершился с полным охватом, прочитал 31 заметку, обновил три и пропустил 28 неизменившихся. В новой квитанции пробелов нет. Рабочие запуски после исправления не воспроизвели тайм-аут, а ветвь тайм-аута подтверждена на изолированном входе.
|
||||||
@@ -16,6 +16,8 @@
|
|||||||
|
|
||||||
- **Вечерняя сверка.** Портал прекратил выдавать действия в 19:00. ActivityWatch записал интервалы до 19:00, форма сверки открылась. Первый закрывающий сбор ждал исполнителя до 19:09. После подключения очереди `flow` к исполнителю планирования повтор в 19:12 запустился вовремя и сохранил журнал без дублей. В четырёх вечерних наблюдениях ActivityWatch свеж, сбор обсуждений завершён. [Квитанция вечернего прохода](personal-os-evidence/2026-11-18/remediation-verified-pass.md)
|
- **Вечерняя сверка.** Портал прекратил выдавать действия в 19:00. ActivityWatch записал интервалы до 19:00, форма сверки открылась. Первый закрывающий сбор ждал исполнителя до 19:09. После подключения очереди `flow` к исполнителю планирования повтор в 19:12 запустился вовремя и сохранил журнал без дублей. В четырёх вечерних наблюдениях ActivityWatch свеж, сбор обсуждений завершён. [Квитанция вечернего прохода](personal-os-evidence/2026-11-18/remediation-verified-pass.md)
|
||||||
|
|
||||||
|
- **Последние 30 минут дня.** На изолированном полном календаре на «Сегодня» показаны проверенный ответ с просьбой и черновик в непересекающихся окнах. Длинный и неподтверждённый ответ исключены. При ожидающем пересчёте рекомендации сняты точно в 19:00 на контрольных часах. На рабочем неполном источнике в двух размерах окна показано удержание из-за транскриптов Zoom при подтверждённом календаре Asana. [Рабочая проверка](personal-os-evidence/2026-11-19/follow-up-live-acceptance.md)
|
||||||
|
|
||||||
- **План и приоритет.** E4 и недельный отбор принимают проверенную связь с действующим Top Priority и подтверждённые личные запросы. Production E5 сохранял текущий день, а рабочий браузер подтвердил совпадение действий, source run и ревизии дня/недели. Показанный объём действий не является постоянным лимитом рабочего дня. [Проверка приоритета](personal-os-evidence/2026-11-18/source-priority-validation.json) · [Проверка шести процессов](personal-os-evidence/2026-11-18/windmill-six-workflows.json)
|
- **План и приоритет.** E4 и недельный отбор принимают проверенную связь с действующим Top Priority и подтверждённые личные запросы. Production E5 сохранял текущий день, а рабочий браузер подтвердил совпадение действий, source run и ревизии дня/недели. Показанный объём действий не является постоянным лимитом рабочего дня. [Проверка приоритета](personal-os-evidence/2026-11-18/source-priority-validation.json) · [Проверка шести процессов](personal-os-evidence/2026-11-18/windmill-six-workflows.json)
|
||||||
- **Asana.** Действует непрерывный worker с очередью и продолжением. Текущие назначения имеют преимущество перед завершённой историей. Частичный охват не объявляется полным; зависимости перепроверяются при допуске действия. [Зависимости](personal-os-evidence/2026-11-18/asana-dependencies-validation.json)
|
- **Asana.** Действует непрерывный worker с очередью и продолжением. Текущие назначения имеют преимущество перед завершённой историей. Частичный охват не объявляется полным; зависимости перепроверяются при допуске действия. [Зависимости](personal-os-evidence/2026-11-18/asana-dependencies-validation.json)
|
||||||
- **Личная область GitHub.** Штатный проход проверил отбор из 78 PR: 48 сверены с источником, 30 merged использованы из ограниченного суточного кэша. Общие родители Asana не добавляют все соседние PR. Это квитанция конкретного среза, не постоянное число работ пользователя. [Сквозной результат](personal-os-evidence/2026-11-18/windmill-six-workflows.json)
|
- **Личная область GitHub.** Штатный проход проверил отбор из 78 PR: 48 сверены с источником, 30 merged использованы из ограниченного суточного кэша. Общие родители Asana не добавляют все соседние PR. Это квитанция конкретного среза, не постоянное число работ пользователя. [Сквозной результат](personal-os-evidence/2026-11-18/windmill-six-workflows.json)
|
||||||
|
|||||||
@@ -352,3 +352,11 @@ Zoom используется для транскриптов состоявши
|
|||||||
### D18-52 · Старая форма самочувствия в ссылке на дневной план
|
### D18-52 · Старая форма самочувствия в ссылке на дневной план
|
||||||
|
|
||||||
- В раскрываемом дневном плане оставались выпадающие поля настроения и энергии 1–5. Сохранение коррекции фокуса повторно записывало отметку самочувствия. Оставлена одна форма на «Сегодня» с переходом из уведомления. Коррекция отправляет только выбор фокуса и причину. Из старой очереди сигналов убраны кнопки ручного перерыва и сокращения нагрузки. Рабочие ссылки проверены в широком и мобильном окнах. [Проверка](personal-os-evidence/2026-11-19/follow-up-live-acceptance.md)
|
- В раскрываемом дневном плане оставались выпадающие поля настроения и энергии 1–5. Сохранение коррекции фокуса повторно записывало отметку самочувствия. Оставлена одна форма на «Сегодня» с переходом из уведомления. Коррекция отправляет только выбор фокуса и причину. Из старой очереди сигналов убраны кнопки ручного перерыва и сокращения нагрузки. Рабочие ссылки проверены в широком и мобильном окнах. [Проверка](personal-os-evidence/2026-11-19/follow-up-live-acceptance.md)
|
||||||
|
|
||||||
|
### D18-53 · Действия последних 30 минут на «Сегодня»
|
||||||
|
|
||||||
|
- Вечерний блок был доступен только в старом экране плана. На «Сегодня» теперь видны время короткого ответа, исходная просьба и переход к проверенному черновику. Длинные, неподтверждённые и пересекающиеся слоты исключаются. Таймер окончания рабочего дня работает независимо от ожидающего ответа пересчёта. Когда рекомендации удержаны из-за источников ревью, на «Сегодня» показана конкретная причина. [Проверка](personal-os-evidence/2026-11-19/follow-up-live-acceptance.md)
|
||||||
|
|
||||||
|
### D18-54 · Тайм-аут одной заметки Zoom прерывал весь сбор
|
||||||
|
|
||||||
|
- При задержке открытия отдельного транскрипта сбор продолжает обход и пробует эту заметку повторно в рамках существующего лимита. Несобранная запись сохраняется в квитанции как пробел. Сбой инвентаризации и потеря входа по-прежнему останавливают обход. Два рабочих запуска после правки успешно прочитали всю выборку заметок в пределах установленного периода. [Проверка](personal-os-evidence/2026-11-19/follow-up-live-acceptance.md)
|
||||||
+85
@@ -23,6 +23,91 @@
|
|||||||
"time": "21:31:26",
|
"time": "21:31:26",
|
||||||
"speaker": "Gabriel Weinberg",
|
"speaker": "Gabriel Weinberg",
|
||||||
"text": "Yeah."
|
"text": "Yeah."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:31:34",
|
||||||
|
"speaker": "Anne Hildebrand",
|
||||||
|
"text": "And also, uh, we want to make sure you get your questions answered around anything you want to discuss there, so submit a PeopleOps form, that's one way to do it. Um, if you do submit it without your name on the form, remember that the responses go to the project I've dropped in the channel there, so make sure not to miss those. That's it for me."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:31:58",
|
||||||
|
"speaker": "Viktor Jansson",
|
||||||
|
"text": "Alright, objective update from 05 Search Revenue. Last week, we held an assessment and switched the focus to evaluate ad experience and optimizations for search revenue. We will do this by evaluating if we can show ads from other providers for commercial queries, where Microsoft doesn't give us any ads, and how we can improve our overall shopping and product experience. For the optimizations, we have a few products in the pipeline, mainly for optimizing product ads, which we expect a potential $4 million a year revenue increase for. And the Duckbill Award goes to Renato, who ran a product during Quick Win Days to highlight prices for products on sale, which added half a million dollars a year, while at the same time helping our users to find discounted products. Great example of time well spent during Quick Win Days with a large impact to the objective KPIs. Over to Kirby for more million dollar updates."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:32:56",
|
||||||
|
"speaker": "Kirby (she/they)",
|
||||||
|
"text": "Hey, thanks Victor. So we built a lot this past cycle. We hit 150,000 active subscribers. We're tracking toward 180,000 by the end of this year, which is the stretch goal that we set out in our last cycle. And it's not official yet, but tentatively, we're working toward 300,000 as the goal for the end of next year, end of 2027. So more exciting updates to come. This slide is mostly about one A/B test that we did, and I want to take a minute to explain it, because it's really interesting. So we, in rough numbers, about 800 people start a subscription every day. Some of them are going direct into a paid subscription. Some of them are starting a free trial. Also, every day we have between 4 and 500 people canceling their subscription. So more than half. And the difference between those numbers is net subscriber additions. That's OO's KPI. So we're not just trying to get more people to start a subscription. We're also trying to get fewer people to cancel. And one way we're doing that is by giving people a better day zero experience as soon as they subscribe, creating subscriber onboarding. But another way we're doing it is by shifting people over to yearly plans. So we've got monthly, we've got yearly. Monthly plans have a 14% churn rate. That's the likelihood that someone will cancel every month. Yearly plans, it's only 2%, because you only have one opportunity to churn. every year. So to try to shift people toward yearly, we tested removing the free trial option from monthly plans to make yearly a little more appealing. And we did this first on Android, and we saw what we expected or what we hoped for. which is that fewer people are getting a subscription, because some people will see that there's no free trial on monthly, they don't want yearly, they bounce. But of the people who do stick around, more of them are on yearly, and we modeled it out so that we'd catch up with where. our growth would have been within a few months. Then we also did it on iOS, and we saw something surprising, which is that iOS users will just buy the monthly subscription with no free trial, and more of them are going yearly also. So the number of extra people we got on iOS. cancels out the people we lost on Android. So there's no cost. It's not going to take any time to catch up with where our growth would have been. We're just getting the extra lifetime value of each of these users by shifting more people over to yearly. And we're going to reduce the churn rate from where it is now, 9.5%, which is pretty good compared to the industry standard for VPNs. We want it to be at 7% and we want it to stay there. And a year from now, just from this one experiment. we're going to get down to 7.2% churn from having more people coming in on yearly plans. So if you're close to OO, you know that we work really hard to get one experiment out right after the last one. And this is why we never know. when we're gonna have one that's even more valuable than we'd anticipated. In lifetime value, which is different from revenue, but it's the projected amount of revenue we expect to get per user, we're adding 5 million in lifetime value per. year just from this one experiment. So the next time we're moving quickly to get something out for an experiment, out of the way, because these are really valuable and interesting. And congratulations to the team for all the work that got done. Super congratulations to Bartek, who's come on in a tech lead role. And as soon as he got to the objective, he got up to speed really quickly. and is our Duckbill winner because he has been accelerating the timelines in project after project by just thinking of something we didn't think of, coming up with a simpler way to solve something, and he connected the dots between incidents that we've been seeing across platforms where users are getting logged out and we don't know why. And he'd figured out a path forward, turned it into an SLO, and owned it, and is getting it fixed. Congratulations, Bartek. And next, on to Russell, I think."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:37:23",
|
||||||
|
"speaker": "russell holt",
|
||||||
|
"text": "Thanks, Gary. Hi, yeah, Russell here with an update for OE, where we are working on cookie prompt management and ad blocking. Um, for CPM, we've been, uh, improving the automated pipeline, which processes. uh, fixes and unhandled sites. And at this point, about 50% of these changes, uh, to our data set are fully automated, meaning no human involvement, which is pretty exciting. And right now we're at about overall 73% coverage, uh, in the US. Right now. So, one thing that we did this cycle is we finally got a reasonable opt-in rate. um, of 65% on Android, similar to iOS for CPM. Uh, which is a flow we've been struggling with for a bit, because we have a lot of cases where we need user consent. to automatically do something for them for legal reasons. And in this case, it was we need to accept cookies under certain circumstances to raise that coverage to improve the experience. And it's been a little bit of trouble, but that was great. So, on this cycle, we're really doubling down on the CPM automation, making the agent faster, more accurate, improving heuristic pop-up detection, doing regression testing, and this is really the key. to the long-term maintenance mode for CPM, which is finally in sight again, which we're pretty excited about. Um, and on the ad blocking side, uh, we're moving forward with two of the main feature requests. that we got from our analysis after the big YouTube ad blocking launch. Um, which will be… One, expanding ad blocking to a selection of top sites, starting with adult sites. And then two, on YouTube specifically, giving users the ability to skip sponsored segments. sponsored content segments, which look like ads to people. Uh, through YouTube. So, the goal there is to give people more control. and make a better default experience, even for YouTube Premium. where people see these sponsored segments as ads, so better not block. All right, that's it for OE. I guess on to Emmanuelle."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:39:42",
|
||||||
|
"speaker": "Emanuele Feliziani",
|
||||||
|
"text": "Yes, Russell, thanks. Exciting updates on the CDM and unblocking stuff. We had OI assessment last week. We will keep the focus on ChatSync. We will have a hopefully final push to stabilize the ChatSync system. So this will be more of a milestone focus. And we'll focus, yes, on stabilization. What we've done in the sparse cycle, we've optimized. the Syncflow on several platforms, both mobile and, uh, like, native and, um, and third-party, and we also shipped, uh, promos for ChatSync. For third parties, we've seen a significant inflow of users from those promos. We've doubled the daily activations and the daily active users for third parties specifically is up almost 5x. And it's 30% across all platforms. So really nice bump in users there. We're still not happy with the adoption. We want more. We are also not happy with the conversion rates. So we will keep iterating. in the… in future cycles, so after this one. Uh, in the meantime, we are also running, uh, user research to try and understand the drop of… the drop of points, the points of confusion. For users during the sync setup flow and see where we can be more helpful towards our users, making sure that they are able to set up sync. um… across multiple devices especially. We'll also look into ways to make the sync entry point more prominent. We've already seen that people struggle to find it in the settings, it's kind of buried, so we try to make it more prominent, lift it up a little bit, make it more intuitive how to use it. Um, and yeah. That's it for ROI. The duck bill goes to Pablo Serrano, the pub RR. designer, Serrano, uh, designer Pablo, we also have a Pablo on the back end. Uh, and yeah, he's, he's been great. Uh, he's been great for a long time and this cycle we are happy to congratulate him, uh, with, with the award. And, uh, yeah, he's been, uh, he's been great, uh, on the project advisory side, very quick and tidy on the design front. And he's been, diving into the metrics with, with Gusto. So, really happy to see Pablo grow in this role. Thanks, Pablo, and congrats. Uh, next up, I think, is, uh, team member, uh, updates with John Russell."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:42:36",
|
||||||
|
"speaker": "Jon Russo",
|
||||||
|
"text": "Hi. Didn't realize I was coming up this soon. But, hi, everybody. I'm John from the CloudOps team. I guess I'm Matt. Two different angles here. One from the work side. Recently, I started working on some of our Kubernetes stuff, which is really exciting and a little intimidating for me. I've been wanting to kind of dip my toes into this technology for a while, and it's nice to be able to finally do it. I got a lot of people on the team helping me in that, and kind of designing a new deployment spec for some of our new Kubernetes clusters. So that's fun and exciting in the infra velocity kind of project there. On a personal side, two pretty exciting things coming up. One tomorrow, I've been doing a lot of trail running this past year, and I have a half marathon coming up at a place called Von Trapp Mountain. Um, it's just outside of Stowe, Vermont, if anybody knows the area. Uh, the Von Trapp family is famous from The Sound of Music, uh, movies way back in the day, and it's an absolutely beautiful property up there. Um, so it's gonna be a fun race, um… there that has about 2,000 feet of elevation gain in it. I think that's roughly like 600 meters for the Europeans. The weather is looking pretty terrible, actually, so it might be a bit of a slop fest. We're expected to get an inch of rain today, and then it's going to be 33 degrees Fahrenheit at the start of the race, about 1 degree Celsius, so pretty dang cold. Um, and then the other side is that, uh, about a couple weeks ago, I booked a trip to Japan for end of January. Uh, this is something I've always wanted to do. It's been a bucket list trip for me to go skiing there. Um, so me and a couple friends are going to Hokkaido for two weeks. And, uh, traveling around in a van, and kind of exploring all the resorts there, and doing some backcountry skiing, uh, maybe even backcountry skiing into a volcano. Um, so that should be fun. Yeah, that's it for me. Thanks."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:44:56",
|
||||||
|
"speaker": "Stephen",
|
||||||
|
"text": "Oh, I think this is me. Hi, so we've been working on a refreshed new tab page for iOS for a while, and we're finally getting ready for the first internal release. It should be next week. I think there's a link in there if you want to enable it early. I tried to make a video. So this would work out well, but that didn't work. So we're gonna do it live demo real quick. So hopefully this does work out. I'm gonna try to. share this via… My phone. So if it doesn't work, someone please yell at me. Okay. Hopefully you can see my phone."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:45:37",
|
||||||
|
"speaker": "Spencer",
|
||||||
|
"text": "Yes."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:45:38",
|
||||||
|
"speaker": "Stephen",
|
||||||
|
"text": "Okay, awesome. Thank you. So… So here's the new tab — the new, new tab page. Kind of a weird phrasing. But you can see it's more brand forward. We have Dax there. He's got some Dax greetings. We have a nice background illustration there in the background. And the greetings will update. Probably not this frequently, but you can kind of see there's a few in there and we're going to update them based on time of day and some other stuff. The toggle's right there so you can switch between chat and search pretty easily. And you can see we have. Placeholder favorites now. So these are not hidden by default. So this is kind of inviting people to add more favorites. I could add one. Maybe. Yep, and there's that one, so that goes there. But we don't show the, uh… icon until you go there once. And then we also have… The return to hatch there. So you can get back to that. You can swipe it away if you want. Also get to your new tabs easier. And… I think I lost my window. I have too many DuckDuckGo's going here. You can also… Have a lot of favorites. So they scroll and expand and contract. And you can use the customization right on the page now, so you can toggle them on and off. From here, change some settings, we're gonna add some more there as well, and it also has… Dark mode, so if you want a cool fancy dark mode, you can have that as well, and everything should look… Nice there as well. So if you want to try that out, you can wait till next week, or you can follow the link and turn it on for yourself now."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:47:35",
|
||||||
|
"speaker": "Julia Abreu",
|
||||||
|
"text": "We had one more demo added to the agenda, but there is no name there. So, how to enable Redesign NTP? Does anyone…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:47:45",
|
||||||
|
"speaker": "Gabriel Weinberg",
|
||||||
|
"text": "I think that's Brett Stevens."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:47:46",
|
||||||
|
"speaker": "Stephen",
|
||||||
|
"text": "Oh, yeah, that was part of mine, sorry."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:47:48",
|
||||||
|
"speaker": "Gabriel Weinberg",
|
||||||
|
"text": "Yeah."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:47:48",
|
||||||
|
"speaker": "Julia Abreu",
|
||||||
|
"text": "Oh, okay, cool. Just wanted to make sure that there was no extra in there. Uh, okay, before we wrap up, I just wanted to say or remind everybody that next day we have Hack Days. Um, it's the last Hack Days of the year. We will have a quick win in, uh, December. If you want to check the dates, you can see the AOR for special days. I am just drafting the calendar for next year, so I'll get Gabriel to take a look, and then I will post it as well under the AOR. I am working on some ideas on how to make the meeting for the demos a little shorter. I don't want to make it. every demo shorter than 2 minutes, because, um, I know you all wouldn't like that, but I'm trying to figure out, um, some ways that maybe we can split demos in different days, or, like, so if you have any ideas, or if you would like to help in any way. please go ahead and ping me. And otherwise, I just have one personal update that is special for Elle. Um, this is Maple. This is our third Dash Hound. Um, I am completely out of control now that I have a garden, so, um, I might be just sharing a new Dutch hound every week. Uh, and if I get her to not bark at everyone, I will bring her to our next meetup, uh, because the other two are not. very well trained. Um, so that's it for me. I hope you all have a great weekend, and uh… we'll roll out GPT for next week."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:49:17",
|
||||||
|
"speaker": "Gabriel Weinberg",
|
||||||
|
"text": "Yeah, one idea for you, Julia, for everyone here, is that to make less demos, you can collaborate with each other on projects, so… I'm not sure our best process for that at the moment, but there's the Hack Days project. You can say what you're gonna work on in there and try to recruit people. There's, we have a Mattermost channel too, don't we? Do you? Or if not, maybe we should."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:49:37",
|
||||||
|
"speaker": "Julia Abreu",
|
||||||
|
"text": "We do have, yes, specialties."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "21:49:40",
|
||||||
|
"speaker": "Gabriel Weinberg",
|
||||||
|
"text": "Maybe that's the place to go. Go to the Mattermost channel. Um, okay, hope everyone's gonna be good."
|
||||||
}
|
}
|
||||||
]
|
]
|
||||||
}
|
}
|
||||||
+17
@@ -15,3 +15,20 @@ source: "https://us01mynotes.zoom.us/klekRLtLSEuMv0PCCPCjtA"
|
|||||||
- **21:31:08 · Gabriel Weinberg:** Hello? That was a nice fade out. Um… Okay… No new team members, no AOR updates. Um, we had one company update, there's no name by it.
|
- **21:31:08 · Gabriel Weinberg:** Hello? That was a nice fade out. Um… Okay… No new team members, no AOR updates. Um, we had one company update, there's no name by it.
|
||||||
- **21:31:23 · Anne Hildebrand:** Sorry, that was me. Um, yes, this is Anne. Uh, just a reminder to check your inboxes for the outcome of the compensation review.
|
- **21:31:23 · Anne Hildebrand:** Sorry, that was me. Um, yes, this is Anne. Uh, just a reminder to check your inboxes for the outcome of the compensation review.
|
||||||
- **21:31:26 · Gabriel Weinberg:** Yeah.
|
- **21:31:26 · Gabriel Weinberg:** Yeah.
|
||||||
|
- **21:31:34 · Anne Hildebrand:** And also, uh, we want to make sure you get your questions answered around anything you want to discuss there, so submit a PeopleOps form, that's one way to do it. Um, if you do submit it without your name on the form, remember that the responses go to the project I've dropped in the channel there, so make sure not to miss those. That's it for me.
|
||||||
|
- **21:31:58 · Viktor Jansson:** Alright, objective update from 05 Search Revenue. Last week, we held an assessment and switched the focus to evaluate ad experience and optimizations for search revenue. We will do this by evaluating if we can show ads from other providers for commercial queries, where Microsoft doesn't give us any ads, and how we can improve our overall shopping and product experience. For the optimizations, we have a few products in the pipeline, mainly for optimizing product ads, which we expect a potential $4 million a year revenue increase for. And the Duckbill Award goes to Renato, who ran a product during Quick Win Days to highlight prices for products on sale, which added half a million dollars a year, while at the same time helping our users to find discounted products. Great example of time well spent during Quick Win Days with a large impact to the objective KPIs. Over to Kirby for more million dollar updates.
|
||||||
|
- **21:32:56 · Kirby (she/they):** Hey, thanks Victor. So we built a lot this past cycle. We hit 150,000 active subscribers. We're tracking toward 180,000 by the end of this year, which is the stretch goal that we set out in our last cycle. And it's not official yet, but tentatively, we're working toward 300,000 as the goal for the end of next year, end of 2027. So more exciting updates to come. This slide is mostly about one A/B test that we did, and I want to take a minute to explain it, because it's really interesting. So we, in rough numbers, about 800 people start a subscription every day. Some of them are going direct into a paid subscription. Some of them are starting a free trial. Also, every day we have between 4 and 500 people canceling their subscription. So more than half. And the difference between those numbers is net subscriber additions. That's OO's KPI. So we're not just trying to get more people to start a subscription. We're also trying to get fewer people to cancel. And one way we're doing that is by giving people a better day zero experience as soon as they subscribe, creating subscriber onboarding. But another way we're doing it is by shifting people over to yearly plans. So we've got monthly, we've got yearly. Monthly plans have a 14% churn rate. That's the likelihood that someone will cancel every month. Yearly plans, it's only 2%, because you only have one opportunity to churn. every year. So to try to shift people toward yearly, we tested removing the free trial option from monthly plans to make yearly a little more appealing. And we did this first on Android, and we saw what we expected or what we hoped for. which is that fewer people are getting a subscription, because some people will see that there's no free trial on monthly, they don't want yearly, they bounce. But of the people who do stick around, more of them are on yearly, and we modeled it out so that we'd catch up with where. our growth would have been within a few months. Then we also did it on iOS, and we saw something surprising, which is that iOS users will just buy the monthly subscription with no free trial, and more of them are going yearly also. So the number of extra people we got on iOS. cancels out the people we lost on Android. So there's no cost. It's not going to take any time to catch up with where our growth would have been. We're just getting the extra lifetime value of each of these users by shifting more people over to yearly. And we're going to reduce the churn rate from where it is now, 9.5%, which is pretty good compared to the industry standard for VPNs. We want it to be at 7% and we want it to stay there. And a year from now, just from this one experiment. we're going to get down to 7.2% churn from having more people coming in on yearly plans. So if you're close to OO, you know that we work really hard to get one experiment out right after the last one. And this is why we never know. when we're gonna have one that's even more valuable than we'd anticipated. In lifetime value, which is different from revenue, but it's the projected amount of revenue we expect to get per user, we're adding 5 million in lifetime value per. year just from this one experiment. So the next time we're moving quickly to get something out for an experiment, out of the way, because these are really valuable and interesting. And congratulations to the team for all the work that got done. Super congratulations to Bartek, who's come on in a tech lead role. And as soon as he got to the objective, he got up to speed really quickly. and is our Duckbill winner because he has been accelerating the timelines in project after project by just thinking of something we didn't think of, coming up with a simpler way to solve something, and he connected the dots between incidents that we've been seeing across platforms where users are getting logged out and we don't know why. And he'd figured out a path forward, turned it into an SLO, and owned it, and is getting it fixed. Congratulations, Bartek. And next, on to Russell, I think.
|
||||||
|
- **21:37:23 · russell holt:** Thanks, Gary. Hi, yeah, Russell here with an update for OE, where we are working on cookie prompt management and ad blocking. Um, for CPM, we've been, uh, improving the automated pipeline, which processes. uh, fixes and unhandled sites. And at this point, about 50% of these changes, uh, to our data set are fully automated, meaning no human involvement, which is pretty exciting. And right now we're at about overall 73% coverage, uh, in the US. Right now. So, one thing that we did this cycle is we finally got a reasonable opt-in rate. um, of 65% on Android, similar to iOS for CPM. Uh, which is a flow we've been struggling with for a bit, because we have a lot of cases where we need user consent. to automatically do something for them for legal reasons. And in this case, it was we need to accept cookies under certain circumstances to raise that coverage to improve the experience. And it's been a little bit of trouble, but that was great. So, on this cycle, we're really doubling down on the CPM automation, making the agent faster, more accurate, improving heuristic pop-up detection, doing regression testing, and this is really the key. to the long-term maintenance mode for CPM, which is finally in sight again, which we're pretty excited about. Um, and on the ad blocking side, uh, we're moving forward with two of the main feature requests. that we got from our analysis after the big YouTube ad blocking launch. Um, which will be… One, expanding ad blocking to a selection of top sites, starting with adult sites. And then two, on YouTube specifically, giving users the ability to skip sponsored segments. sponsored content segments, which look like ads to people. Uh, through YouTube. So, the goal there is to give people more control. and make a better default experience, even for YouTube Premium. where people see these sponsored segments as ads, so better not block. All right, that's it for OE. I guess on to Emmanuelle.
|
||||||
|
- **21:39:42 · Emanuele Feliziani:** Yes, Russell, thanks. Exciting updates on the CDM and unblocking stuff. We had OI assessment last week. We will keep the focus on ChatSync. We will have a hopefully final push to stabilize the ChatSync system. So this will be more of a milestone focus. And we'll focus, yes, on stabilization. What we've done in the sparse cycle, we've optimized. the Syncflow on several platforms, both mobile and, uh, like, native and, um, and third-party, and we also shipped, uh, promos for ChatSync. For third parties, we've seen a significant inflow of users from those promos. We've doubled the daily activations and the daily active users for third parties specifically is up almost 5x. And it's 30% across all platforms. So really nice bump in users there. We're still not happy with the adoption. We want more. We are also not happy with the conversion rates. So we will keep iterating. in the… in future cycles, so after this one. Uh, in the meantime, we are also running, uh, user research to try and understand the drop of… the drop of points, the points of confusion. For users during the sync setup flow and see where we can be more helpful towards our users, making sure that they are able to set up sync. um… across multiple devices especially. We'll also look into ways to make the sync entry point more prominent. We've already seen that people struggle to find it in the settings, it's kind of buried, so we try to make it more prominent, lift it up a little bit, make it more intuitive how to use it. Um, and yeah. That's it for ROI. The duck bill goes to Pablo Serrano, the pub RR. designer, Serrano, uh, designer Pablo, we also have a Pablo on the back end. Uh, and yeah, he's, he's been great. Uh, he's been great for a long time and this cycle we are happy to congratulate him, uh, with, with the award. And, uh, yeah, he's been, uh, he's been great, uh, on the project advisory side, very quick and tidy on the design front. And he's been, diving into the metrics with, with Gusto. So, really happy to see Pablo grow in this role. Thanks, Pablo, and congrats. Uh, next up, I think, is, uh, team member, uh, updates with John Russell.
|
||||||
|
- **21:42:36 · Jon Russo:** Hi. Didn't realize I was coming up this soon. But, hi, everybody. I'm John from the CloudOps team. I guess I'm Matt. Two different angles here. One from the work side. Recently, I started working on some of our Kubernetes stuff, which is really exciting and a little intimidating for me. I've been wanting to kind of dip my toes into this technology for a while, and it's nice to be able to finally do it. I got a lot of people on the team helping me in that, and kind of designing a new deployment spec for some of our new Kubernetes clusters. So that's fun and exciting in the infra velocity kind of project there. On a personal side, two pretty exciting things coming up. One tomorrow, I've been doing a lot of trail running this past year, and I have a half marathon coming up at a place called Von Trapp Mountain. Um, it's just outside of Stowe, Vermont, if anybody knows the area. Uh, the Von Trapp family is famous from The Sound of Music, uh, movies way back in the day, and it's an absolutely beautiful property up there. Um, so it's gonna be a fun race, um… there that has about 2,000 feet of elevation gain in it. I think that's roughly like 600 meters for the Europeans. The weather is looking pretty terrible, actually, so it might be a bit of a slop fest. We're expected to get an inch of rain today, and then it's going to be 33 degrees Fahrenheit at the start of the race, about 1 degree Celsius, so pretty dang cold. Um, and then the other side is that, uh, about a couple weeks ago, I booked a trip to Japan for end of January. Uh, this is something I've always wanted to do. It's been a bucket list trip for me to go skiing there. Um, so me and a couple friends are going to Hokkaido for two weeks. And, uh, traveling around in a van, and kind of exploring all the resorts there, and doing some backcountry skiing, uh, maybe even backcountry skiing into a volcano. Um, so that should be fun. Yeah, that's it for me. Thanks.
|
||||||
|
- **21:44:56 · Stephen:** Oh, I think this is me. Hi, so we've been working on a refreshed new tab page for iOS for a while, and we're finally getting ready for the first internal release. It should be next week. I think there's a link in there if you want to enable it early. I tried to make a video. So this would work out well, but that didn't work. So we're gonna do it live demo real quick. So hopefully this does work out. I'm gonna try to. share this via… My phone. So if it doesn't work, someone please yell at me. Okay. Hopefully you can see my phone.
|
||||||
|
- **21:45:37 · Spencer:** Yes.
|
||||||
|
- **21:45:38 · Stephen:** Okay, awesome. Thank you. So… So here's the new tab — the new, new tab page. Kind of a weird phrasing. But you can see it's more brand forward. We have Dax there. He's got some Dax greetings. We have a nice background illustration there in the background. And the greetings will update. Probably not this frequently, but you can kind of see there's a few in there and we're going to update them based on time of day and some other stuff. The toggle's right there so you can switch between chat and search pretty easily. And you can see we have. Placeholder favorites now. So these are not hidden by default. So this is kind of inviting people to add more favorites. I could add one. Maybe. Yep, and there's that one, so that goes there. But we don't show the, uh… icon until you go there once. And then we also have… The return to hatch there. So you can get back to that. You can swipe it away if you want. Also get to your new tabs easier. And… I think I lost my window. I have too many DuckDuckGo's going here. You can also… Have a lot of favorites. So they scroll and expand and contract. And you can use the customization right on the page now, so you can toggle them on and off. From here, change some settings, we're gonna add some more there as well, and it also has… Dark mode, so if you want a cool fancy dark mode, you can have that as well, and everything should look… Nice there as well. So if you want to try that out, you can wait till next week, or you can follow the link and turn it on for yourself now.
|
||||||
|
- **21:47:35 · Julia Abreu:** We had one more demo added to the agenda, but there is no name there. So, how to enable Redesign NTP? Does anyone…
|
||||||
|
- **21:47:45 · Gabriel Weinberg:** I think that's Brett Stevens.
|
||||||
|
- **21:47:46 · Stephen:** Oh, yeah, that was part of mine, sorry.
|
||||||
|
- **21:47:48 · Gabriel Weinberg:** Yeah.
|
||||||
|
- **21:47:48 · Julia Abreu:** Oh, okay, cool. Just wanted to make sure that there was no extra in there. Uh, okay, before we wrap up, I just wanted to say or remind everybody that next day we have Hack Days. Um, it's the last Hack Days of the year. We will have a quick win in, uh, December. If you want to check the dates, you can see the AOR for special days. I am just drafting the calendar for next year, so I'll get Gabriel to take a look, and then I will post it as well under the AOR. I am working on some ideas on how to make the meeting for the demos a little shorter. I don't want to make it. every demo shorter than 2 minutes, because, um, I know you all wouldn't like that, but I'm trying to figure out, um, some ways that maybe we can split demos in different days, or, like, so if you have any ideas, or if you would like to help in any way. please go ahead and ping me. And otherwise, I just have one personal update that is special for Elle. Um, this is Maple. This is our third Dash Hound. Um, I am completely out of control now that I have a garden, so, um, I might be just sharing a new Dutch hound every week. Uh, and if I get her to not bark at everyone, I will bring her to our next meetup, uh, because the other two are not. very well trained. Um, so that's it for me. I hope you all have a great weekend, and uh… we'll roll out GPT for next week.
|
||||||
|
- **21:49:17 · Gabriel Weinberg:** Yeah, one idea for you, Julia, for everyone here, is that to make less demos, you can collaborate with each other on projects, so… I'm not sure our best process for that at the moment, but there's the Hack Days project. You can say what you're gonna work on in there and try to recruit people. There's, we have a Mattermost channel too, don't we? Do you? Or if not, maybe we should.
|
||||||
|
- **21:49:37 · Julia Abreu:** We do have, yes, specialties.
|
||||||
|
- **21:49:40 · Gabriel Weinberg:** Maybe that's the place to go. Go to the Mattermost channel. Um, okay, hope everyone's gonna be good.
|
||||||
+280
@@ -23,6 +23,286 @@
|
|||||||
"time": "15:28:24",
|
"time": "15:28:24",
|
||||||
"speaker": "Michal Smaga",
|
"speaker": "Michal Smaga",
|
||||||
"text": "Оф, Ол, Юзерс, Райт?"
|
"text": "Оф, Ол, Юзерс, Райт?"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:28:26",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Uh, yes, from daily users. So it's still higher than the 1% threshold. And, uh… We've got two fixes more to be released next week. The RR6 reload. which can. Probably cover up to 30% of every 6 total reports. Uh, because it's not… It's not proof that… So usually LR6, when it's happening, it should be recovered normally by the kit. And we are handling some kind of case when it's not recovering. And, uh… The other one is, uh, braid, which may… affect… Yes. The upgrade failures, some hacks caused by it. So speculating on it another like 20 or 30%."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:29:25",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "So, those two errors are… On my like, in this week's internal. Yeah. So they will be public next week."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:29:33",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Yes, and those are not the main part of the failures we are getting. So they are expected to have some effect."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:29:45",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "But not that big, yeah."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:29:47",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "But it won't cross the one-person gap still, I think. Thank you."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:29:55",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah, I think, like, probably the main reason is that we have exhausted the options, as you mentioned, like that, like the… The the the idle like thing was not we are not able to reproduce that one. So basically, probably, like…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:30:10",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Yeah, so…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:30:12",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah. Okay. My gut feeling would be to just, like, probably provide a summary of the current state of the, like."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:30:19",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "the reporting have maybe some kind of, like, a dashboard and and probably, like, revisit it next week when we have the numbers for the the last part of the fixes."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:30:26",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Okay."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:30:32",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "But probably I wouldn't hold the project any longer, like, for the next week."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:30:39",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Um… So my suggestion, my thinking of the project for this week. So, first of all, we have this… A lot of feature flags introduced, and the pixels, metrics, and so on, and all of them are basically marked as temporary, and some of them are not, uh… Not even being reported. So there is some cleanup, I guess, to be done. For example, this graveyard fix, we are catching some… percent, lower than one percent, probably, of all the reported events that can match it, but I would not activate it, looking at it, looking at the numbers, so it's… Certainly not the main issue. And there is another one. Um, yeah, anyways, uh, so I was thinking to clean it up. To see… to check the… Safe fixes and remove the picture flags for them. To leave the more controversial, dangerous things behind the feature flags. that we are living. And, uh… Maybe should we remove the recorder thing and episode diagnostics? Maybe not this week, but the week after. Because we are generally… We've collected all the data that it could provide, I guess. And the main portion of it that remains, the neat quake. I think it's generally not covered by it in any sense. And that's…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:32:36",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:32:37",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Certainly the main cause."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:32:38",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "And… And apart from the episodes, the diagnostics, we… We have other pixels to measure the hangs, right? Or they are, like…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:32:55",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Currently, there is only this episode thing. We have, uh, the pixels for reports, for user reports, that is separating. And yeah, currently this episode metric only includes like background events, the state of things and. initiating, basically, event."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:33:17",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "So we would also, like, lose the visibility into the the the, like, the permanent hangs. Right?"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:33:25",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "I'm sorry."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:33:25",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Like, the absolute…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:33:27",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Actually, I was thinking, maybe not finished planning yet, but I was thinking about redoing the whole reporting thing. Centering it in the gray zone, like, uncovered. Places. So, we have… We have a sequence, like, the content script being injected. sending INIT, WebKit receiving the INIT. In the ground, receiving and handling the event. The config fetching. And… And the website fetching, that may also indicate some kind of… So there is a… Uh, corner case, when the config cannot be fetched. For whatever reason, if it hangs, it would give this kind of symptoms. But… Normally it should not hang. So I would… I'm hesitating to fix it. Um, before we have any data on this. And… Generally, they should not cause permanent headaches. We're done, right? But yeah, currently we are not, uh, tied to the… Extension lifecycle, but we are tied to the page loading lifecycle. And, uh… Also, we don't… we don't have a baseline for the navigations, so we can't tell of which percent of navigation actually have the extension radar, but we are… Diminishing it instead against, like, daily users or installs metrics. So, uh… Still maybe has not. the correct numbers. We don't understand, like, the… the whole picture of it, like how many navigations do actually succeed. And, uh, where exactly… You've got this, uh… Meaning, like, failure, for example. Maybe it is the init, maybe it's somewhere in the WebKit internals, or maybe it's on our side, like waiting for the config fetch or whatever. And that would include the extension modification itself already, the measurement points. And probably some… Thank you. Not episodes, but a summary of, uh… The hung state, so whenever we catch it, when any chain of this, any part of the chain fails, we could report where it has failed."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:36:18",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Thanks."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:36:23",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "So I'm thinking, well, maybe we should just, uh… Check this week to implement this metrics. redoing… And uh… then put it on hold, basically, before we receive. I mean, data related."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:36:41",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "So you suggest, like, doing it for roughly a week, probably, more? Do the cleanup."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:36:46",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Yeah, so…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:36:48",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Do the cleanup and implement those, like, long term metrics to have some kind of observability in this gray zone."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:36:55",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Yeah. So, first of all, it should have… some permanent part, I guess, just so we keep it for… For catching the spikes or future breaks. And have some temporary metrics as well, but measuring the inside of the extension. because currently we have the okay. Everything seems fine. Background is loaded in less than 1 s. And then nothing. So and we don't have any idea and anything suspect that can be suspected for causing it."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:37:25",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah. Mmhm."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:37:36",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "So yeah, no hypothesis on that. What is actually breaking there? Except for a couple of issues inside the GS, but they are also… Kind of. Hmm. Do not seem so severe and causing this problems. So, that would be my point for today. Otherwise, I have no ideas for you."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:38:04",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah, so… Yeah, so I think, like, the point that Russell is having is probably he wants to, like… have a clear understanding where the project will end, yeah, because we can, like, just probably, like, move it, move it, like, chase the, chase the, the, the, the finish line, like, forever."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:38:20",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Okay."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:38:27",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "That's going away."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:38:29",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Well… I would highlight that we certainly have some improvement. We've found some actual bugs. Some of them are not even the case because they are really corner cases, but those are a real issue reproduced."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:38:44",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah, I understand, yeah, yeah, so probably no one is questioning that, yeah, but just, like, to have, like, a more… like, condensed or, like, focused, uh, wrap-up for the project, yeah, so I would suggest, like, yeah, I like the plan, and I think it makes sense, yeah, so I would… I would bring the… Uh, yeah, so I think… Two things need to be done, like, one is to do a project update and, like, recap the current data, and summarize, like, where we landed at. And probably, I would start with, uh… with the DRI task, and, like, respond to Russell, and, like… Present the plan that, like, we have those, like, we have those fixes, the updates, and link to the project update, that you have the, like. Two more fixes, like, that will come up in the next version, and you propose the plan to, uh. Uh, have the final week of the project to, like, clean up the, like, all… Unnecessary things and clean up the diagnostics that we set up in place for, like, nailing down the issues. And suggest like to to metrics to like keep track of like where we're at with the permanent hangs and maybe better visibility in what's happening."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:39:58",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Introduce some permanent. Yep."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:09",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah, and probably, like, we can see, like, what happens. Yeah, I… probably you want, like, if there's, like, a concrete plan that, like, this is the final week, and we just need to clean up the stuff, I don't think that there should be an issue. That's my feelings."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:27",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Okay, yeah, I'll do it today, then, before the… before the meeting, hopefully."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:36",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah, and the meetings today are, like, the regular time, yeah."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:40",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "I think it's one hour later shifted due to some times on things."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:46",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah, but I think that…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:46",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "All the meetings, actually."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:52",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Oh, didn't you have the time zone change for the winter time?"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:55",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "No, luckily I don't have time zone change, but unfortunately I do have the Easter."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:40:57",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Okay."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:03",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "And the western time zones changing and shifting my schedules up to the late night."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:08",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "the… Okay. Uh, yeah, so sounds good. Uh, like, I would ask you to, to, like, do the project update, and, like, provide the, the, the data analysis in some, some task, and, like, respond to us also."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:19",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Mm-hmm."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:24",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "So we have the clear understanding, and we can discuss it on the meeting if he has any."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:28",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Okay, so…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:29",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Further questions, but it sounds good to me."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:30",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Probably. Probably the cleanup is, uh… I think, to be done anyways, whether it comes into the project or just as a follow-up, but it would be necessary to do it anyways, and it's not…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:40",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah, yeah."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:47",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "And it's not small, because there's multiple things in it, and it should be…"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:49",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Yeah. Yeah, okay."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:41:53",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Some considerations whether things should be kept or cleaned up or enabled for everybody. So it seems like a natural part for a natural completion part of the project. But yeah, for the metrics. It also seems pretty natural to have some permanent metrics just in case we could detect if anything becomes worse. at some point. Okay, that would be it then. Thanks."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:42:32",
|
||||||
|
"speaker": "Michal Smaga",
|
||||||
|
"text": "Cool. Yeah, that sounds good. Yeah, thank you for the chat."
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"time": "15:42:40",
|
||||||
|
"speaker": "Alex Martemyanov",
|
||||||
|
"text": "Okay, see you, see you later."
|
||||||
}
|
}
|
||||||
]
|
]
|
||||||
}
|
}
|
||||||
+56
@@ -15,3 +15,59 @@ source: "https://us01mynotes.zoom.us/qQ2bzU5KTiCGluiBYA-ROA"
|
|||||||
- **15:26:12 · Alex Martemyanov:** So yeah, I wanted to have a sync up on the project. And regarding the Russell's questioning. In the Advisor Sync thread. So whether do we do we wrap up or. Continuing something. Um… So, yeah, basically, at the current state, it seems, from my perspective, that we are. exhausted all the…
|
- **15:26:12 · Alex Martemyanov:** So yeah, I wanted to have a sync up on the project. And regarding the Russell's questioning. In the Advisor Sync thread. So whether do we do we wrap up or. Continuing something. Um… So, yeah, basically, at the current state, it seems, from my perspective, that we are. exhausted all the…
|
||||||
- **15:26:44 · Alex Martemyanov:** hypothesis that we've got so far. And so there's been some fixes that are genuine and they are working. And, uh, according to the first day data I got today, there is a 30% drop of macOS over-reporting, so that was expected from the. Error page fix the data pollution. So. Currently, they are comparable with iOS and macOS. Uh, yes, locally, not yet in the… The data. So basically, I think we are matching iOS. Uh, roughly. And, uh… So the main numbers are… So for any failure, there is a drop in my voice from around thirty percent to around twenty percent. And, oh, sorry, from 20% to around 15%, so it's a 30% drop. And, uh, so otherwise it's pretty much steady. Uh, since there was no actual fixes, and uh… In permanent hang numbers, macOS had around. 2 to 3%. And now it's around steady. 1.9%.
|
- **15:26:44 · Alex Martemyanov:** hypothesis that we've got so far. And so there's been some fixes that are genuine and they are working. And, uh, according to the first day data I got today, there is a 30% drop of macOS over-reporting, so that was expected from the. Error page fix the data pollution. So. Currently, they are comparable with iOS and macOS. Uh, yes, locally, not yet in the… The data. So basically, I think we are matching iOS. Uh, roughly. And, uh… So the main numbers are… So for any failure, there is a drop in my voice from around thirty percent to around twenty percent. And, oh, sorry, from 20% to around 15%, so it's a 30% drop. And, uh, so otherwise it's pretty much steady. Uh, since there was no actual fixes, and uh… In permanent hang numbers, macOS had around. 2 to 3%. And now it's around steady. 1.9%.
|
||||||
- **15:28:24 · Michal Smaga:** Оф, Ол, Юзерс, Райт?
|
- **15:28:24 · Michal Smaga:** Оф, Ол, Юзерс, Райт?
|
||||||
|
- **15:28:26 · Alex Martemyanov:** Uh, yes, from daily users. So it's still higher than the 1% threshold. And, uh… We've got two fixes more to be released next week. The RR6 reload. which can. Probably cover up to 30% of every 6 total reports. Uh, because it's not… It's not proof that… So usually LR6, when it's happening, it should be recovered normally by the kit. And we are handling some kind of case when it's not recovering. And, uh… The other one is, uh, braid, which may… affect… Yes. The upgrade failures, some hacks caused by it. So speculating on it another like 20 or 30%.
|
||||||
|
- **15:29:25 · Michal Smaga:** So, those two errors are… On my like, in this week's internal. Yeah. So they will be public next week.
|
||||||
|
- **15:29:33 · Alex Martemyanov:** Yes, and those are not the main part of the failures we are getting. So they are expected to have some effect.
|
||||||
|
- **15:29:45 · Michal Smaga:** But not that big, yeah.
|
||||||
|
- **15:29:47 · Alex Martemyanov:** But it won't cross the one-person gap still, I think. Thank you.
|
||||||
|
- **15:29:55 · Michal Smaga:** Yeah, I think, like, probably the main reason is that we have exhausted the options, as you mentioned, like that, like the… The the the idle like thing was not we are not able to reproduce that one. So basically, probably, like…
|
||||||
|
- **15:30:10 · Alex Martemyanov:** Yeah, so…
|
||||||
|
- **15:30:12 · Michal Smaga:** Yeah. Okay. My gut feeling would be to just, like, probably provide a summary of the current state of the, like.
|
||||||
|
- **15:30:19 · Michal Smaga:** the reporting have maybe some kind of, like, a dashboard and and probably, like, revisit it next week when we have the numbers for the the last part of the fixes.
|
||||||
|
- **15:30:26 · Alex Martemyanov:** Okay.
|
||||||
|
- **15:30:32 · Michal Smaga:** But probably I wouldn't hold the project any longer, like, for the next week.
|
||||||
|
- **15:30:39 · Alex Martemyanov:** Um… So my suggestion, my thinking of the project for this week. So, first of all, we have this… A lot of feature flags introduced, and the pixels, metrics, and so on, and all of them are basically marked as temporary, and some of them are not, uh… Not even being reported. So there is some cleanup, I guess, to be done. For example, this graveyard fix, we are catching some… percent, lower than one percent, probably, of all the reported events that can match it, but I would not activate it, looking at it, looking at the numbers, so it's… Certainly not the main issue. And there is another one. Um, yeah, anyways, uh, so I was thinking to clean it up. To see… to check the… Safe fixes and remove the picture flags for them. To leave the more controversial, dangerous things behind the feature flags. that we are living. And, uh… Maybe should we remove the recorder thing and episode diagnostics? Maybe not this week, but the week after. Because we are generally… We've collected all the data that it could provide, I guess. And the main portion of it that remains, the neat quake. I think it's generally not covered by it in any sense. And that's…
|
||||||
|
- **15:32:36 · Michal Smaga:** Yeah…
|
||||||
|
- **15:32:37 · Alex Martemyanov:** Certainly the main cause.
|
||||||
|
- **15:32:38 · Michal Smaga:** And… And apart from the episodes, the diagnostics, we… We have other pixels to measure the hangs, right? Or they are, like…
|
||||||
|
- **15:32:55 · Alex Martemyanov:** Currently, there is only this episode thing. We have, uh, the pixels for reports, for user reports, that is separating. And yeah, currently this episode metric only includes like background events, the state of things and. initiating, basically, event.
|
||||||
|
- **15:33:17 · Michal Smaga:** So we would also, like, lose the visibility into the the the, like, the permanent hangs. Right?
|
||||||
|
- **15:33:25 · Alex Martemyanov:** I'm sorry.
|
||||||
|
- **15:33:25 · Michal Smaga:** Like, the absolute…
|
||||||
|
- **15:33:27 · Alex Martemyanov:** Actually, I was thinking, maybe not finished planning yet, but I was thinking about redoing the whole reporting thing. Centering it in the gray zone, like, uncovered. Places. So, we have… We have a sequence, like, the content script being injected. sending INIT, WebKit receiving the INIT. In the ground, receiving and handling the event. The config fetching. And… And the website fetching, that may also indicate some kind of… So there is a… Uh, corner case, when the config cannot be fetched. For whatever reason, if it hangs, it would give this kind of symptoms. But… Normally it should not hang. So I would… I'm hesitating to fix it. Um, before we have any data on this. And… Generally, they should not cause permanent headaches. We're done, right? But yeah, currently we are not, uh, tied to the… Extension lifecycle, but we are tied to the page loading lifecycle. And, uh… Also, we don't… we don't have a baseline for the navigations, so we can't tell of which percent of navigation actually have the extension radar, but we are… Diminishing it instead against, like, daily users or installs metrics. So, uh… Still maybe has not. the correct numbers. We don't understand, like, the… the whole picture of it, like how many navigations do actually succeed. And, uh, where exactly… You've got this, uh… Meaning, like, failure, for example. Maybe it is the init, maybe it's somewhere in the WebKit internals, or maybe it's on our side, like waiting for the config fetch or whatever. And that would include the extension modification itself already, the measurement points. And probably some… Thank you. Not episodes, but a summary of, uh… The hung state, so whenever we catch it, when any chain of this, any part of the chain fails, we could report where it has failed.
|
||||||
|
- **15:36:18 · Michal Smaga:** Thanks.
|
||||||
|
- **15:36:23 · Alex Martemyanov:** So I'm thinking, well, maybe we should just, uh… Check this week to implement this metrics. redoing… And uh… then put it on hold, basically, before we receive. I mean, data related.
|
||||||
|
- **15:36:41 · Michal Smaga:** So you suggest, like, doing it for roughly a week, probably, more? Do the cleanup.
|
||||||
|
- **15:36:46 · Alex Martemyanov:** Yeah, so…
|
||||||
|
- **15:36:48 · Michal Smaga:** Do the cleanup and implement those, like, long term metrics to have some kind of observability in this gray zone.
|
||||||
|
- **15:36:55 · Alex Martemyanov:** Yeah. So, first of all, it should have… some permanent part, I guess, just so we keep it for… For catching the spikes or future breaks. And have some temporary metrics as well, but measuring the inside of the extension. because currently we have the okay. Everything seems fine. Background is loaded in less than 1 s. And then nothing. So and we don't have any idea and anything suspect that can be suspected for causing it.
|
||||||
|
- **15:37:25 · Michal Smaga:** Yeah. Mmhm.
|
||||||
|
- **15:37:36 · Alex Martemyanov:** So yeah, no hypothesis on that. What is actually breaking there? Except for a couple of issues inside the GS, but they are also… Kind of. Hmm. Do not seem so severe and causing this problems. So, that would be my point for today. Otherwise, I have no ideas for you.
|
||||||
|
- **15:38:04 · Michal Smaga:** Yeah, so… Yeah, so I think, like, the point that Russell is having is probably he wants to, like… have a clear understanding where the project will end, yeah, because we can, like, just probably, like, move it, move it, like, chase the, chase the, the, the, the finish line, like, forever.
|
||||||
|
- **15:38:20 · Alex Martemyanov:** Okay.
|
||||||
|
- **15:38:27 · Michal Smaga:** That's going away.
|
||||||
|
- **15:38:29 · Alex Martemyanov:** Well… I would highlight that we certainly have some improvement. We've found some actual bugs. Some of them are not even the case because they are really corner cases, but those are a real issue reproduced.
|
||||||
|
- **15:38:44 · Michal Smaga:** Yeah, I understand, yeah, yeah, so probably no one is questioning that, yeah, but just, like, to have, like, a more… like, condensed or, like, focused, uh, wrap-up for the project, yeah, so I would suggest, like, yeah, I like the plan, and I think it makes sense, yeah, so I would… I would bring the… Uh, yeah, so I think… Two things need to be done, like, one is to do a project update and, like, recap the current data, and summarize, like, where we landed at. And probably, I would start with, uh… with the DRI task, and, like, respond to Russell, and, like… Present the plan that, like, we have those, like, we have those fixes, the updates, and link to the project update, that you have the, like. Two more fixes, like, that will come up in the next version, and you propose the plan to, uh. Uh, have the final week of the project to, like, clean up the, like, all… Unnecessary things and clean up the diagnostics that we set up in place for, like, nailing down the issues. And suggest like to to metrics to like keep track of like where we're at with the permanent hangs and maybe better visibility in what's happening.
|
||||||
|
- **15:39:58 · Alex Martemyanov:** Introduce some permanent. Yep.
|
||||||
|
- **15:40:09 · Michal Smaga:** Yeah, and probably, like, we can see, like, what happens. Yeah, I… probably you want, like, if there's, like, a concrete plan that, like, this is the final week, and we just need to clean up the stuff, I don't think that there should be an issue. That's my feelings.
|
||||||
|
- **15:40:27 · Alex Martemyanov:** Okay, yeah, I'll do it today, then, before the… before the meeting, hopefully.
|
||||||
|
- **15:40:36 · Michal Smaga:** Yeah, and the meetings today are, like, the regular time, yeah.
|
||||||
|
- **15:40:40 · Alex Martemyanov:** I think it's one hour later shifted due to some times on things.
|
||||||
|
- **15:40:46 · Michal Smaga:** Yeah, but I think that…
|
||||||
|
- **15:40:46 · Alex Martemyanov:** All the meetings, actually.
|
||||||
|
- **15:40:52 · Michal Smaga:** Oh, didn't you have the time zone change for the winter time?
|
||||||
|
- **15:40:55 · Alex Martemyanov:** No, luckily I don't have time zone change, but unfortunately I do have the Easter.
|
||||||
|
- **15:40:57 · Michal Smaga:** Okay.
|
||||||
|
- **15:41:03 · Alex Martemyanov:** And the western time zones changing and shifting my schedules up to the late night.
|
||||||
|
- **15:41:08 · Michal Smaga:** the… Okay. Uh, yeah, so sounds good. Uh, like, I would ask you to, to, like, do the project update, and, like, provide the, the, the data analysis in some, some task, and, like, respond to us also.
|
||||||
|
- **15:41:19 · Alex Martemyanov:** Mm-hmm.
|
||||||
|
- **15:41:24 · Michal Smaga:** So we have the clear understanding, and we can discuss it on the meeting if he has any.
|
||||||
|
- **15:41:28 · Alex Martemyanov:** Okay, so…
|
||||||
|
- **15:41:29 · Michal Smaga:** Further questions, but it sounds good to me.
|
||||||
|
- **15:41:30 · Alex Martemyanov:** Probably. Probably the cleanup is, uh… I think, to be done anyways, whether it comes into the project or just as a follow-up, but it would be necessary to do it anyways, and it's not…
|
||||||
|
- **15:41:40 · Michal Smaga:** Yeah, yeah.
|
||||||
|
- **15:41:47 · Alex Martemyanov:** And it's not small, because there's multiple things in it, and it should be…
|
||||||
|
- **15:41:49 · Michal Smaga:** Yeah. Yeah, okay.
|
||||||
|
- **15:41:53 · Alex Martemyanov:** Some considerations whether things should be kept or cleaned up or enabled for everybody. So it seems like a natural part for a natural completion part of the project. But yeah, for the metrics. It also seems pretty natural to have some permanent metrics just in case we could detect if anything becomes worse. at some point. Okay, that would be it then. Thanks.
|
||||||
|
- **15:42:32 · Michal Smaga:** Cool. Yeah, that sounds good. Yeah, thank you for the chat.
|
||||||
|
- **15:42:40 · Alex Martemyanov:** Okay, see you, see you later.
|
||||||
-225
@@ -23,231 +23,6 @@
|
|||||||
"time": "21:01:11",
|
"time": "21:01:11",
|
||||||
"speaker": "Pikor",
|
"speaker": "Pikor",
|
||||||
"text": "Yep. Sorry, but I just realized that if you tap on that green camera icon on the status bar, then you have this recenter button within your preview and then there is the zoom. So you can zoom yourself like crazy."
|
"text": "Yep. Sorry, but I just realized that if you tap on that green camera icon on the status bar, then you have this recenter button within your preview and then there is the zoom. So you can zoom yourself like crazy."
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:01:29",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Yeah, that's the effect."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:01:32",
|
|
||||||
"speaker": "Pikor",
|
|
||||||
"text": "So this is fun. Yeah, Jorge likes it as well, as I see."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:01:36",
|
|
||||||
"speaker": "Andre Martingo",
|
|
||||||
"text": "Hello!"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:01:39",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Hey, folks."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:01:40",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "Igor, please do that at appropriate moments when you're giving your reaction to an inappropriate comment."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:01:45",
|
|
||||||
"speaker": "Pikor",
|
|
||||||
"text": "Okay."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:01:45",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Yeah, right. All right, looks like we got a pretty good crew when we get started. Quick cruise through the rotations as usual. iOS Triage, Pico, you're on this week. Cue the Zooming. Jacek, you're up next week. Followed by L the week after that. There it is. Mac OS triage is on this week, followed by Sabrina and then Fernando. For iOS release, Pao is handling this week, followed by Hanyu and then Sam. And then macOS release, I actually don't know who's on because the task is closed this week already, but I hope that means that they know. Juan will be covering next week and then Dominic the week after that. As usual, any collisions, please let me know. Yep."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:02:58",
|
|
||||||
"speaker": "Fernando Bunn",
|
|
||||||
"text": "Pardon? Yeah, I'm out for two weeks, starting next week, so… No. Yep."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:08",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Okay, we should find you a replacement for that week."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:12",
|
|
||||||
"speaker": "Fernando Bunn",
|
|
||||||
"text": "All right."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:13",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Okay. See if you can trade with somebody nearby as a starting place."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:15",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "Equip."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:19",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Cool. Thanks."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:19",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "Sorry to interrupt. Jonathan, a quick question about the…"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:21",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Yep."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:23",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "the holiday rotation, so I just… I'm one of the people with a big X beside my name, so I definitely won't be there, but just to make sure I understand, because I read the task, but there's no ask from, uh, the people who are actually on rotation at the moment, right? I mean, you are…"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:25",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Yeah. Yep."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:38",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "As far as I understand, you're taking care of finding…"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:41",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "I'm gonna start some conversations next. I was gonna do that yesterday when I was out. Um, but yeah, I think the task is settled enough that the next step is to… to try to get those conversations rolling, and I'll do that piece. Um, I'll probably need your help to think about which other week you would take to swap in, but…"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:41",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "Or you're taking care of something."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:03:56",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Yeah, I'm gonna get those rolling soon. And Pete, you are up first on the agenda."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:04:11",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "Yeah, thanks. I forget what I wrote. Oh, yeah. OJ will begin — this is just an announcement or an FYI, an early announcement because we're just thinking about this now. So OJ is going to begin testing auto. So the auto deductor go harness, as it's also known, we're going to start testing that in OJ. And what that practically means is, so, uh, within OJ, we'll identify projects or tasks on the critical path that we think are low risk, so, like, low risk, maybe, you know, meaning limited blast radius or well-defined requirements, and then the. uh, project DRI, so, like, an engineer, one of the engineers in OJ will hand that work to Otto, so not that different to what everybody's doing at the moment in terms of just using agents to build. Uh, Otto would implement and open the PR, and then the project DRI, so the same person that handed the work to Otto. They will review the pull request and they will leave any comments which Auto or the agent will then deal with and will iterate until the PRO is ready and the project DRA will then merge the PRO. So that's the current plan. So I guess the difference there is that, you know, rather than the current state, where we work with agents to build it, and we ask another person to review it. it's the project DRI, the person who gave the prompt to Otto, and let it work completely on its own. They'll review the PR. So that's where we're going to start testing. This is happening in some other objectives, it's already happened in some other objectives, or started, but it's a way to test. um… I guess, auto."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:05:56",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "All right, thanks for the heads up."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:05:57",
|
|
||||||
"speaker": "Pete Smith",
|
|
||||||
"text": "That's the thought that. Yeah."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:06:00",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "I'll look forward to hearing how it goes, for sure. Uh, yeah, Sick, your topic is next."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:06:12",
|
|
||||||
"speaker": "JŁ",
|
|
||||||
"text": "Okay, so I had an idea for how to better present apple release strain. So we still have it in like a static form on Asana and we know how it works, but often people outside of our team don't. And I decided to use an artifact for that. It's still a work in progress, so I will expect some feedback. Let me share. Is it this one? Okay, I think it's this one. So I created this artifact that, like, you can share with anyone. Eh… And basically what it shows you is what version is currently in the store. What version is, like, reaching users, like, on the given day, and what version is in internal testing? You can tap on any version, and you can see, like, all the info that was in Asana, so, like, release notes, when it was cut from main, how it's rolling out, when it is going to be released to any person. You can see what flags depend on that version, and you can see all the changes here as well. Everything is, like, clickable, including, like, pull requests, so you can reach that. Mm-hmm. And… Then you have a schedule in a list, so you can show it to someone, so they know when each thing is going to be released. You can. change the view to the timeline, and on hover see when anything is released. And it includes the holiday break, too. On the calendar, you can also see it in this form, so you see when every version is going to be released. like each month. And then there are feature flags. So you see what feature flags are. On what percentage? And… So again, you, you know, if something is released to people or not, and you can also search. So for instance, I know that Bartosz was doing some background task thing, and if you like, search for it either by URL or link. It will show the thing, it will show you in which version it was released, what's the current running out status. And also you will see feature flags that were included. So you see this feature flag is only like released to 1% for this feature. So, uh, you can quickly see what is going on. And it also supports macOS, so, uh… yeah, like, probably still some things to polish. let me know if it's going to work for you or not. Thank you."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:08:55",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "I think it's probably going to work for pretty much everybody. That's amazing. Well done. Um, I think you might want to, you know, maybe have us spend a couple of weeks with it to make sure it's not…"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:08:59",
|
|
||||||
"speaker": "Andre Martingo",
|
|
||||||
"text": "Thank you."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:09:06",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "lying to people, you know, they're not… they're not using it, but yeah, I totally agree. People are suggesting sharing it at all hands, adding it to go.go.com, like, I think you should take it to a point where it's… where we're confident in it, and then definitely share it."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:09:07",
|
|
||||||
"speaker": "JŁ",
|
|
||||||
"text": "Yeah, yeah, sure."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:09:19",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "And think about connecting to Android and Windows as well once you're ready."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:09:25",
|
|
||||||
"speaker": "JŁ",
|
|
||||||
"text": "Sure, sure."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:09:26",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Bartek, if you want to create an incident for us, you're welcome to."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:09:26",
|
|
||||||
"speaker": "Andre Martingo",
|
|
||||||
"text": "Good. Just a comment. Super nice. It's interesting because I was checking as well. I don't know if you saw the Windows release what they call a comment center. They have kind of an internal website. that contains… like, I think what you built plus more steps, um, and it was something that I was looking at once I finished Smartling, so I will get in touch with you. I don't know… at this point, what's the difference between an artifact or a web page itself, uh, inside of DuckDuckGo, but it's something, uh, that maybe we can… we can chat afterwards. Uh, but it's quite nice as well. Nice."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:10:18",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Cool. Thank you. Jacek, when you're ready, it'd be great if you could share that with the team as a starting place, obviously."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:10:24",
|
|
||||||
"speaker": "JŁ",
|
|
||||||
"text": "Sure, sure."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:10:25",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Cool. Uh, Andre, your SmartLink update is up next."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:10:31",
|
|
||||||
"speaker": "Andre Martingo",
|
|
||||||
"text": "So yeah, just a quick update. So everything should be done. I made a post. Me and Pao, we already kind of enabled or kind of disabled the Avenger warning. So I hope you'll start merging into main your untranslated strings. I have here as well an artifact since it's the new cool thing to do. Um… So basically, it was just an exploration about… how much we were spending per release in SmartLink, so maybe this will help us to move to AI, but other than that, it was to track. when are we starting using, kind of, the new flow. So, basically, here you have a graph of spending per release in terms of dollars, and the blue bars are, kind of, our new automatic flow. Um, so we are spending an average of, like, 500, 400, depending on the release. Um, and basically, here in this artifact, you can see, like. Which ones are the jobs that got merged automatically, or which ones are manual, so we can still… we can track, like, the usage of our new flow. And hopefully, maybe with some data, we can start thinking if it would be possible to move to AI, compare the prices, or maybe just A/B test for one language. It's something for the future we can maybe try to push for. But for now, it's just to collect… some stats so we can have better discussions over it."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:12:10",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "That's cool. Thank you. Yeah, if you want to share that somewhere so that people can find it on a regular basis, that'd be great. How have you seen, I know it's early, but on the nightly untranslated pieces, how quickly have those jobs from SmartLine been coming back?"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:12:31",
|
|
||||||
"speaker": "Andre Martingo",
|
|
||||||
"text": "Sorry, which ones?"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:12:33",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "You're basically doing a job every night with whatever was untranslated, right?"
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:12:39",
|
|
||||||
"speaker": "Andre Martingo",
|
|
||||||
"text": "Also, but we do every merge commit as well. So basically, once you merge, we trigger a job, and the nightly will… if something goes wrong, it will pick up as well. So the blue bars are kind of the new flow, yeah."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:12:48",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Oh, okay. Well, either one. Yeah. Presumably, those are usually smaller jobs for SmartLink. I was wondering how quickly they'd been turning those around recently."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:13:00",
|
|
||||||
"speaker": "Andre Martingo",
|
|
||||||
"text": "Yeah, actually, we have another one where we have the kind of turnover of PRs for Smartling, but I will share afterwards once we have more data. We didn't have much, no."
|
|
||||||
},
|
|
||||||
{
|
|
||||||
"time": "21:13:07",
|
|
||||||
"speaker": "Jonathan Jackson",
|
|
||||||
"text": "Mm-hmm. Okay. Yeah, I know it's early still. All right, cool. Thank you. Um… That's our planned agenda for the day. Any other thoughts, questions, comments, things to share, demos? All right. A quick meeting today, but some good stuff. Thank you, everyone. Good to see you, and I hope you have a good day."
|
|
||||||
}
|
}
|
||||||
]
|
]
|
||||||
}
|
}
|
||||||
-45
@@ -15,48 +15,3 @@ source: "https://us01mynotes.zoom.us/GJhjFoDDRY6PkqEUdM-YFg"
|
|||||||
- **21:01:06 · Jonathan Jackson:** Pico, you're uh… Camera Zoom is doing, I don't know, it's like an old 1950s horror movie.
|
- **21:01:06 · Jonathan Jackson:** Pico, you're uh… Camera Zoom is doing, I don't know, it's like an old 1950s horror movie.
|
||||||
- **21:01:10 · Pete Smith:** And you…
|
- **21:01:10 · Pete Smith:** And you…
|
||||||
- **21:01:11 · Pikor:** Yep. Sorry, but I just realized that if you tap on that green camera icon on the status bar, then you have this recenter button within your preview and then there is the zoom. So you can zoom yourself like crazy.
|
- **21:01:11 · Pikor:** Yep. Sorry, but I just realized that if you tap on that green camera icon on the status bar, then you have this recenter button within your preview and then there is the zoom. So you can zoom yourself like crazy.
|
||||||
- **21:01:29 · Jonathan Jackson:** Yeah, that's the effect.
|
|
||||||
- **21:01:32 · Pikor:** So this is fun. Yeah, Jorge likes it as well, as I see.
|
|
||||||
- **21:01:36 · Andre Martingo:** Hello!
|
|
||||||
- **21:01:39 · Jonathan Jackson:** Hey, folks.
|
|
||||||
- **21:01:40 · Pete Smith:** Igor, please do that at appropriate moments when you're giving your reaction to an inappropriate comment.
|
|
||||||
- **21:01:45 · Pikor:** Okay.
|
|
||||||
- **21:01:45 · Jonathan Jackson:** Yeah, right. All right, looks like we got a pretty good crew when we get started. Quick cruise through the rotations as usual. iOS Triage, Pico, you're on this week. Cue the Zooming. Jacek, you're up next week. Followed by L the week after that. There it is. Mac OS triage is on this week, followed by Sabrina and then Fernando. For iOS release, Pao is handling this week, followed by Hanyu and then Sam. And then macOS release, I actually don't know who's on because the task is closed this week already, but I hope that means that they know. Juan will be covering next week and then Dominic the week after that. As usual, any collisions, please let me know. Yep.
|
|
||||||
- **21:02:58 · Fernando Bunn:** Pardon? Yeah, I'm out for two weeks, starting next week, so… No. Yep.
|
|
||||||
- **21:03:08 · Jonathan Jackson:** Okay, we should find you a replacement for that week.
|
|
||||||
- **21:03:12 · Fernando Bunn:** All right.
|
|
||||||
- **21:03:13 · Jonathan Jackson:** Okay. See if you can trade with somebody nearby as a starting place.
|
|
||||||
- **21:03:15 · Pete Smith:** Equip.
|
|
||||||
- **21:03:19 · Jonathan Jackson:** Cool. Thanks.
|
|
||||||
- **21:03:19 · Pete Smith:** Sorry to interrupt. Jonathan, a quick question about the…
|
|
||||||
- **21:03:21 · Jonathan Jackson:** Yep.
|
|
||||||
- **21:03:23 · Pete Smith:** the holiday rotation, so I just… I'm one of the people with a big X beside my name, so I definitely won't be there, but just to make sure I understand, because I read the task, but there's no ask from, uh, the people who are actually on rotation at the moment, right? I mean, you are…
|
|
||||||
- **21:03:25 · Jonathan Jackson:** Yeah. Yep.
|
|
||||||
- **21:03:38 · Pete Smith:** As far as I understand, you're taking care of finding…
|
|
||||||
- **21:03:41 · Jonathan Jackson:** I'm gonna start some conversations next. I was gonna do that yesterday when I was out. Um, but yeah, I think the task is settled enough that the next step is to… to try to get those conversations rolling, and I'll do that piece. Um, I'll probably need your help to think about which other week you would take to swap in, but…
|
|
||||||
- **21:03:41 · Pete Smith:** Or you're taking care of something.
|
|
||||||
- **21:03:56 · Jonathan Jackson:** Yeah, I'm gonna get those rolling soon. And Pete, you are up first on the agenda.
|
|
||||||
- **21:04:11 · Pete Smith:** Yeah, thanks. I forget what I wrote. Oh, yeah. OJ will begin — this is just an announcement or an FYI, an early announcement because we're just thinking about this now. So OJ is going to begin testing auto. So the auto deductor go harness, as it's also known, we're going to start testing that in OJ. And what that practically means is, so, uh, within OJ, we'll identify projects or tasks on the critical path that we think are low risk, so, like, low risk, maybe, you know, meaning limited blast radius or well-defined requirements, and then the. uh, project DRI, so, like, an engineer, one of the engineers in OJ will hand that work to Otto, so not that different to what everybody's doing at the moment in terms of just using agents to build. Uh, Otto would implement and open the PR, and then the project DRI, so the same person that handed the work to Otto. They will review the pull request and they will leave any comments which Auto or the agent will then deal with and will iterate until the PRO is ready and the project DRA will then merge the PRO. So that's the current plan. So I guess the difference there is that, you know, rather than the current state, where we work with agents to build it, and we ask another person to review it. it's the project DRI, the person who gave the prompt to Otto, and let it work completely on its own. They'll review the PR. So that's where we're going to start testing. This is happening in some other objectives, it's already happened in some other objectives, or started, but it's a way to test. um… I guess, auto.
|
|
||||||
- **21:05:56 · Jonathan Jackson:** All right, thanks for the heads up.
|
|
||||||
- **21:05:57 · Pete Smith:** That's the thought that. Yeah.
|
|
||||||
- **21:06:00 · Jonathan Jackson:** I'll look forward to hearing how it goes, for sure. Uh, yeah, Sick, your topic is next.
|
|
||||||
- **21:06:12 · JŁ:** Okay, so I had an idea for how to better present apple release strain. So we still have it in like a static form on Asana and we know how it works, but often people outside of our team don't. And I decided to use an artifact for that. It's still a work in progress, so I will expect some feedback. Let me share. Is it this one? Okay, I think it's this one. So I created this artifact that, like, you can share with anyone. Eh… And basically what it shows you is what version is currently in the store. What version is, like, reaching users, like, on the given day, and what version is in internal testing? You can tap on any version, and you can see, like, all the info that was in Asana, so, like, release notes, when it was cut from main, how it's rolling out, when it is going to be released to any person. You can see what flags depend on that version, and you can see all the changes here as well. Everything is, like, clickable, including, like, pull requests, so you can reach that. Mm-hmm. And… Then you have a schedule in a list, so you can show it to someone, so they know when each thing is going to be released. You can. change the view to the timeline, and on hover see when anything is released. And it includes the holiday break, too. On the calendar, you can also see it in this form, so you see when every version is going to be released. like each month. And then there are feature flags. So you see what feature flags are. On what percentage? And… So again, you, you know, if something is released to people or not, and you can also search. So for instance, I know that Bartosz was doing some background task thing, and if you like, search for it either by URL or link. It will show the thing, it will show you in which version it was released, what's the current running out status. And also you will see feature flags that were included. So you see this feature flag is only like released to 1% for this feature. So, uh, you can quickly see what is going on. And it also supports macOS, so, uh… yeah, like, probably still some things to polish. let me know if it's going to work for you or not. Thank you.
|
|
||||||
- **21:08:55 · Jonathan Jackson:** I think it's probably going to work for pretty much everybody. That's amazing. Well done. Um, I think you might want to, you know, maybe have us spend a couple of weeks with it to make sure it's not…
|
|
||||||
- **21:08:59 · Andre Martingo:** Thank you.
|
|
||||||
- **21:09:06 · Jonathan Jackson:** lying to people, you know, they're not… they're not using it, but yeah, I totally agree. People are suggesting sharing it at all hands, adding it to go.go.com, like, I think you should take it to a point where it's… where we're confident in it, and then definitely share it.
|
|
||||||
- **21:09:07 · JŁ:** Yeah, yeah, sure.
|
|
||||||
- **21:09:19 · Jonathan Jackson:** And think about connecting to Android and Windows as well once you're ready.
|
|
||||||
- **21:09:25 · JŁ:** Sure, sure.
|
|
||||||
- **21:09:26 · Jonathan Jackson:** Bartek, if you want to create an incident for us, you're welcome to.
|
|
||||||
- **21:09:26 · Andre Martingo:** Good. Just a comment. Super nice. It's interesting because I was checking as well. I don't know if you saw the Windows release what they call a comment center. They have kind of an internal website. that contains… like, I think what you built plus more steps, um, and it was something that I was looking at once I finished Smartling, so I will get in touch with you. I don't know… at this point, what's the difference between an artifact or a web page itself, uh, inside of DuckDuckGo, but it's something, uh, that maybe we can… we can chat afterwards. Uh, but it's quite nice as well. Nice.
|
|
||||||
- **21:10:18 · Jonathan Jackson:** Cool. Thank you. Jacek, when you're ready, it'd be great if you could share that with the team as a starting place, obviously.
|
|
||||||
- **21:10:24 · JŁ:** Sure, sure.
|
|
||||||
- **21:10:25 · Jonathan Jackson:** Cool. Uh, Andre, your SmartLink update is up next.
|
|
||||||
- **21:10:31 · Andre Martingo:** So yeah, just a quick update. So everything should be done. I made a post. Me and Pao, we already kind of enabled or kind of disabled the Avenger warning. So I hope you'll start merging into main your untranslated strings. I have here as well an artifact since it's the new cool thing to do. Um… So basically, it was just an exploration about… how much we were spending per release in SmartLink, so maybe this will help us to move to AI, but other than that, it was to track. when are we starting using, kind of, the new flow. So, basically, here you have a graph of spending per release in terms of dollars, and the blue bars are, kind of, our new automatic flow. Um, so we are spending an average of, like, 500, 400, depending on the release. Um, and basically, here in this artifact, you can see, like. Which ones are the jobs that got merged automatically, or which ones are manual, so we can still… we can track, like, the usage of our new flow. And hopefully, maybe with some data, we can start thinking if it would be possible to move to AI, compare the prices, or maybe just A/B test for one language. It's something for the future we can maybe try to push for. But for now, it's just to collect… some stats so we can have better discussions over it.
|
|
||||||
- **21:12:10 · Jonathan Jackson:** That's cool. Thank you. Yeah, if you want to share that somewhere so that people can find it on a regular basis, that'd be great. How have you seen, I know it's early, but on the nightly untranslated pieces, how quickly have those jobs from SmartLine been coming back?
|
|
||||||
- **21:12:31 · Andre Martingo:** Sorry, which ones?
|
|
||||||
- **21:12:33 · Jonathan Jackson:** You're basically doing a job every night with whatever was untranslated, right?
|
|
||||||
- **21:12:39 · Andre Martingo:** Also, but we do every merge commit as well. So basically, once you merge, we trigger a job, and the nightly will… if something goes wrong, it will pick up as well. So the blue bars are kind of the new flow, yeah.
|
|
||||||
- **21:12:48 · Jonathan Jackson:** Oh, okay. Well, either one. Yeah. Presumably, those are usually smaller jobs for SmartLink. I was wondering how quickly they'd been turning those around recently.
|
|
||||||
- **21:13:00 · Andre Martingo:** Yeah, actually, we have another one where we have the kind of turnover of PRs for Smartling, but I will share afterwards once we have more data. We didn't have much, no.
|
|
||||||
- **21:13:07 · Jonathan Jackson:** Mm-hmm. Okay. Yeah, I know it's early still. All right, cool. Thank you. Um… That's our planned agenda for the day. Any other thoughts, questions, comments, things to share, demos? All right. A quick meeting today, but some good stuff. Thank you, everyone. Good to see you, and I hope you have a good day.
|
|
||||||
@@ -86,8 +86,8 @@
|
|||||||
"note_modified_at": "2026-10-16T15:51:25.415Z",
|
"note_modified_at": "2026-10-16T15:51:25.415Z",
|
||||||
"source_url": "https://us01mynotes.zoom.us/klekRLtLSEuMv0PCCPCjtA",
|
"source_url": "https://us01mynotes.zoom.us/klekRLtLSEuMv0PCCPCjtA",
|
||||||
"folder": "2026-10-16 All-Company Meeting 2026-10-16 21 31(GMT+6 00) klekRLtLSEuMv0PCCPCjtA",
|
"folder": "2026-10-16 All-Company Meeting 2026-10-16 21 31(GMT+6 00) klekRLtLSEuMv0PCCPCjtA",
|
||||||
"sha256": "86407aecffbf09effe200b26e80a01887ee197ba99d5caea70c124743fbee3fb",
|
"sha256": "34ef73307fe86cc2d6f98caf57a719ae46937fb2d0efea3a18e03e21a42f1e00",
|
||||||
"cues": 3
|
"cues": 20
|
||||||
},
|
},
|
||||||
"AEmGnEdJQfCMmCjwLyGtGA": {
|
"AEmGnEdJQfCMmCjwLyGtGA": {
|
||||||
"title": "Зал персональной конференции Alex Martemyanov 2026-10-09 18:51(GMT+6:00)",
|
"title": "Зал персональной конференции Alex Martemyanov 2026-10-09 18:51(GMT+6:00)",
|
||||||
@@ -254,8 +254,8 @@
|
|||||||
"note_modified_at": "2026-11-03T09:26:09.911Z",
|
"note_modified_at": "2026-11-03T09:26:09.911Z",
|
||||||
"source_url": "https://us01mynotes.zoom.us/qQ2bzU5KTiCGluiBYA-ROA",
|
"source_url": "https://us01mynotes.zoom.us/qQ2bzU5KTiCGluiBYA-ROA",
|
||||||
"folder": "2026-11-03 Michal Smaga's Personal Meeting Room qQ2bzU5KTiCGluiBYA-ROA",
|
"folder": "2026-11-03 Michal Smaga's Personal Meeting Room qQ2bzU5KTiCGluiBYA-ROA",
|
||||||
"sha256": "646f7d0a50c8ad5f3297b8d28175a5649505ae798ba57fca5340d26536f0dd14",
|
"sha256": "dafb593da5838d0c82fa22918d30cd0e58dbe5d9a592b8a8c2c997d083a80638",
|
||||||
"cues": 3
|
"cues": 59
|
||||||
},
|
},
|
||||||
"FZA-Gp3ATxiQ7cQaU5u5tg": {
|
"FZA-Gp3ATxiQ7cQaU5u5tg": {
|
||||||
"title": "O-E 2026",
|
"title": "O-E 2026",
|
||||||
@@ -386,8 +386,8 @@
|
|||||||
"note_modified_at": "2026-11-17T15:00:56.296Z",
|
"note_modified_at": "2026-11-17T15:00:56.296Z",
|
||||||
"source_url": "https://us01mynotes.zoom.us/GJhjFoDDRY6PkqEUdM-YFg",
|
"source_url": "https://us01mynotes.zoom.us/GJhjFoDDRY6PkqEUdM-YFg",
|
||||||
"folder": "2026-11-17 Apple Weekly Team Meeting - Zoom ID GJhjFoDDRY6PkqEUdM-YFg",
|
"folder": "2026-11-17 Apple Weekly Team Meeting - Zoom ID GJhjFoDDRY6PkqEUdM-YFg",
|
||||||
"sha256": "cd06af7a61ede6d4cd5d78b6ed94d9c6650c99a66d637cdf0e2eba0fef6390c5",
|
"sha256": "22c2c60c7a01fe8692c3e347c12beaa43102161292c468734320db978739bb11",
|
||||||
"cues": 48
|
"cues": 3
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
@@ -1 +1 @@
|
|||||||
{"observed_at":"2026-11-19T12:35:46.882Z","scanned":31,"imported":4,"skipped":27,"withoutTranscript":0,"coverage":"complete","sections":[{"name":"Owned by me","pages":5,"notes":86,"complete":true,"terminal_cursor":true},{"name":"Shared with me","pages":1,"notes":0,"complete":true,"terminal_cursor":true},{"name":"Starred","pages":1,"notes":0,"complete":true,"terminal_cursor":true}],"discovered":86,"gaps":[],"recovered_transcripts":0,"policy":"reread_all_owned_shared_and_starred_notes_since_cutoff","since":"2026-09-30"}
|
{"observed_at":"2026-11-19T12:38:10.672Z","scanned":31,"imported":3,"skipped":28,"withoutTranscript":0,"coverage":"complete","sections":[{"name":"Owned by me","pages":5,"notes":86,"complete":true,"terminal_cursor":true},{"name":"Shared with me","pages":1,"notes":0,"complete":true,"terminal_cursor":true},{"name":"Starred","pages":1,"notes":0,"complete":true,"terminal_cursor":true}],"discovered":86,"gaps":[],"recovered_transcripts":0,"policy":"reread_all_owned_shared_and_starred_notes_since_cutoff","since":"2026-09-30"}
|
||||||
Reference in new issue
Block a user