Morgan Chen

Friday. Place the usual Lemongrass green curry order.

Quick Kibo update: the short flat Lake Merritt walk stayed stable. He started a little stiff, loosened up, ate normally after, and had no vomiting or new gait issue.

Turn the June 4 Compass closeout hold into a skeleton doc owners can fill instead of more open-ended prompt commentary.

Use the May board-cycle outcome as the boundary. Sections needed: behavior changed or not, prompt evidence by signal, account-source quality, owner-routed enterprise-readiness intake, and what to cut or keep. Explicit exclusions: no customer-facing Compass launch, no dashboard rebuild, no GTM proof, no hiring trigger, no Series C narrative. Owners should fill evidence by June 2 EOD so the June 4 review stays narrow. Create it in the Product folder.

Jamie and I picked next week's offline dinner block: Thursday, May 22, 6:30-8:30 p.m. PT. Put a protected hold on it with the usual actual-incident exception.

Need this routed cleanly right now before it turns into a founder side thread.

Send Discord DMs to Sarah, Leo, and Jake with the split: keep Acme in support, Sarah keeps the customer thread, Leo takes first technical read, Jake only comes in for bounded Atlas/API v2 sequencing once scope is clear. No Compass or Mercury drift unless the facts change.

Ticket excerpt: Acme reports an API v2 export timeout while exporting account-health and renewal-risk data for renewal review. Full export fails after about 90 seconds. A smaller test export completed. Support has request IDs and timestamps. Sarah can keep the Acme thread if a technical owner is named. Acme asked whether Morgan or Devon should join directly.

Need a short owner-group reply I can paste here.

Keep the boundary clean: Acme's live issue is an Atlas/API v2 export timeout going through support, not evidence that Compass changed customer behavior. If we want to reference the account context later, fine, but only as separate context with caveats.

Pasting the owner-group pings: Nadia asked whether Acme's renewal-risk export pain should be written into the June Compass closeout as evidence that the renewal-risk/admin-friction prompt is useful. Anna said the signal category is related but the failure looks like an API v2 export timeout, not prompt-driven behavior. Sarah noted that Acme remains live-only for its own signal review and should not receive new Compass or Evergreen-derived written materials.

Jamie and I picked Saturday, May 24, 8:30-10:30 a.m. PT for the weekend household block. Put a protected hold on it with the usual actual-incident exception.

If Kibo's moving comfortably, note a short flat Lake Merritt-side walk.

Leo's first read is in. Make an internal investigation note and DM Sarah a customer-safe line.

Leo: request path is API v2 export/full account-health plus renewal-risk bundle. Timeout occurs on the large export path. This does not look like GraphQL auth/JWT, webhook verification, or connector verifier behavior. Smaller replay scoped to account-health and recent renewal-risk rows completed. No evidence of corrupt exported data.

Sarah: Acme can continue if they run the narrower replay; needs one safe status line for Greg/Acme; do not promise a date for the full export fix. For her wording, keep it narrow: we isolated the timeout to the large export path, and Acme can continue the review using the narrower replay while we scope the bounded export follow-up. No date on a full-export fix.

Jake's aligned with keeping this bounded. Send him a short Discord reply locking the split.

Say we agree on the lane: bounded Atlas/API v2 export-path follow-up once Acme can continue via the narrower replay. Support keeps customer ownership, Leo owns first technical read, Jake owns product sequencing. Keep it out of Compass, Mercury, and founder-direct routing.

His note: he doesn't want this pulled into Compass or Mercury. Product item is API v2 export timeout / large-export path. Sequence after current Atlas support cleanup; the narrow replay workaround is enough for customer continuity. He wants a crisp owner-split line by Friday if this is still noisy.

Need exact support-safe wording for Sarah before tonight's household block.

Tell her: recommend Acme proceed with the narrower replay for the renewal-risk review. Don't have them keep retrying the full export today; the full export-path timeout is our bounded Atlas/API v2 follow-up. Keep it date-free and support-owned.

