Journal

Today

Bottleneck: arrivals from the audiences. Best evidence is unchanged since Oct 14 (no new read on a weekday): 0 of 11 outside Default Alive players on build 20 or later had given an email, and the game has no remote push (PostHog, read Oct 14). The enginery.co site had 0 subscribers from 63 visitors (read Oct 14). New since yesterday: the build that asks for the email earlier is now in TestFlight as 1.0.1 (40).

1. Read the App Store agreements page and renew both agreements before Oct 23 (DAVE-171) → run the business: both agreements end in 7 days, four days before the launch. Every listing, and the launch, depends on them. Last read Sept 26.
2. Play 1.0.1 (40), then swap it into App Store review in place of 1.0 (32) by Oct 18 to 20 (SIM-76) → row 31, row 34, row 32, Default Alive: tomorrow. Build 40 holds the earlier email ask, the one-time ask for players who passed it, and the role and room questions. It passed TestFlight beta review on Oct 15. Before the swap: the App Privacy answer for the role question, and the store version string changes from 1.0 to 1.0.1. This build turns launch-week arrivals into people we can reach again.
3. Send the reply to the tester whose two fixes shipped (SIM-77) → Default Alive: tomorrow. The fixes cleared beta review in build 37, and build 40 carries them too. The reply is now free to go. A kept promise to an engaged player is how spread starts, and spread is the only arrival source that is not our own posts.
4. Commit and merge the systems-thinking video page, then send one email to the 136 by hand (DPC-13) → row 35. The page is still uncommitted on the davepaola.com branch systems-thinking-video. Count replies and clicks at 7 days. It tells us if the one list with people in it still answers.
5. Open the two accounts for the pet-owner character and connect TikTok before the Oct 24 launch (PMR-64) → row 36. All 17 launch and week 1 videos are made. Without the accounts, the character cannot post, and the pet-owner audience has no way back.

Canaries

• Fired: the App Store agreements end Oct 23, 7 days out, last read Sept 26 (move 1).
• Default Alive 1.0 is in App Store review, 2.3 days in the queue. Clear (the bar is 9 days).
• Uploaded and not submitted for more than seven days: clear. Default Alive 1.0.1 (40) was uploaded Oct 15 and is in TestFlight; it waits for the swap in move 2.
• No app moved to “ready for distribution” since the last standup.

Since yesterday

• The Default Alive build with the earlier email ask, the one-time ask for players who passed it, and the role and room questions is merged and in TestFlight as 1.0.1 (40). It passed beta review at once.
• The two tester-reported fixes cleared beta review in 1.0.1 (37), so the tester can now get them.
• The pet-owner character has all 17 videos made for launch week and week 1, after a rewrite of every script into natural speech. Its launch is Sat Oct 24.
• The Android port of Default Alive passed its first full build: four screen tasks merged, every check green, and the second wave of five tasks started.

Today

Bottleneck: arrivals from the audiences. Best evidence: the people who arrive cannot be reached again. On Oct 14, 0 of 11 outside Default Alive players on build 20 or later had given an email, and the game has no remote push (PostHog, read Oct 14). The enginery.co site had 0 subscribers from 63 visitors (read Oct 14). The 136-person davepaola.com list has had no send in 46 or more days.

1. Read the App Store agreements page and renew both agreements before Oct 23 (DAVE-171) → run the business: both agreements end Oct 23, four days before the launch. Every listing, and the launch, depends on them. Last read Sept 26.
2. Ship build 34 to TestFlight, play it, then put it in App Store review in place of 1.0 (32) by Oct 18 to 20 (SIM-76) → row 31, row 34, row 32, Default Alive: tomorrow. Build 34 holds the earlier email ask (at the first paying customer), a one-time ask for players who passed it, and the role and room questions. It turns launch-week arrivals into a stock we can reach. The code is done on a branch; the App Privacy answer for the role question is owed before submission.
3. Close the loop with the tester whose two fixes shipped in TestFlight build 37 (SIM-77) → Default Alive: tomorrow. Build 37 waits for Apple’s beta review. When it clears, the reply goes out and the fixes are recorded. A kept promise to an engaged player is how spread starts, and spread is the only arrival source that is not our own posts.
4. Send one email to the 136 with the systems-thinking video (DPC-13) → row 35. Merge the video page first, then a hand send. Count replies and clicks at 7 days. It tells us if the one list with people in it still answers.
5. Watch the three pilot videos for the pet-owner character and give the Gate 1 call (PMR-64) → row 36. The pilot used 147.75 of 150 credits. The character is the only way back to the pet-owner audience, because Pet Med Reminder keeps no addresses.

Canaries

• Fired: the App Store agreements end Oct 23, 8 days out, last read Sept 26 (move 1).
• Default Alive 1.0 is in App Store review, 1.3 days in the queue. Clear (the bar is 9 days).
• Uploaded and not submitted for more than seven days: clear. Default Alive 1.0.1 (37) is in TestFlight beta review, uploaded Oct 14.
• No app moved to “ready for distribution” since the last standup.

Since yesterday

• Default Alive 1.0 went into App Store review, with the Launch Week event (Oct 27 to Nov 2), on manual release.
• Two tester-reported fixes reached TestFlight as 1.0.1 (37), built in the cloud. One binary now serves TestFlight and the App Store.
• The email ask, the role question and the room question are built for the next build. The ask moved earlier, to the first paying customer, so players who never close a seed also see it.
• The systems-thinking explainer v2 is rendered (4:05), and its page is built for davepaola.com. An email to the list is drafted as a vault file.
• A plan for an AI character for pet owners was adopted, and a three-video pilot was rendered.
• The enginery.co link fix that keeps the room tag on the way to the App Store is live.
• Android work started again in the cloud: the shared game core now compiles for Android, and four screen tasks run in parallel.

Today

1. Send the Series A build to App Store review as 1.0 (SIM-20) → D5. Due by Thursday Oct 15. Version 1.0 reads “prepare for submission”, so the old approved build is off it, and builds through 28 are uploaded. The steps left are to play the TestFlight build, attach the App Store build and submit.
2. Create the Default Alive app in Play Console and open the Android tester sign-up (SIM-68) → D7. The target is 12 names by Oct 18, from 0. Google’s 14-day test clock starts only at the twelfth tester.
3. Write the real-player rule and its same-count check into the cycle instruments (SIM-69) → D6. The new dashboard is live and its counts match the old read. The check is not yet written where the Friday review reads it.

Canaries

• Fired: Pet Med Reminder 1.12.1 cleared App Store review and is out (it releases on approval). It unblocks no work this cycle, because Pet Med Reminder carries no objective.
• Fired, carried: the App Store agreements expire Oct 23, 9 days out, last read Sept 26. Every listing depends on them.
• The App Store status tool timed out on Apple’s API three times this morning. The states above come from the release check, which reached Apple. Shown only for three apps.
• Uploaded and not submitted for more than seven days: clear.

Since yesterday

• Cycle 3 adopted: one product, one week, Default Alive only. Seven key results across current players, the share loop and launch readiness.
• Default Alive shipped five TestFlight builds in a day: the garage one step at a time and tester fixes; the daily garage, seed codes and pixel share cards; the Series A stage; focus or fight with switching costs; and the game’s own math functions, so Android numbers match. Series B followed in the evening and waits for beta review.
• The real-player view and the Growth loop dashboard went live and give the same counts as the old filter.
• The 1.0 submission was paused in the evening to include Series B. A new App Store build 1.0 (28) was prepared for it; the submission waits for a go.

Today

1. Write the Cycle 3 OKRs (DAVE-218) → no live cycle. Yesterday’s close is the input: 1 of 10 key results met, rates held where people arrived, every count missed on arrivals.
2. Decide whether to release Default Alive 1.0, now approved by Apple (SIM-20) → run the business. It reads approved and waiting for a manual release. Nothing is public until someone presses release.
3. In one App Store Connect sitting, read the Pet Med Reminder store figures (PMR-37, PMR-57) and the agreements page (DAVE-171) → run the business. The listing hold ended Oct 10 and the ad test ended yesterday. The agreements expire Oct 23 by the last reading.

Canaries

• Fired: Default Alive 1.0 cleared App Store review. It moved from waiting for review (10.3 days yesterday) to approved, pending a manual release. Task 2.
• Fired: Pet Med Reminder store reads, carried from yesterday. Both windows have closed and neither is read. Task 3.
• Fired: App Store agreements expire Oct 23, 10 days out, and the last reading is Sept 26. Every listing depends on them. Task 3.
• Pet Med Reminder 1.12.1: waiting for review, 7.3 days (submitted Oct 6). Inside the 9-day budget. It releases automatically on approval.
• Uploaded and not submitted for more than seven days: clear.

Since yesterday

• Cycle 2 closed: 1 of 10 key results met. Default Alive players come back the next day on their own; every other key result missed, most on arrivals.
• Default Alive 1.0 was approved by App Store review after about 10 to 11 days in the queue.
• Nothing else changed the list.

Today

1. Write the Cycle 3 OKRs (DAVE-218) → no live cycle. Today’s close is the main input: one key result of ten met. Pet Med Reminder and Paroletta beat their first-win rates where people arrived, and every count missed because few people arrived.
2. Look at Default Alive 1.0 in App Store Connect: it is 10.3 days in the review queue (SIM-20) → run the business. The budget is 5 to 9 days. Past that, a submission can be stuck rather than queued. It uses manual release, so approval does not publish it.
3. Read the Pet Med Reminder store figures for the close (PMR-37, PMR-57) → run the business. The listing hold ended Oct 10 and the ad test ends today at noon. Both reads need the signed-in App Store Connect: impressions for Sept 12 to Oct 10, and downloads by ad campaign for Oct 5 to 12. The subtitle change waits for the first read.

Canaries

• Fired: Default Alive 1.0 is 10.3 days in the App Store review queue (submitted Oct 2). This is task 2.
• Fired: two dated gates on Pet Med Reminder. The store listing hold ended Oct 10 and the Instagram ad test ends today at noon. Both reads are task 3.
• Pet Med Reminder 1.12.1: waiting for review, 6.3 days (submitted Oct 6). Inside the budget. It releases automatically on approval.
• Uploaded and not submitted for more than seven days: clear.
• No app changed state since the last standup (Oct 10).

Since yesterday

• The Cycle 2 close ran. One key result of ten is met: Default Alive players come back the next day on their own.
• The Pet Med Reminder landing page change of Oct 8 did not raise the tap rate. Before it, 1 of 123 phone visits tapped through to the App Store. After it, 2 of 142. By the rule set before the change, the page was not the leak; the audience was.
• No change in any app’s review state.

Today

1. Write the Cycle 3 OKRs (DAVE-218) → no live cycle. The Cycle 2 close on Monday is the main input: the cycle doc says the Default Alive results are its baseline for Cycle 3, not a verdict.
2. Read the Pet Med Reminder impressions for the 28-day window Sept 12 to Oct 10 in App Store Connect, when today’s data lands, and record it before the listing changes (PMR-37) → run the business. The window closes today, so the store listing hold ends. The subtitle change waits for this read.

With no live cycle, no other task is listed or filed.

Canaries

• Fired: the Pet Med Reminder store listing hold ends today (PMR-37, a dated gate in the cycle doc). It is task 2.
• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 8.3 days (submitted Oct 2). Inside the 5 to 9 day budget. The canary fires tomorrow, Oct 11. On approval it is not released (manual release, held for v2).
• Pet Med Reminder 1.12.1 (27): WAITING_FOR_REVIEW, 4.3 days (submitted Oct 6). It releases automatically on approval.
• App Store agreements: active to Oct 23, 13 days away (read Sept 26, not re-read).
• Uploaded and not submitted for more than seven days: clear. Default Alive builds 17 and 18 went to TestFlight on Oct 3, for testers, not for review.
• No app changed state since yesterday.

Since yesterday

• The Cycle 2 window ended. No cycle is live until Cycle 3 is adopted.
• No change in any app’s review state.

Today

1. Send the five follow-ups to the foster coordinators. They are drafted and unsent. Then stop (PMR-59) → P1 and P2 (first doses and habitual crossings). The follow-up date was Oct 8. No coordinator has replied in 8 days; the bar is 2 of 5 within 14 days.
2. Compare the landing page before and after the Oct 8 rewrite: App Store taps per phone visitor, before and after 15:43 EDT on Oct 8 (PMR-57) → P1 (external installs with a first dose inside 48 hours). Before the rewrite, about 2 of 124 phone visitors tapped through. The ads run to Oct 12, so this is the first read of the change.
3. Record the Show Notes store page figures (page views, first downloads, conversion since Sept 28) in SN-58 (SN-58) → S1 (strangers who read a summary of a show they follow). The close needs them to rule whether S1 had a population.

Default Alive has no task today that moves a key result. G3 is settled at 3 of 5. G2 passes as read on Oct 5. G1 reads 15 against 25 and needs 10 new outside players to start the clock today. Show Notes and Paroletta start nothing new, as the Oct 1 amendment says.

Canaries

Clear.

• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 7.3 days (submitted Oct 2). Inside the 5 to 9 day budget. The canary fires on Oct 11. On approval it is not released (manual release, held for v2).
• Pet Med Reminder 1.12.1 (27): WAITING_FOR_REVIEW, 3.3 days (submitted Oct 6). It releases automatically on approval. The listing hold ends tomorrow, Oct 10 (PMR-37).
• App Store agreements: active to Oct 23, 14 days away (read Sept 26, not re-read).
• Uploaded and not submitted for more than seven days: clear. Default Alive builds 17 and 18 went to TestFlight on Oct 3, for testers, not for review.

Since yesterday

• The tap test settled that a tap on the App Store button inside Instagram’s in-app browser is counted. The near-zero tap rate is real: the landing page, not the ad, is where the funnel breaks. The page was rewritten the same day (PMR-57).
• G3 (installs a person sent) is settled at 3 against a target of 5, from the founder’s check of his own sends (SIM-12).
• The five foster follow-ups were drafted and not sent (PMR-59).

Today

1. Mark the three open “a person sent me the game” answers as sent by the founder or not (SIM-12) → G3 (new installs that a person sent). The list is in SIM-12 from Oct 7. One of the five answers is already out (a tester the founder invited), and two probably are one phone. G3 reads 3 at most against a target of 5. The close on Oct 12 needs the marks.
2. Settle whether the Pet Med Reminder ads lead to installs: tap the App Store button inside Instagram’s in-app browser and check that PostHog counts it, then read downloads per ad (ct=ig-a, ct=ig-b) in App Analytics (PMR-57) → P1 (external installs with a first dose inside 48 hours). Carried from Oct 7. The ads run to Oct 12, so this read is what the close uses.
3. Send the one follow-up to each foster coordinator who has not replied, log the replies, then stop (PMR-59) → P1 and P2 (first doses and habitual crossings). Oct 8 is the date set for the follow-up.

Show Notes and Paroletta start nothing new, as the Oct 1 amendment says.

Canaries

Clear.

• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 6.3 days (submitted Oct 2). Inside the 5 to 9 day budget; the canary fires on Oct 11. On approval it is not released (manual release, held for v2).
• Pet Med Reminder 1.12.1 (27): WAITING_FOR_REVIEW, 2.3 days (submitted Oct 6). It releases automatically on approval. Name, subtitle and keywords are unchanged, so the listing hold to Oct 10 (PMR-37) is intact.
• App Store agreements: active to Oct 23, 15 days away (read Sept 26, not re-read).
• Uploaded and not submitted for more than seven days: clear. Default Alive builds 17 and 18 went to TestFlight on Oct 3.

Since yesterday

• Nothing changed the list. No Enginery product work was recorded on Oct 7.
• The App Store queue moved one day: both apps in review are still waiting, inside their budgets.

Today

1. Prepare the G3 check for the close: list the five “a person sent me the game” answers by install and date (SIM-12) → G3 (new installs a person sent). Done at 07:13 and written to SIM-12. Of the five answers, one is a tester the founder invited himself, and two look like one phone installed twice. G3 reads at most 3 against a target of 5, before the founder marks the rest.
2. Settle whether the Pet Med Reminder ads lead to installs: check that the App Store tap is counted inside Instagram’s in-app browser, and read downloads per ad from App Analytics (PMR-57) → P1 (external installs with a first dose inside 48 hours). The day 1 read found about 42 phone visitors from the ads and 0 or 1 App Store taps. Either the landing page loses them or the tap is not measured. App Analytics data for the first two days should now be in. An install today still has its 48 hours inside the window.
3. Read the Show Notes store page by hand: page views and first downloads since 1.5.0 (SN-58) → S1 (strangers who read a summary of a followed show). Done by the founder on Oct 7.

Show Notes and Paroletta start nothing new, as the Oct 1 amendment says.

Canaries

Clear.

• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 5.3 days (submitted Oct 2). Inside the 5 to 9 day budget. When Apple approves it, it is not released (manual release, held for v2).
• Pet Med Reminder 1.12.1 (27): submitted Oct 6, WAITING_FOR_REVIEW, 1.3 days. It adds the Meta SDK for app promotion ads (PMR-60). It releases automatically on approval. Name, subtitle and keywords are unchanged, so the store listing hold to Oct 10 (PMR-37) is intact.
• App Store agreements: active to Oct 23, 16 days away (read Sept 26, not re-read).
• Uploaded and not submitted for more than seven days: clear. Default Alive builds 17 and 18 went to TestFlight on Oct 3.

Since yesterday

• The Pet Med Reminder ads delivered on day 1: 35 link clicks for $13.35, about the same for both ads. Almost no visitor tapped through to the App Store, or the tap was not counted.
• A redirect straight to the App Store was tried and refused by Meta: its Traffic objective does not accept App Store links. The current test keeps its landing pages to its Oct 12 end.
• Pet Med Reminder 1.12.1 went to App Review on Oct 6 with Meta’s SDK, so a second round can use app promotion ads that open the store page directly and count installs. Review takes 5 to 9 days, so this serves the next cycle, not P1.
• The ad account now has a $200 spending limit. It was not set before Oct 6.

Today

1. Prepare the G3 check for the close: list the five “a person sent me the game” answers by install and date (SIM-12) → G3 (new installs a person sent). The founder ruled on Oct 5 that a link he sent himself does not count. The list lets him mark each answer against his own sends before Oct 12, so the G3 grade rests on checked answers. No garage change and no new posts go out for G1 in the last days, by the founder’s Oct 5 call.
2. Confirm the Pet Med Reminder round test is delivering: both ads out of Meta review, first impressions and clicks per ad (PMR-57) → P1 (external installs with a first dose inside 48 hours). The ads were published on Oct 5 for 7 days. An install needs to land before Oct 7 to have its 48 hours inside the window, so a stalled ad review costs P1 its last chance.
3. Read the Show Notes store page by hand: page views and first downloads since 1.5.0 (SN-58) → S1 (strangers who read a summary of a followed show). Carried from Oct 5. It says whether the break is the page or the traffic to it, which is the S1 finding for the close.

Show Notes and Paroletta start nothing new, as the Oct 1 amendment says.

Canaries

Clear.

• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 4.3 days (submitted Oct 2). Inside the 5 to 9 day budget. When Apple approves it, it is not released (manual release, held for v2).
• App Store agreements: active to Oct 23, 17 days away (read Sept 26, not re-read).
• Uploaded and not submitted for more than seven days: clear. Default Alive builds 17 and 18 went to TestFlight on Oct 3.

Since yesterday

• Default Alive 1.0 will not be released when Apple approves it. Default Alive v2 (a web version and a rewrite, with the server running the game) is the next work on the game. The garage change planned for G1 (SIM-50) is on hold for v2.
• The Default Alive end calls with five players are dropped. The learning question (do people play to learn how startups work, or for the tycoon fun?) is answered at the close from the written tester feedback.
• Two possible internal installs are excluded from G1 and G2. G1 reads 15 and G2 reads 11 of 13, which still passes.
• The Pet Med Reminder round test went live on Instagram on Oct 5: two ads, $10 a day each, for 7 days, iPhone only.

Today

1. Show the garage’s options one at a time in the first minutes (SIM-50) → G1 (outside players who start the clock at the seed close). G1 reads 17 against at least 25. On build 18, which carries the build 17 tips, 3 of 5 new players still left the garage within 4 minutes. Four testers say the game shows too much at once. This is a design call; the standup recommends it, and Dave can overrule.
2. Hold the end calls with five players (SIM-12) → G2 (players back the next day on their own) and the learning question. Carried for the third day. G2 already reads 13. The calls answer the learning question (learn how startups work, or tycoon fun?) and the second half of the playtest criteria. Hold them before the Oct 12 close.
3. Launch the Pet Med Reminder round test now that 1.12.0 is live (PMR-57) → run the business, and P1 (external installs that log a first dose inside 48 hours). 1.12.0 went live at 05:32 EDT today. The ads are built and dated Oct 6. P1 needs installs before Oct 7 to have their 48 hours inside the window.

Show Notes and Paroletta start nothing new, as the Oct 1 amendment says.

Canaries

Fired: Pet Med Reminder 1.12.0 cleared review. It went live on 2026-10-05 at 05:32 EDT, after 7.6 days in the queue. It unblocks the round test (PMR-57), the coordinator follow-up on Oct 8 (PMR-59) and the P3 read: rage clicks on the fixed Care Item Editor can be counted from today. Filed as task 3.

• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 3.3 days (submitted Oct 2). Inside the 5 to 9 day budget.
• App Store agreements: active to Oct 23, 18 days away (read Sept 26, not re-read).
• Uploaded and not submitted for more than seven days: clear.

Key results

Read 2026-10-05 between 06:05 and 06:30 EDT, product analytics, external tier. Day 10 of 14.

Since yesterday

• Pet Med Reminder 1.12.0 is live. It makes the round of doses the unit: one notification and one tap per round, a catch-up screen, and reminders that do not stop after 48 hours. It also carries the Care Item Editor fix.
• G3 reached its target of 5: five new players say another person sent them the game.
• The garage read on build 18 (yesterday’s task 1) is done in this run. The one-time tips did not hold the opening: 3 of 5 new players left within 4 minutes. That gives today’s task 1.

Today

1. Read the garage funnel for new players on build 17 and later, against 14 of 27 before the fix (SIM-48) → G1 (outside players who start the clock at the seed close). Build 17 put a one-time tip over the first “what to build”, price check and launch cards, the step where most players stopped. It reached testers on Oct 3. The read shows whether the step still loses the most players, and the hint_closed reasons. One day of installs is a small sample. Read it anyway: the cycle says to read the funnel daily and fix the step that loses the most players, and 5 days are left. The Oct 5 review reads it again.
2. Plan the end calls with five players (SIM-12) → G2 (players back the next day on their own) and the learning question. Carried from yesterday. Pick five players who started the clock and whose next day has passed, so a call cannot count as a prompted return. Hold the calls before the Oct 12 close.

Pet Med Reminder has no task: 1.12.0 is in the review queue. Show Notes and Paroletta start nothing new this cycle. Their next read is the Oct 5 review.

Canaries

Clear.

• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 2.3 days (submitted Oct 2). The App Store release is not cycle work.
• Pet Med Reminder 1.12.0: WAITING_FOR_REVIEW, 7.3 days (submitted Sept 27). Inside the 5 to 9 day budget. Past 9 days on Oct 6.
• App Store agreements: active to Oct 23, 19 days away.
• Uploaded and not submitted for more than seven days: clear. Default Alive builds 17 and 18 are TestFlight builds, approved for beta.

Since yesterday

• Build 17 (SIM-48) fixed the garage’s opening with a one-time tip on the first three cards and a clearer garage line. It is on TestFlight. Yesterday’s task 1 shipped.
• Build 18 (SIM-49) made the Series A bar wait for the update instead of ending the run, and added a main menu with one save per company. It is on TestFlight. The first players reach the bar about Oct 8 to 10.
• New builds now upload from the command line, with no manual step. Each build’s What’s new entry publishes to enginery.co.
• The game’s next architecture is decided: a server runs the game, and the web version comes after the game reaches the IPO. This is Cycle 3 and later work, and no cycle key result changes.

Today

1. Fix the garage’s opening, where players stop at the first “what to build” card (SIM-48) → G1 (outside players who start the clock at the seed close). The funnel read (SIM-47) found that 13 outside players never started the clock, 10 of them left within 11 minutes, and the largest group (5) left with the “what to build” card on screen. None of the 13 went broke. Players whose garage run ended started again and reached the clock. So the step to fix is the first decision, not the late raise. The fix ships in the next TestFlight build.
2. Plan the end calls with five players (SIM-12) → G2 (players back the next day on their own) and the learning question. A call is a contact, and a return after a contact does not count for G2. So each call goes after that player’s next day has passed. Pick the five from players who started the clock, and hold the calls before the Oct 12 close.

Pet Med Reminder has no task: 1.12.0 is in the review queue. Show Notes and Paroletta start nothing new this cycle. Their next read is the Oct 5 review.

Canaries

Clear.

• Default Alive 1.0 (16): WAITING_FOR_REVIEW, 1.3 days (submitted Oct 2). The App Store release is not cycle work.
• Pet Med Reminder 1.12.0: WAITING_FOR_REVIEW, 6.3 days (submitted Sept 27). Inside the 5 to 9 day budget. Past 9 days on Oct 6.
• App Store agreements: active to Oct 23, 20 days away.
• Uploaded and not submitted for more than seven days: clear.

Since yesterday

• Default Alive 1.0 (16) went to App Store review on Oct 2, with manual release.
• TestFlight 0.1 (15), with the launch fix, reached the Playtest group. A player on build 15 started the clock the same evening.
• The garage funnel read is done (SIM-47). It disproved the morning’s guess that a late raise loses players. The first “what to build” card loses the most. That is why task 1 changed.
• Readings from the same pass, not grades: G1 at 14 outside clock starts against a target of 25. G2 at 8 of 10 back the next day: the rate passes, the count does not yet.

Today

1. Upload Default Alive TestFlight 0.1 (15) and give it to the Playtest group (SIM-43) → G1 (outside players who start the clock at the seed close). Build 15 makes “Launch now” always launch. On builds 7 and 11, a company whose first customer had left could not launch, so the player never reached the seed close. The build is archived and open in the Organizer. App Store Connect has builds up to 14, so 15 is not uploaded yet. Upload 1.0 (16) in the same pass for the store version.
2. Read the garage funnel: the last event before each outside player left (SIM-47) → G1. On Oct 1, 0 of 7 players who arrived after 10:20 EDT started the clock. The read names the step that loses the most players, by build. It is a reading, not a grade.

Pet Med Reminder has no task: 1.12.0 is in the review queue, and its launch day waits on Apple. Show Notes and Paroletta start nothing new this cycle. Their next read is the Oct 5 review.

Canaries

Clear.

• Pet Med Reminder 1.12.0: WAITING_FOR_REVIEW, 5.3 days (submitted Sept 27). Inside the 5 to 9 day budget.
• Default Alive 1.0: PREPARE_FOR_SUBMISSION, build 1.0 (8) attached Oct 1. The App Store submission is not cycle work.
• App Store agreements: active to Oct 23, 21 days away.
• Uploaded and not submitted for more than seven days: clear.

Since yesterday

• Sim Startup’s public name is Default Alive.
• The Cycle 2 weekly review is answered (in the Sept 28 standup), with Sim Startup in it. Its first change for Sim Startup is to ship the launch-card fix to testers. That change became today’s task 1.
• Builds 0.1 (11) and 0.1 (13) went to testers on Oct 1 with the first launch-card fix (SIM-43) and the garage’s full mail list (SIM-46). A tester who updated an old save still could not launch. The cause was found that night and fixed in build 15.
• Two testers say the garage is hard to follow, and a third ran short of cash. That is why the funnel read is task 2.
• The three Show Notes founder-welcome notes went out (SN-51, closed).

Today

1. Send the three Show Notes founder-welcome notes to the earlier sign-ups who never added a show (SN-51) → S1 (strangers who read a summary of a show they follow) and the Show Notes learning question. The notes are written and wait as drafts. Each asks one question: what were you hoping it would do? Carried for the second day.
2. Answer the weekly review questions in the Sept 28 standup (15 minutes) → every key result. Carried for the third day. The Oct 5 review, 4 days away, replaces them if they are still open then.

Pet Med Reminder has no task. Its cycle work is 1.12.0, which is in the review queue (4.3 days). Paroletta has no task. The next read is the Oct 5 review.

Canaries

Clear.

• Pet Med Reminder 1.12.0: WAITING_FOR_REVIEW, 4.3 days (submitted Sept 27). Inside the 5 to 9 day budget.
• App Store agreements: active to Oct 23, 22 days away.
• Uploaded and not submitted: clear for the three products. Every uploaded build is live, in review or replaced.

Since yesterday

• Nothing changed the list. No release moved. Yesterday’s work was outside the three products.

Today

1. Send the three Show Notes founder-welcome notes to the earlier sign-ups who never added a show (SN-51) → S1 (strangers who read a summary of a show they follow) and the Show Notes learning question. The notes are written and wait as drafts. Each asks one question: what were you hoping it would do? The answers go into the loop page’s statement of why people want the product.
2. Answer the weekly review questions in the Sept 28 standup (15 minutes) → every key result. Carried for the second time. If they are still open on Oct 5, the Oct 5 review replaces them.

Pet Med Reminder has no task. Its cycle work is 1.12.0, which is in the review queue (3.3 days). Paroletta has no task. The next read is the Oct 5 review.

Canaries

Clear.

• Pet Med Reminder 1.12.0: WAITING_FOR_REVIEW, 3.3 days (submitted Sept 27). Inside the 5 to 9 day budget.
• App Store agreements: active to Oct 23, 23 days away.
• Uploaded and not submitted: clear. Every uploaded build is live, in review or replaced.

Since yesterday

• Nothing changed the list. No release moved, and yesterday’s work was outside the three products.

Today

1. Start the founder welcome for new external Show Notes accounts (SN-51) → S1 (strangers who read a summary of a show they follow) and the Show Notes learning question. Show Notes 1.5.0 went live yesterday, so strangers now read a summary before they give an email. Each new account from now on gets one personal note with one question: what were you hoping it would do? The cycle’s learning question (does the loss of skipped episodes bother anyone?) counts an answer to this note as evidence. SN-51 is in the backlog and was never started.
2. Answer the weekly review questions in the Sept 28 standup (15 minutes) → every key result. Carried from yesterday. The ## Review section is still empty, so no change is chosen for the week.

Pet Med Reminder has no task. Its cycle work is 1.12.0, which is in the review queue (2.3 days). The outreach and the message test wait for it to go live. Paroletta has no task. Its first-guess fix is live, and the next read is the Oct 5 review.

Canaries

One fired.

• Show Notes 1.5.0 is live (READY_FOR_DISTRIBUTION, released Sept 28 at 09:31 EDT). It carries the first win before sign-in, the caught-up screen and the new store page. The S1 and S2 counts start at that time, and the S1 event is proven. It unblocks the founder welcome, task 1.
• Pet Med Reminder 1.12.0: WAITING_FOR_REVIEW, 2.3 days. Inside the 5 to 9 day budget.
• App Store agreements: active to Oct 23, 24 days away.
• Uploaded and not submitted: clear.

Since yesterday

• Show Notes 1.5.0 went live on Sept 28 at 09:31 EDT, about 23.5 hours after submission. The time comes from the public App Store lookup. S1’s first 30 and S2’s distinct days count from then.
• The event that counts S1 (followed_summary_read) fired for the first time on Sept 28 at 09:56 EDT, on 1.5.0, from an internal test read. The event works. Internal reads are excluded from the grade.
• The App Store download reports refuse access for a named cause: the role of the key that asks for them. The fix is one new key.

Today

1. Weekly review (15 minutes): answer the three questions per product at the end of this doc → every key result. The answers go under ## Review here, and each chosen change is filed with an owner and a review date.
2. Read one followed-show summary in Show Notes 1.4.0 on your phone (SN-57) → S1 (strangers who read a summary of a show they follow). The event that counts S1 has not fired once since 1.4.0 went live. This is probably because nobody has read a summary on that build, but nothing proves the event works. It must work before 1.5.0 opens the S1 window.

Pet Med Reminder has no task. Its cycle work is 1.12.0, which is in the review queue. The outreach and the ad test wait for 1.12.0 to go live, by an earlier decision. Paroletta has no task. Its first-guess fix is live on web and iOS, and the new arrivals show it: no first-visit how-to screen has fired since the fix.

Canaries

Clear.

• Pet Med Reminder 1.12.0: WAITING_FOR_REVIEW, 1.3 days. Inside the 5 to 9 day budget.
• Show Notes 1.5.0: WAITING_FOR_REVIEW, 1.3 days. Inside the budget.
• Nothing moved to live since yesterday.
• App Store agreements: active to Oct 23, 25 days away (read Sept 26).
• Uploaded and not submitted: clear. Every uploaded build is live, in review or replaced by a later one.

Key results

Since yesterday

• Pet Med Reminder 1.11.0 was withdrawn from review. Its Care Item Editor had no time picker, so no dose time could be set. 1.12.0 replaced it and is in review. 1.12.0 makes the round of doses the unit: one notification and one tap per round, a catch-up screen, and reminders that do not stop after 48 hours. It also carries the editor fix, so P3’s graded population starts with 1.12.0.
• Two custom App Store product pages for Pet Med Reminder are in review with 1.12.0, for a message test after launch.
• Show Notes 1.5.0 was submitted Sept 27. It carries the first win before sign-in, the caught-up screen and the new store page.
• The Paroletta first-guess fix is reaching players: since the release, no web or iOS arrival was shown the how-to screen on first visit.

Today

1. Upload Show Notes 1.5.0 build 18 and submit it (SN-56) → S1. Show Notes 1.4.0 cleared review last night, so the review slot is free and nothing has to be pulled. 1.5.0 carries the first-win flag on, the caught-up screen and the new store page. The cycle doc says the store page is the first break for S1, and S1’s window opens the day 1.5.0 goes live. The build is archived; the listing packet and the runbook are ready.
2. Read I1 on the fixed build at tomorrow’s baseline (ITALIAN-147) → I1. Paroletta 1.7 is live, so the first-guess fix is out on web and iOS. The Monday read takes the 28-day figure as the cycle doc says, and also splits arrivals at the two release times, so the baseline and the fix are not mixed.

Pet Med Reminder has no task today. Its cycle work shipped in 1.11.0, which has been in the review queue for 1.3 days.

Canaries

• Show Notes 1.4.0 cleared review. Live 2026-09-26 at 19:09 EDT, about five hours after submission. It unblocks the 1.5.0 submission (task 1). The S1 event followed_summary_read ships in 1.4.0 and can fire from now on.
• Paroletta 1.7 cleared review. Live 2026-09-26 at 19:05 EDT, about five hours after submission. It puts the iOS half of the first-guess fix in front of arrivals (task 2). The web half has been live since 13:22 EDT Sept 26.
• Pet Med Reminder 1.11.0: WAITING_FOR_REVIEW, 1.3 days. Inside the 5 to 9 day budget.
• App Store agreements: clear. Read Sept 26: both active to Oct 23, 26 days away. The canary now watches Oct 23.
• Uploaded and not submitted: clear. Pet Med Reminder build 19 and Paroletta build 15 are superseded by submitted builds 24 and 16.

Since yesterday

• Three releases were submitted on Sept 26: Pet Med Reminder 1.11.0, Show Notes 1.4.0 and Paroletta 1.7. Two cleared review the same evening. Pet Med Reminder is still in the queue.
• Pet Med Reminder 1.11.0 carries all of the cycle’s Pet Med Reminder work: the Care Item Editor fix, onboarding that ends on the Today screen, the run of dosing days, and the rating prompt on the habitual crossing. P3’s window starts when 1.11.0 goes live.
• The Apple agreements were signed Sept 26 and run to Oct 23.
• Show Notes 1.5.0 is built and archived. The server side of the caught-up screen is deployed.
• SN-41 was closed early at 3 of 29. The Show Notes store page is free to change, and 1.5.0 changes it.

Today

1. Submit Pet Med Reminder 1.10.0 (PMR-44) → P1. The cycle doc lists this first under the Pet Med Reminder work: “submit 1.10.0 today”. Build 19 has been uploaded and unsubmitted for 19 days, so this is also a fired canary. The PMR-37 hold binds the submission: name, subtitle and keywords stay as they are until Oct 10. Promo text, description, What’s New and screenshot captions are free. Two App Information age-rating questions (socialMedia, socialMediaAgeRestricted) are still blank, and the submission will probably ask. Do task 2 first, because a lapsed agreement can block the submit.
2. Read the App Store agreements page and record the reading (DAVE-171) → run the business. The recorded expiry of both agreements was Sept 25, yesterday. Nobody has read the page since Sept 8, and every listing depends on it. The cycle doc lists this under “Instruments owed”.
3. Re-verdict the Care Item Editor on 1.9.0 (PMR-46, with PMR-19 and PMR-41) → P3. This is the second Pet Med Reminder work item, and it also serves P1: the editor is the step in front of the first dose. P3’s target is zero rage clicks after the fix is live, so the fix has to ship early in the cycle to have a window.

Canaries

• Agreements past expiry, unread (DAVE-171). Recorded expiry Sept 25. The last reading (Sept 8) showed Paid Apps “Pending User Info”. The API version states say nothing about agreements. Fires until DAVE-171 records a reading.
• Build uploaded, not submitted, past seven days. Pet Med Reminder 1.10.0 build 19, uploaded Sept 7 per the PMR-44 activity log, 19 days. release_check.py confirms build 19 is on Apple’s servers and 1.9.0 is live.
• No state changed since the dry run. Nothing is in WAITING_FOR_REVIEW.
• Not readable: release_check.py also shows Paroletta build 15 uploaded and not attached to a live version. No WCP item records its upload date, so the seven-day rule cannot be checked.

Since yesterday

• Cycle 2 adopted (Sept 26, 10:02). It has three objectives, one per product, and seven key results: P1 to P3, S1 and S2, I1 and I2. WCP Cloud, Lettura and Ascolto are personal projects and carry no objective.
• DAVE-191 (write the Cycle 2 OKRs) is done.
• Show Notes: Dave decided at adoption that the first win comes before the email. A stranger adds one show and reads a summary without an account. SN-46’s email bar is superseded by S1.
• Verdicts written: Vocabolario iterates as the number, Categorie’s growth thesis is killed, and the WCP Cloud registry closed early at 0 of 5 (WCPC-193). The evangelism hold is lifted.
• PMR-49 filed: the rating prompt on the habitual-bar crossing, after the editor fix.
• The standup replaced the pulse (DAVE-201). Board freeze and pointer re-pointing are due by Sept 29.

We ported Pet Med Reminder to Android this week, or rather astra did. By Monday I had a pitch for it, a big one, six weeks of appetite with maybe two of actual building and the other four sitting with Google, since the Play account is a personal one and you need twelve testers opted in for fourteen days before you can even ask for production. Monday evening I handed the pitch to astra and a couple of hours later there was a working app, and by the next morning a signed bundle, screenshots, the listing, all of it. So the two weeks of building were a day, and the four weeks of Google are still four weeks. What I keep thinking about is that astra disagreed with the pitch in six places and was right every time. I had the iOS test count wrong, I said we didn’t need an in-app privacy policy (Play requires one), I promised a max delay on alarms that nobody can promise, I claimed we send no identifiers when the analytics SDK sends one, and so on. Meanwhile I did the three Play Console declarations myself and got two of them wrong. Said the app collects no data, which it does. Set the audience to 13 and up, which pulls the whole thing under the Families policy. Both fixed now, but the thing is I had 119 tests for the app and nothing for the forms except me looking at them.

I am picking Italian back up alongside the rest of Enginery. The evening chapter came by email, and an email sender has no way to know whether I read yesterday’s, so it sent another one. Two weeks into having a newborn I had fourteen unread chapters and a review schedule that had ticked over thirteen times without me listening once. On a walk there was no way to tell which one was next.

Filotto would cover this eventually, and Filotto is parked, because it is not ready for anyone who is not me. I did not want to wait for it to keep listening. So the chapters go to a small backend now, and a one-screen app shows the one at the head of the queue and nothing behind it. Tapping Fatto is the only thing that moves the queue along.

While I was in there I wanted the car. I built the audio for the dog walk, and the walk is the slot that keeps losing to a newborn. The twenty minutes of driving happen either way. From the same queue I can put today’s chapter on the dashboard, play only, nothing to tap while the car is moving. Due piccioni, uno bracco.

CarPlay turned out not to be mine to schedule. You cannot switch it on in Xcode and you cannot switch it on in App Store Connect. You send Apple a form asking for the CarPlay audio entitlement and then wait days to weeks. Makes sense, and it is still annoying. I am building against the simulator and I sent the form the same day rather than waiting until the code was done.

I wrote all of this down before building any of it. The pitch, the architecture proposal, the CarPlay follow-on and the six screens are up here as written.

My issues list has carried the same row since Aug 31: publish something into an ecosystem registry. Today was the fifth morning it read zero attempts, so I stopped trying to do it and asked instead why it wasn’t getting done. It wasn’t avoidance. Publishing a listing looks like a twenty minute job and it is really four, because whatever you’re listing has to be pulled out of wherever it lives, licensed, and written up well enough for a stranger before you ever get to the twenty minutes. Both of the things I had lined up needed a week of that first, which is why I kept not starting. Then I remembered that WCP Cloud is already hosted, so there is nothing to pull out and nothing to document, because what you publish is an address and not code. The thing with the deadest repo turned out to be the easiest one to list. It’s live now. Every user WCP has ever had came from me talking about it in Slack, so this is the first time somebody can find it without already looking at me.

After waking up and finding Opus hallucinating, the product pulse operating on old stale data and context, I decided to streamline the system.

I’m already happy with how our OKRs tracker is working (mostly) but I’ve learned that I need to hand-rewrite the OKRs every time. They’re always too complex, too wordy, which ironically is the same problem I’ve had in the past with human teams. C’est la vie.

It hit me today that I really just need to use my time-tested system, codified here:

• OKRs at the company level
• Per-product OKRs that align/nest within them
• KPIs derived from these key results, but also to just ‘Run the business’
• One single issues list that we keep prioritized that

Every morning the pulse skill will read all of this, including fetching the latest KPIs, and we’ll look at the issues list and decide what to do today.

And, a lesson that I repeatedly learn: the writing of the OKRs is the single highest-leverage activity as a leader because it forces me to clarify my thinking and put it into writing. Thinking is bottom-up, but communicating is top-down.

At the coffee shop today I was struck by the lack of OKRs. Mentor of mine used to say that OKRs are the solution to every management problem ever, which I’ve come to agree with. However, many people tend to have an allergic reaction to OKRs. I’ve found this is almost always because the goals themselves are poor, the KRs don’t cleanly map, and/or the objectives don’t make sense or don’t nest underneath any product. And, sometimes it’s because the OKRs were themselves a mechanism to micromanage rather than provide organizational clarity.

Note to self that I should publish my own OKR patterns. Which reminds me - perhaps one of the reasons claude-generated OKRs, even living in the context of my assistant ontology vault, still suffer: they don’t have the context of past examples drawn directly from my own career and experience.

Adding that, I found I was able to use many Bloc, LeakAware, OrangeQC, Sierra Rails, QueryClips, and The Agency of Learning examples that really helped claude tighten them up into a decent template. I’ll hopefully share that pattern at some point here as well.

The Italian portfolio is not building three apps. It is building one known-word list per person, with evidence tiers and provenance: which words you can produce, which you only recognize, and which surface proved it. Every product accrues to that list and spends from it.

Today, though, the portfolio ships one list per learner per surface, keyed to identifiers that never join. One person on three devices counts as three learners, Filotto builds its own list and never reads Paroletta’s, and the bridge between them was designed but not built. The number the portfolio grades itself on was summing anonymous fragments and calling that the asset.

So the list gets one home and one anchor. If the hypothesis under test is that the list is the valuable object, an architecture where the valuable object can be orphaned by a cleared cookie is not testing it.

Ten shipped things and three old names between them: Zen Space Labs (a defunct company, still the seller line on the App Store), Sierra Rails, and Paola Software. As of today Enginery owns all of it. It holds the iOS and Android apps, the web properties, the Work Context Protocol, and the client work.

The name is not decoration. Latin ingenium meant the cleverness begotten inside you, and it descends along two limbs: Italian kept ingegno, which still means both ingenuity of mind and a contrivance or machine, while English turned it into engine, kept the machine, and let the mind sense die out around 1600. Enginery is the English word left standing on that seam.

Products keep their own names. Enginery is the publisher, not a prefix.

Android was a generation behind: it matched the game as of Aug 12, not Aug 20. It was missing the farm card, both new modes, the per-artifact share codes, and the shared streak key. All of that was built and shipped in a single day, and the signed release went into Google Play closed-testing review under the Enginery developer account, with the store listing complete in two languages.

App signing enrolled by uploading the existing key rather than letting Google mint a new one, which would have orphaned every sideload install. And Play requires twelve testers to stay opted in for fourteen continuous days before an app can be promoted to production, so the real gate is recruiting, not code. Production lands around mid-September.

The plan said iOS would stay on TestFlight until Categorie and Vocabolario were finished, with one App Store submission in mid-September carrying everything. Instead the build went to review early with both modes aboard but switched off, which meant nothing about the shipped app changed except that it now contains two features nobody can see yet. Apple approved it today.

The App Store cohort clock starts now rather than in three weeks, so the activation and week-two-return numbers get an extra three weeks of real store traffic before the first grade. And the input-flow fix is in front of store users now instead of in September. It was one of the two known causes of people touching the game and never producing a word.

Filotto keeps several tiers of evidence about a word. Some are strong: you answered it, or you marked it known on the calibration grid. One is a statistical inference from a recognition checklist, useful for placing a learner and not a record of any particular word. The chapter generator had to say, in code, which tier the shipping gate measures against.

Both candidates ran against a volume already read and judged. The generous boundary scored it at 99.4 percent, comfortably easy. The narrow one scored it at 95.5 percent, and its reader had said that chapter read harder than usual. The generous gate would have called a chapter too easy on the evening its reader called it hard.

So coverage counts proved words only. A statistical inference about two thousand words is not two thousand words of comprehension.

The nightly reader is live: a chapter generated for one reader, verified against their known words, and mailed out. Nothing in that pipeline says whether anybody reads it.

A tracking pixel measures image loading, not reading. Apple Mail pre-fetches images for much of the plausible audience and other clients block them outright, so the number would be wrong in both directions at once. Instead every send mints a token, the email carries the title and a two-line lead-in, and the link goes to that reader’s own page. Fetching the page is the open, recorded server-side.

The chapter text is not in the email, so reading it takes a click. If opens come in under 50 percent, the first hypothesis is that this decision caused it, and the variant to test is putting the text back.