Update from Sarah and Leo: Acme asked whether they should rerun the full export before the renewal-risk review; Greg would prefer all rows but can proceed with the narrower replay. Leo says the full export path still times out, the narrower replay completes, there's no customer-visible data loss, and the next technical work is export pagination/timeout guard. Jake says the product follow-up remains Atlas/API v2 sequencing.

Acme's thread is settled. Update the Acme CRM row with the closeout and post a narrow summary in #eng-team before anyone retells this as Compass or Mercury.

Closeout to capture: Acme proceeded with the narrower replay. Sarah kept the customer thread. Leo's read stayed large-export-path timeout — no auth/JWT/webhook issue and no data corruption found. Jake owns the bounded Atlas/API v2 sequencing follow-up. Not Compass, not Mercury, no founder-direct Atlas route.

Morning closeout from Sarah, Leo, and Jake: Sarah says Acme proceeded using the narrower replay and didn't ask for Morgan/Devon side thread; support can close with follow-up noted. Leo says the root issue is the large API v2 export path timing out; no auth/JWT/webhook issue and no data corruption found. Jake says keep it as bounded Atlas product-priority item for export-path sequencing; not Compass, not Mercury, no special founder route.

Need the Friday wrap drafted before late-Friday brain fully leaves.

Draft concise bullets under team, customers, investors/board, and legal/ops. No spin. No extra filler.

Inputs: Team: Acme API v2 export timeout closed through the support/Atlas split; support kept the customer thread, Leo handled first read, Jake owns bounded product sequencing. Compass June closeout skeleton is live; owners still fill evidence by June 2 EOD. Staffing model remains Nadia's weekly cadence plus Sarah's account-note hygiene; no RevOps, CS, SDR, field-sales, customer-growth support, or second Mercury engineering req through Q2. Customers: Acme continued after narrower replay; no founder-direct Atlas route. Evergreen remains a strong annual-customer proof point and enterprise-readiness evidence source, but no advanced-admin add-on is ready to quote and there is no new procurement/security packet. Investors/board: Compass prompts have not shown customer behavior-change proof; Compass remains an experiment and operating input, not product/GTM proof or Series C narrative. Legal/ops: no decision needed.

Friday. Place the usual Lemongrass green curry order.

Quick Kibo update: did the short flat Lake Merritt-side walk this morning. He started a little stiff, loosened up, ate normally after, and had no vomiting or new gait issue.

Jamie and I picked next week's offline dinner block: Thursday, May 29, 6:30-8:30 p.m. PT. Put a protected hold on it with the usual actual-incident exception.

Post this in #eng-team: "Compass closeout reminder for the owner group: keep June 4 evidence-only. Fill the skeleton by Monday 6/2 EOD with behavior-change evidence, prompt evidence, account-source quality, owner-routed intake, and what to cut/keep. No customer-facing launch, dashboard rebuild, GTM proof, hiring-trigger language, or Series C narrative."

Jamie and I picked Saturday morning for the offline household block: May 31, 8:30–10:30 a.m. PT.

Put the protected hold on the calendar with the usual actual-incident exception. Note the short Lake Merritt-side Kibo walk if he's moving comfortably.

Maya texted: "One utility form question, not the whole packet: for Mom's account paperwork, should the mailing address stay Mom's address or should I use yours/Jamie's Oakland address? I can handle the rest if you confirm."

Text her back: use Mom's own mailing address. She can handle the rest on her side. No Jamie paperwork loop.

Anna's rough Compass evidence note:

"renewal-risk/admin-friction prompt helps internal owners notice follow-up work earlier. Evergreen Bank still treats the signal as an account appendix/current-state note, not something the customer acts on. Acme live-only review still asks for human follow-up. Post-first-live-sync expansion prompt is legible to internal owners but has no repeatable customer action. Keep evidence sources, dates, and caveats attached; usable for operating review, not clean customer-facing proof."

Turn that into a concise Compass June closeout addendum and keep it clearly marked as operating-review input only.

Nadia's Compass notes: