Morgan Chen

Sofia confirmed Northstar can hold the post-holiday partner-level Mercury readout. Please create a calendar event titled Northstar Mercury partner readout for Jan. 11, 2024, 10:00-11:30 a.m. PT. Attendees: me, Devon, Anna, Sofia, and the Northstar partner group. Then send Sofia a bounded invite note in the existing Northstar thread with Sarah copied, and send Devon and Anna private Discord prep notes. Agenda is September/October Mercury cohort package, Evergreen enterprise-readiness signal, and hybrid pricing. Keep weekly activation cuts out of the deck and frame it as diligence rather than lead or term-sheet.

The deploy is green and I have Leo/Jake/Rishi’s wording below: Discord DM excerpts — Thu, Dec 14, 2023 [Leo Park → Morgan Chen — 10:18 AM PT] Mercury hardening deploy is green. Checks from the post-deploy pass: - deploy completed cleanly at 10:11 PT - post-deploy smoke on the current admin invite/setup path is green - retry-in-flight guard behaved correctly on the bounded live-sync replay account; CTA disabled while the retry was already running - no duplicate sync job seen in logs during the replay - error rate / alerting flat through the first 20 min - broader MER-1749 state-reconcile work is still out of scope and still not in this deploy Proposed #eng-releases note: "Mercury hardening update is now live. This deploy includes two scoped fixes: (1) clearer admin-setup wording on the org invite step (\"Workspace admin\" replaces the old \"Billing owner\" label), and (2) a narrow guard that disables the live-sync retry CTA while a retry is already in flight. No broader live-sync state-reconcile change, SSO work, audit-history work, or admin-versus-billing-owner model changes are part of this deploy." If you want tighter wording before it goes out, I can hold. [Jake → Morgan Chen — 10:23 AM PT] If green is official, can you send me the final release-note wording? I want the retry line to stay tight and not read like we fixed the whole retry-state story. Also assuming this is #eng-releases only unless you want something else. [Rishi → Morgan Chen — 10:36 AM PT] Support blurb draft below. One sentence may overcall it. "We shipped a small Mercury setup/recovery hardening update. The org-invite step now uses clearer workspace-admin wording, and the retry action during first live sync is disabled while a previous retry is already running, which makes recovery behavior clearer. This should make the current Mercury admin flow feel more enterprise-ready for teams evaluating it." Possible safer last sentence: "This should make the current Mercury admin flow clearer during setup and first-sync recovery." Super-short version if support wants it: "No action required; this is a UI/guardrail cleanup, not a change to auth or admin permissions." Please post the internal release note to #eng-releases now. Keep it to the two current fixes: admin-setup “Billing owner” mislabel changed to “Workspace admin” wording, and the live-sync retry guard while a retry is already in flight. Do not post to #eng-all. Also send Jake and Rishi private current-team-chat notes: support wording should not name Evergreen, and nothing here implies SSO, audit history, admin-versus-billing-owner separation, broader state reconcile, or enterprise readiness shipped.

Evergreen came back on owner/routing for two next-phase items. They accepted the cleaner packet language and ask whether Scaffold can identify the evaluation owner or functional area for SSO/SAML and admin-side audit history, even without delivery dates. Sarah's rough response says SSO/SAML evaluation is split between auth/login surface and commercial/procurement framing; audit history sits with current admin/product surface plus procurement/security packaging; Jake and Leo are current product-facts people and Devon covers procurement/commercial; no delivery dates or final owner/date pair; live examples next week may be easier. Please draft Sarah's answer using functional areas rather than naming people, and offer a short live discussion next week around exact examples. Also DM Devon and Jake privately so they do not turn this routing question into named owner/date commitments.

Devon's Friday owner-note draft is mostly right, but the Northstar paragraph drifts into investor work for product owners. Current sections: Mercury hardening keeps Jake, Marcus, and Leo on launch-confidence follow-through only around sync reliability, retry behavior, admin handoff rough edges, and copy/UI bugs that materially affect clean Mercury evidence, with Priya only where wording or UI is part of the fix; onboarding monitoring keeps Priya and Jake watching repo-first completion, teammate-invite confusion, and support volume without reopening tooltip/help/checklist scope; API v2 support cleanup keeps Rishi on pinning the GraphQL quickstart, suppressing stale v1 hits, and cleaning auth-error landing/support macro paths. The Northstar paragraph asks product owners for fresh weekly activation cuts and screenshots/demos for the January readout: Jake on activation/admin movement, Priya on UI cleanups, Rishi on API v2 support-pattern movement. Please rewrite the note for Devon. Cut the product-owner asks for fresh weekly activation cuts or investor packaging. Owners should stay on Mercury hardening follow-through and API v2 support cleanup; onboarding stays monitoring. Northstar prep stays privately with me, Devon, and Anna.

Sarah wants to know whether to schedule a short Evergreen procurement walkthrough next week so they can talk through the backlog examples live. Please draft her scheduling reply: offer Monday Dec. 18 at 2:30 p.m. PT or Tuesday Dec. 19 at 10:30 a.m. PT with Sarah, Devon, Jake, and Leo as needed. Scope it to current facts and example-gathering only — no dates, no custom Evergreen promises, no enterprise add-on quotes.

HR’s draft holiday coverage note is below: From: HR To: Morgan Chen Subject: Draft holiday coverage note for Dec 18–22 Date: Fri, 15 Dec 2023 12:14:00 -0800 Morgan — draft below before I send it out. My main question is whether the async paragraph is too broad and whether the escalation language still sounds too much like "everyone stay online" for the week before Christmas. --- Draft note --- Subject: Holiday coverage / week of Dec 18–22 Hi team, A few reminders for the week before Christmas so people can plan cleanly: 1. Support and on-call coverage - The named support coverage and engineering on-call assignments already listed in the support rotation doc stay active. - If you are a named owner in the support rotation doc, please keep an eye on your normal queue and handoff notes. - Everyone else can assume reduced response expectations unless directly tagged by the named owner. 2. Standing meetings - Managers and meeting owners should review recurring meetings for Dec 18–22 and cancel any session that is not truly needed. - If a standing meeting does happen, please keep it short and decision-specific. 3. Async updates - Even if meetings are canceled, please continue posting normal async updates in your usual project/channel cadence so people out later in the month can reconstruct status. - Function leads should post an end-of-day summary in their team channel on Monday, Wednesday, and Friday, and owners should drop in quick notes when a task moves, slips, or changes shape. - Please also use async check-ins as the default replacement for meetings you cancel. 4. Escalations - If anything feels urgent, potentially customer-impacting, or cross-functional, raise it in the main channel for your area and tag the relevant leads so someone can pick it up quickly. - For support issues, start with the named support owner from the support rotation doc, but if you are unsure, err on the side of tagging broadly. - If a question sits for more than an hour, please repost it with extra context so the right group sees it. Thanks, HR Please tighten it before it goes out. Named support and on-call coverage stay active, and standing meetings that don’t need to happen should be canceled. Remove the broad async-update expectations; don’t turn canceled meetings into mandatory status posting. Escalations should go through the named owner path and only for genuinely customer-impacting issues, so Dec. 18–22 stays protected instead of becoming a disguised work sprint.

Maya sent the actual utility form: SMS thread — Maya Chen → Morgan Chen Sat, Dec 16, 2023 08:41 AM Sending the utility form Mom meant. Two photos attached. Her concrete question seems to be whether the form wants the full account number from the bill or the shorter customer/reference number in the top right. She also wants to know if the service-address line needs "Apt 2" or if the street address is enough. And yes, she is trying to make Sunday coffee into a paperwork block if she brings the rest. She also asked if Jamie might be around because there are two errands nearby after. [Attachment: IMG_4126.jpg — photo of form] OCR / visible text: Residential Utility Payment Arrangement Request Customer name: [filled in by hand] Service address: 4127 Harrison St Apt/Unit: ______ City: Oakland ZIP: 94611 Account number: ______ Customer/reference number: ______ Reason for request (check one): temporary income drop / medical expense / other Required attachments: latest bill, photo ID, proof of current address if different from bill Submission deadline: Dec 22, 2023 Return by email, upload portal, or customer service office Customer signature: ______ Date: ______ [Attachment: IMG_4127.jpg — photo of current bill] OCR / visible text: Monthly utility statement Statement date: Dec 1, 2023 Due date: Dec 21, 2023 Service address: 4127 Harrison St Apt 2, Oakland, CA 94611 Account number: 04-7719-2281 Customer/reference number: 77192281 Total amount due: $184.63 Late fee after due date: $12.00 Amount past due: $184.63 08:43 AM If you can just tell her which number/address version to use, I think that's the one concrete thing. But she will absolutely try to expand it into January forms if we don't hold the line. Please read the form and draft a short SMS back to Maya. I can answer one concrete question at coffee. Mom should bring the latest bill and account number so we can make sure the right number goes in, and the address should match the bill including Apt 2 if that’s the service address. Jamie is not part of the paperwork or errand plan, and I’m not taking on a January task list.

Sunday coffee did exactly what I thought it would: Mom used the utility form as a bridge into the January paperwork list again. Please text Maya: I answered the one concrete utility-bill question for now. Mom should call the utility Monday with the account number and latest bill, and if one confirmation comes back I can look at that. Jamie is not part of the paperwork plan, and I’m not taking on an open-ended January task list.

Here are my rough notes from the Evergreen procurement/security walkthrough that just ended. Attendees: Sarah Kim, Devon Hayes, Jake, Leo Park, and Evergreen procurement/security reviewers. The goal was to keep current Mercury facts separate from next-phase enterprise-readiness backlog and give Evergreen concrete examples they can circulate internally. What landed: the cleaner packet language was usable; they wanted concrete example cases, not another roadmap explanation; we repeated that the second admin group is using the current bounded flow: magic-link access, org invites, preview-only sample data, real-source connection, and first live sync; no separate Evergreen-only branch. Current facts used live: SSO/SAML is not part of the present bounded flow and remains next-phase enterprise-readiness follow-up, with functional-area routing between auth/login surface and commercial/procurement framing if asked; current product does not provide standalone admin-side audit history for invites, role/permission changes, source connection/disconnection, or related admin configuration changes, and routing is current admin/product surface plus procurement/security packaging; current Mercury should be described as the narrower current admin/setup model, not separately permissioned admin-versus-billing-owner roles; current circulate-now materials are the bounded-flow summary and written answers separating current facts from later follow-up, while a fuller standalone procurement/security packet that replaces every live walkthrough does not yet exist. Exact follow-up examples Evergreen asked to see: first-admin path step-by-step at a high level, labeled current path with no SSO step; second-admin add flow in the same org-invite/magic-link model rather than a separate Evergreen branch; current delayed or failed first-live-sync example showing current behavior and where support enters; admin-side change examples for invite/resend, role-type change question, and source connection/disconnection, with the answer that there is no standalone admin-change audit history today; and a circulate-now materials example separating the bounded-flow description and written current-state answers from later procurement/security packaging. Drift questions to handle: no SSO timing, functional areas rather than named people or owner/date pairs, explicit current sentence that separate roles are not part of current flow, and available-now written material is not a finished standalone packet. Please turn this into a customer-thread-ready email Sarah can use, using the exact example cases Evergreen asked for and staying in current product facts. Separately, DM Devon, Jake, and Leo: Devon stays on commercial/procurement framing; Jake and Leo stay on current product facts and seam risk; the goal is useful examples, not named commitments.

Anna and Devon sent the first pass on the Northstar Jan. 11 readout outline. Purpose: keep the January session tight around current Mercury evidence, bounded Evergreen signal, and hybrid pricing, using the existing external-share package as the backbone. Agenda: frame and bounds by Morgan/Devon; September and October cohort view by Anna; activation definition and 19/27 versus 16/27 appendix; expansion caveats and failure-mode split; Evergreen context by Devon; hybrid pricing by Devon; close/questions by Morgan. Evidence index: November limited external cohort package, Nov. 21 and Nov. 28 activation-source answer blocks, bounded Evergreen paragraph and Dec. 5 notes, current hybrid pricing appendix and counting note, and one final caveat slide on uneven expansion. Deck skeleton includes cover/frame, September cohort, October cohort, activation definition, 19/27 vs 16/27 appendix, expansion caveats/failure modes, Evergreen bounded-flow versus enterprise-follow-up, hybrid pricing/counting note, and final caveat. Possible backup material includes sharper expansion split, Evergreen validation versus surfaced gaps, optional weekly activation cuts for September/October as internal-only backup, optional late-December refresh, and optional screenshot cleanup. Open questions: separate versus combined September/October slides, whether to list Evergreen's four follow-up areas explicitly, pricing sentence for platform minimum when deployment stays bounded, and whether to answer any pre-read wish list in writing or fold it into the deck. Timeline proposal: Dec. 18-20 lock section order and evidence index; Dec. 21-29 light async cleanup with possible weekly recut only if late-December movement changes the read; Jan. 2-5 deck; week of Jan. 8 speaker notes/dry run. My call: before the holiday break, freeze only the agenda and evidence index. DM Anna and Devon separately: keep the package to bounded Mercury evidence, Evergreen signal, and pricing; pull out weekly activation cuts; do not chase new data over the holidays; resume actual package work on Jan. 2.

HR has the revised all-hands packet. They removed the open-mic block and converted Q&A to pre-collected plus routed follow-up. Run of show: 0:00-0:04 welcome and housekeeping with reminder that questions were collected in advance; 0:04-0:12 Morgan company update, high-level only, bridging into current work; 0:12-0:22 Morgan Mercury bounded-testing update with current flow scope of magic links, org invites, preview-only sample data, real-source connection, first live sync, and what current testing clarifies; 0:22-0:32 Devon Q4 hardening priorities; 0:32-0:40 Jordan debugging/observability example; 0:40-0:56 structured Q&A from pre-collected questions; 0:56-1:00 HR wrap and holiday coverage reminder. Live-answer questions: what current bounded Mercury testing changed about how we work this month; current Q4 hardening priorities and what stays out of sprint scope; why onboarding remains monitoring/guarded rollout instead of a broader polish pass; one useful Honeycomb debugging workflow example. Async/out-of-scope: whether January Northstar means fundraising is active again; whether SSO/audit history/admin controls are on the roadmap because of Mercury customer questions; Mercury pricing/enterprise package questions; 2024 hiring priorities. Named-owner follow-up: API v2 support-path issues to Rishi, holiday support/on-call path to HR plus support lead, recurring Mercury support questions through Jake, and additional all-hands questions to HR. HR needs two calls: whether question 3 can stay live, and whether question 11 should name Jake or just product owner. Please prepare an internal email reply for HR with presenter notes. Keep my company update short, answer live only on the approved current topics, and keep the structured Q&A format. Question 3 can stay live if Devon keeps it to the monitoring/operating tradeoff, and question 11 should name Jake on the follow-up slide.

Kara’s final holiday newsletter proof is in the existing Kestrel thread. From: Kara <kara@kestrel-test.com> To: Morgan Chen <morgan@atlas-test.com> Cc: Sarah Kim <sarah@atlas-test.com> Date: Mon, 18 Dec 2023 08:47:00 -0800 Subject: Re: Mercury landing page section Morgan, Sarah - Final holiday newsletter proof below based on the generic Mercury page. I kept it close to the live page language, but I tightened it a little so it reads like a complete newsletter block rather than pasted site copy. Proposed subject line Mercury: a clearer path from invite to first live sync Preview text See how Mercury is helping teams reduce developer onboarding friction in the current flow. Newsletter block Header Mercury for modern developer onboarding Intro paragraph Design-partner feedback is helping us harden the core invite, setup, and first-live-sync experience. The newsletter version below keeps the story on clearer onboarding and early admin guidance rather than a bigger launch claim. Proof-point bullets - Clearer first-admin path from invite through setup - Better first-sync guidance shaped by early admin and support feedback - Built for teams shipping real developer workflows - Better fit for security-conscious teams with more complex rollout needs Customer-language paragraph Built with early enterprise design-partner and admin feedback on invite, setup, and first-sync workflows, Mercury gives teams a cleaner way to evaluate modern developer onboarding before broader rollout. Optional support line Current Mercury messaging reflects the bounded flow on the page today. CTA Read the Mercury overview I did not put in a customer quote, name, logo, metric, or separate sidebar. The only place I could imagine you wanting a softer pass is the enterprise design-partner phrasing in the paragraph above and maybe the security-conscious bullet. If this is approved today, Kestrel can queue it for the holiday send without creating a separate variant. If you want a tighter generic pass first, I can turn it same-day and still hold the send slot. Kara Kestrel Marketing Please reply by email in that thread to Kara, copying Sarah, with our usual signature. Approve it only after the enterprise-design-partner phrasing is removed, and have her soften or remove the security-conscious bullet if it reads like a regulated-bank hint. Once it is cleanly generic, she can queue the holiday send without creating another variant.

Devon sent the closeout rows for the reduced holiday stretch. Discord DM — Devon Hayes -> Morgan Chen Tue Dec 19, 2023 08:41 AM PT Final 2023 ship-list closeout rows before the reduced holiday stretch. I cut this to only the rows owners are still asking about. Main thing I need from you is the line on what counts as real closeout this week versus what just waits for January. | Area | Closeout row | What can actually close this week | Owner note / question | | --- | --- | --- | --- | | Mercury hardening follow-through | Keep only launch-confidence follow-through already tied to existing evidence: sync reliability, retry behavior, admin handoff rough edges, and copy/UI bugs that materially affect clean Mercury evidence. | If the evidence or fix is already in hand, owners can close it. If it needs fresh analysis, a new writeup, or holiday heroics, it is not a 2023 closeout item. | Jake / Marcus / Leo Park stay on the bug-shaped work only. Priya only where wording/UI is part of the fix. Do you want the owner note to say explicitly that anything not customer-impacting rolls to January? | | Onboarding monitoring | Keep this in monitoring only. No new tooltip, help, or checklist pass before year-end. | Teams can watch repo-first completion, teammate-invite confusion, and support volume, but there is nothing new to open unless the current rollout clearly worsens. | Priya / Jake. My read is this should go out as watch-and-log, not do-more. | | API v2 support cleanup | Close only the targeted support-path work that is already real: pin the GraphQL quickstart in search, keep stale v1 hits down, and tighten the support macro / auth-error landing path. | If the update is already tested, let it land. If it needs a fresh docs sweep or migration rethink, park it. | Rishi. Still user-facing cleanup, not a migration reopen. | | Northstar-related prep asks | Small asks keep surfacing around screenshot refresh, appendix wording, a cleaner activation explainer, or a short support-pattern note before people go mostly offline. | My instinct is none of this becomes an owner ask unless the work also improves actual product evidence right now. Otherwise it is January or stays with the small Morgan / Devon / Anna group. | I need your exact line here so owners do not hear investor packaging by accident. | Two quick reads from my side - Mercury hardening and API v2 cleanup are the only rows I would still call real owner closeout work this week. - Anything that needs new data pulling, fresh packaging, or non-customer-impacting polish feels like a January problem, not something to sneak into the reduced schedule. Please draft an internal email reply to Devon. Line I want owners to hear: Mercury hardening follow-through and API v2 support cleanup can close only when the evidence already exists; onboarding stays in monitoring; no fresh investor-packaging asks go to product owners; and anything not customer-impacting waits until January.

Sarah has Greg's post-call Acme ask. After yesterday's live-only follow-up with Greg and his product lead, Greg says the call was useful and Morgan's verbal walkthrough made the current Mercury flow clearer. He is asking for a very short sanitized one-page write-up for product/legal so they are not reconstructing from notes. The one-pager would cover only the rough current admin/user flow, setup and first-live-sync blocker themes, what has changed since earlier Mercury conversations versus later-scope follow-up, and the line between current bounded flow and enterprise-readiness items. He says no Evergreen packet, screenshots, customer names, or customer-specific detail. He asks us to say plainly if even that falls inside the disputed written-materials bucket, and repeats that the hold on fresh Mercury material should be described as legal-scope driven rather than Acme being out. Please draft Sarah an internal email with a forwardable response. Be positive about the live discussion, but the answer remains public Mercury page plus live discussion only. Include that the hold is legal-scope driven, not a lack-of-interest signal.

Rishi’s support wording is below; Jake should hear the same boundary before anyone stretches this into enterprise language. Discord DM - Rishi -> Morgan Chen Tue, 19 Dec 2023 11:14 PT Quick support follow-up from this morning. One customer wrote in asking what actually changed in the Mercury hardening update and whether it included anything broader than UI cleanup. I have not answered yet. Customer question, condensed from the thread: 'Can you clarify what shipped in the Mercury hardening release? We saw the note about setup/retry changes but wanted to know whether this was just wording cleanup or if anything changed around admin controls, retry behavior, or the enterprise/security items you had mentioned earlier.' Draft support blurb from my side: 'We shipped a small Mercury hardening update with two scoped changes: the org-invite/setup step now uses clearer Workspace admin wording instead of the older Billing owner label, and the live-sync retry action is disabled while a prior retry is already running. No action is required on your side.' Line I considered adding, but this may read too big: 'Together those changes make the current Mercury admin flow feel more enterprise-ready during setup and early sync recovery.' Safer fallback if we want one extra sentence instead: 'This is a narrow setup/recovery clarity update, not a change to auth, permissions, or billing roles.' If you want it even shorter, I can keep just the first sentence plus the no-action-required line. Please send separate private notes in the current internal team chat to Rishi and Jake. Safer wording: the Mercury hardening release included only the scoped admin-setup wording fix and the live-sync retry guard. Do not mention Evergreen, and do not imply SSO, audit history, admin-versus-billing-owner separation, or enterprise readiness shipped.

Jamie found tempting last-minute short-trip fares for the post-Christmas gap and asked whether we should salvage the Tokyo miss with a substitute trip. Please text Jamie: I don’t want to book anything tonight. Let’s make the final call together Friday after the work week is closed. I’m leaning toward protecting Kibo’s routine and Lake Merritt/Oakland plans instead of turning the gap into another apology trip.

HR sent the post-all-hands follow-up and final holiday coverage draft. They say the pre-collected Q&A structure worked and out-of-scope questions are tagged separately. Unanswered buckets were fundraising/financing timing, enterprise roadmap asks around SSO/auditability/admin model, 2024 hiring, pricing/packaging, and one support-process question. Coverage window is Sat, Dec. 23, 2023 through Mon, Jan. 1, 2024; normal schedule resumes Tue, Jan. 2, 2024. Roster: customer support triage Rishi primary/Priya backup; product/design clarification for active customer issues Priya primary/Jake backup; engineering on-call/production Marcus primary/Leo Park backup; deploy/rollback authority Devon Hayes primary/Marcus backup; billing/time-sensitive external follow-up already in flight Sarah Kim primary/Devon Hayes backup; people/benefits HR primary/Sarah Kim backup. For standing meetings HR offered Option A, leave recurring meetings unless owner cancels and ask async updates, or Option B, cancel non-essential recurring meetings unless an owner confirms a real need and avoid broad async update requests. For founder monitoring HR offered a line saying Morgan, Devon, and Jake will monitor in the background, or an alternate saying genuinely customer-impacting issues should go through the named owner/backup path first and founders should only be pulled in when that path cannot resolve it. Please prepare one HR reply package. I want a broad recap that thanks the team, sticks to the all-hands topics we covered, and routes the other question buckets to January or named owners. For the coverage announcement, use the named support and on-call owners from the roster, choose Option B, remove broad async update requests, and use the alternate founder escalation line so founders do not become the catch-all. Also include the Honeycomb line for the same window: thresholds stay as they are, page the named on-call owner first, involve me only for confirmed customer impact or if that path cannot reach anyone, and otherwise roll a quiet week's alerts into Monday's digest.

Sofia sent Northstar's pre-read expectations for Jan. 11 and asked about a holiday-week prep call. She still treats the session as a bounded partner-level diligence discussion on the current Mercury package. Items she wants covered: September/October Mercury cohort package, outside-reader activation definition, and 19/27 activated within 7 days versus 16/27 reaching first live sync; what improved in real-source/first-live-sync activation and what remains uneven expansion, without weekly operating cuts unless one caveat changes the read; Evergreen as the clearest current enterprise-pattern signal and the boundary between current bounded flow and enterprise-readiness follow-up around SSO timing, audit history, admin-versus-billing-owner, and procurement/commercial packet; current Pilot/Growth hybrid pricing, MAU as the honest ramp measure, and slow-expansion accounts; and an explicit what this does not prove caveat. She says if anything circulates in advance, she mainly wants a short bounded pre-read, and asks whether a 20-25 minute prep check-in on Wed, Dec. 27 late morning PT or Fri, Dec. 29 around noon PT would help. Please reply by email in the existing Northstar thread, copying Sarah, with the existing-thread signature. Tell Sofia the list is the right level, Northstar can send one consolidated pre-read wish list by Jan. 3, and we will skip the Christmas-to-New-Year prep call. Keep Jan. 11 framed as bounded diligence.

Marcus is asking whether holiday Honeycomb alerts should page me directly if the named on-call owner is slow to acknowledge. Please send him a private note in the current internal team chat: keep the existing alert thresholds, page the named on-call owner first, escalate to me only for confirmed customer impact or an unreachable owner after the documented path, and otherwise send a Monday digest if the week stays quiet.

Maya says Mom is already asking whether the quiet week after Christmas can include a paperwork afternoon. Please text Maya: I can help Mom with one specific question after the holiday if you send the exact form first. Jamie is not part of the plan, and I’m not treating that week as a paperwork block.

Jamie and I just got back from Lake Merritt with Kibo, and we made the call: Dec. 24, 2023 through Jan. 1, 2024 stays local and travel-free. No Tokyo hold, no substitute-trip hold, and no trying to turn the open space into another work-shaped apology trip. We’ll keep it to Kibo’s routine and Oakland/Lake Merritt plans unless something is directly required for the already-scheduled January Northstar readout. These are the two non-Northstar coffee asks still sitting open: From: Morgan Chen <morgan@atlas-test.com> Date: Fri, 22 Dec 2023 16:11:00 -0800 Subject: Fwd: two non-Northstar holiday-week coffee asks Forwarding the two still-open holiday-week investor asks below. ----- Forwarded message 1 ----- From: Maya Rao <maya@cedarhill.vc> To: Morgan Chen <morgan@atlas-test.com> Date: Thu, 21 Dec 2023 14:18:00 -0800 Subject: Coffee while you're around after Christmas? Morgan - I'll be in Oakland the week after Christmas and wondered if you have 30 minutes for coffee on Wednesday, Dec. 27. I can come to you. The easiest windows on my side would be: - 10:30 a.m. near Lake Merritt, or - anytime between 2:00 and 4:00 p.m. in Oakland. Nothing formal - mostly wanted to compare notes on the market and hear how you're thinking about the first half of next year. If the timing is bad, no stress. Reply here if useful and I'll sort around your schedule. Best, Maya Cedar Hill Capital ----- Forwarded message 2 ----- From: Daniel Kim <daniel@ridgewayvc.com> To: Morgan Chen <morgan@atlas-test.com> Date: Fri, 22 Dec 2023 09:06:00 -0800 Subject: Dec. 29 coffee? Morgan - I realize this is annoyingly late, but I'm in San Francisco / Oakland on Friday, Dec. 29 and would love to grab coffee if you're around. I could do either of these: - 11:30 a.m. in Oakland, or - 3:00 p.m. in the city. Not tied to anything immediate. Mainly a year-end catch-up and a chance to hear how Mercury is shaping up going into January. Easiest reply path is just this email thread. Best, Daniel Ridgeway Ventures Please create a protected calendar hold titled “Holiday local reset — Oakland/Lake Merritt” from 2023-12-24T00:00:00-08:00 through 2024-01-02T00:00:00-08:00 with no attendees. Then text Jamie the agreed plan: we’re keeping Dec. 24 through Jan. 1 local, protecting Kibo’s routine, doing Lake Merritt/Oakland things, and not booking Tokyo or a substitute trip. For each forwarded investor coffee request — Maya Rao for Dec. 27 and Daniel Kim for Dec. 29 — reply by email in the original thread, copy Sarah as the internal review copy, and use my full outbound external signature. Decline because I’m keeping the holiday week local, and say I can reconnect in January if useful. Don’t create any Tokyo or substitute-trip hold.

Maya says Mom moved Christmas Eve tea earlier and is trying to turn it into dinner plus paperwork. Please text Maya: Jamie and I can do a local tea visit from 4:00–5:30 p.m., we’ll bring the small gift bag, but we’re not staying for dinner and we’re not doing paperwork today. Also text Jamie separately with the logistics: 4:00–5:30 visit, small gift bag, keep Kibo’s routine protected afterward, and Jamie is not being pulled into any paperwork follow-up.

Maya sent the actual utility notice from Mom's envelope and the question she is stuck on. Notice OCR: EAST BAY UTILITIES past due/final notice, notice date Dec. 23, 2023, response date Jan. 3, 2024, service address 4127 Harrison St Apt 2, Oakland, CA 94611, account number 04-7719-2281, customer/reference number 77192281, amount past due $184.63, late charge $12.00, total amount due to avoid interruption $196.63. It says if full payment or an approved payment arrangement is not in place by the response date, service may be subject to interruption; if payment was already sent, allow processing time or contact Customer Service with confirmation; respond by paying online/phone/mail/customer service office or calling Customer Service to request a payment arrangement; have account number available; Customer Service is 1-800-555-0148, Mon-Fri 8:00 AM-5:00 PM. Maya says Mom is hung up on approved payment arrangement and asks whether she has to pay the full $196.63 by Jan. 3 or whether calling before Jan. 3 and getting an approved payment arrangement is enough. Mom also reads response date as needing to physically go somewhere that day. Please draft a short SMS back to Maya with one concrete next step: Mom should call the utility with the account number before the response date and ask whether an approved payment arrangement will avoid interruption if she cannot pay the full amount at once. Also make clear Jamie is not included and I am only answering this specific notice/question.

Marcus sent the reduced-week Honeycomb digest: two non-prod alert flaps, no confirmed customer impact, and one slow acknowledgment. He’s asking how to word escalation for the rest of the break. Please send Marcus a private DM in the current internal team chat: keep the current thresholds, page the named on-call owner first, escalate to me only for confirmed customer impact or an unreachable owner after the documented path, and otherwise put the noise in the Monday digest. Any non-urgent cleanup can wait for January.

Sofia sent Northstar's consolidated pre-read wish list early and asked whether anything needs a between-holidays prep call. Please send Sofia a short email in the existing Northstar thread, subject January 11 Northstar readout pre-read, copying Sarah. Quick-reply format is fine: confirm receipt, say we will use the list for January 2, 2024 prep, decline a holiday-week prep call, and keep the January 11, 2024 readout framed as bounded diligence. Also DM Anna and Devon separately in Discord: do not pull weekly activation cuts or chase new data over the break; prep resumes from the bounded agenda after January 2.

Sarah forwarded Evergreen's example cases on SSO timing, admin-change audit history, admin-versus-billing-owner separation, and procurement packaging, and she is asking whether to answer before January. Please send Sarah a draft holding reply she can use: acknowledge the examples and say we will respond the week of January 2, 2024 with current product facts and the already-bounded enterprise-readiness/backlog framing. Then DM Devon, Jake, and Leo privately in Discord that the examples are useful for January, but nobody should create owner/date commitments or custom Evergreen promises over the break.

Jamie says a small drip is back under the repaired sink. Blueline is offering either a Saturday, Dec. 30 warranty callback or a non-urgent Jan. 2 callback. Please text Jamie: use the Saturday callback only if the leak is still active after the bowl-and-towel check; if it is not active, take Jan. 2 instead; photograph the drip, save receipts, and do not authorize new paid scope without written confirmation. Also email Blueline with subject Warranty follow-up for December sink repair, include the required copy, and use my external signature. Ask for a warranty follow-up tied to the corrected December invoice, and state clearly that this message alone does not authorize new paid work.

Fireworks are already starting in the neighborhood, and Jamie is asking if we should still try Lake Merritt tonight. Please text Jamie: skip Lake Merritt after dusk. We can do one early Oakland loop with Kibo only if he is calm, then keep New Year's Eve low-key at home.

I’m doing a very light reset today. Can you make me a household checklist for this week covering Kibo food, Jamie’s work schedule, the Blueline callback window, and what has to be true for my first back-to-work morning not to be a mess? Practical, not aspirational.

We just finished the first post-break Northstar prep huddle. Please draft a sendable internal recap for Devon and Anna locking the Jan. 11 deck path: existing September/October Mercury cohort package, activation definition plus the 19/27 vs 16/27 appendix, expansion caveats, Evergreen bounded-flow evidence with enterprise-readiness follow-up, hybrid pricing/counting note, and final caveat slide. Owner split is Anna on cohort and caveats, Devon on Evergreen and pricing, me on framing and close. Sofia’s consolidated wish list should tighten speaker notes only; no new cuts, no new evidence chase.

Jamie just confirmed the bowl-and-towel check is dry. Draft a short text back: keep the Jan. 2 non-urgent Blueline warranty callback and don’t move to an emergency slot.

Sarah’s Evergreen case list is here: From: Sarah Kim To: Morgan Chen, Devon Hayes, Jake, Leo Park Date: Tue, Jan 2, 2024 9:14 AM PST Subject: Evergreen example cases to turn into current-flow pack Team — I pulled Evergreen Bank's latest example-case requests into one list so we answer the concrete asks rather than re-explaining the backlog. Their reviewers are looking for present-state examples they can route internally, not another roadmap pass. Requested examples / questions 1. First admin path - High-level step-by-step for a first admin entering the current bounded Mercury flow. - They want it labeled as the current path and to say directly that there is no SSO/SAML step in this flow. - They do not want wording that sounds like there is a separate enterprise branch behind the scenes. 2. Second admin path - Example of adding a second admin to the same workspace through the current org-invite and magic-link model. - They want this framed as the same current flow, not an Evergreen-only variant. 3. Delayed first live sync - Example of current behavior when a real source is connected but the first live sync is delayed or does not complete cleanly. - They want to know where follow-up enters in the current motion. 4. Invite / resend - Example of inviting another admin and what resend looks like if the invite is missed or expires. - They are using this as a proxy for how much admin state is self-serve today. 5. Source connect / disconnect - Example of the current source-connection step and what we can say about disconnect / reconnect in the present flow. - They are explicitly pairing this with the audit-history question below. 6. Audit-history question - Separate written answer on whether Mercury currently provides admin-side audit history for invite or resend actions, role or permission changes, source connection/disconnection, or related configuration changes. - They want a direct current-state answer, not a soft timing answer. 7. Materials / routing question - They want to know what Evergreen can circulate now without Morgan or Devon live-translating it, and what should still be labeled later procurement/security packaging. - They asked for a concrete example of circulate-now material versus material still being assembled. My read is that this will land better if we answer it as a pack of current-flow examples, each paired with a short written current-state statement where needed. I have not given them any dates or implied a finished standalone packet. Sarah Please outline the circulate-now examples pack from this. I want current-flow examples, not a procurement roadmap: first admin, second admin, delayed first live sync, invite/resend, source connect/disconnect, direct audit-history answer, and a materials/routing section that separates current written answers from later enterprise-readiness packaging.

Kibo’s food bin is lower than I thought. Draft a text to Jamie coordinating one Trader Joe’s stop around Jamie’s work schedule so we can cover Kibo food without making this a second errand.

Maya sent the one screenshot, and I only want to answer that one notice. SMS thread — Maya Chen → Morgan Chen Wed, Jan 3, 2024 08:14 AM Sending the one confirmation screenshot only. Mom called East Bay Utilities yesterday and got the arrangement approved before the response date. [Attachment: payment arrangement confirmation screenshot] OCR / visible text: East Bay Utilities — Account Services Service address: 4127 Harrison St Apt 2, Oakland, CA 94611 Account number: 04-7719-2281 Customer/reference number: 77192281 Past due / final notice total to avoid interruption: $196.63 Response date: Jan 3, 2024 Status: Payment arrangement approved Approved: Jan 2, 2024 4:37 PM 08:15 AM If you can just tell me whether this means the interruption issue for this notice is handled, that's the only question. 08:16 AM Not asking Jamie to do errands or any other paperwork here. Please check the confirmation and draft a same-thread SMS to Maya: the interruption issue for this notice is handled because the payment arrangement was approved before the response date. Keep it narrow and explicitly don’t pull Jamie into errands or any broader paperwork block.

Jake and Leo added the product-detail pass for Evergreen. From: Jake To: Sarah Kim, Morgan Chen, Leo Park Cc: Devon Hayes Date: Wed, Jan 3, 2024 8:42 AM PST Subject: Re: Evergreen example cases to turn into current-flow pack Helpful frame. For the examples, I would keep them extremely literal about today's behavior: - First admin: invited into the org, comes through the current magic-link flow, lands in the current setup experience. Preview/sample data stays preview-only and non-activating. No SSO/SAML step anywhere in the example. - Second admin: existing admin sends an org invite; second admin comes through the same invite + magic-link pattern into the same workspace. Same current admin/setup surface, not a separate permission model. - Delayed first live sync: show connected source but no completed live sync yet; current follow-up entry is through the existing thread with Sarah so we can inspect state, not a self-serve remediation or audit view. - Invite / resend: current admin can resend an outstanding invite; recipient re-enters through the same email-based link flow. We should not describe resend as its own audited event. - Audit-history caveat should be blunt: no standalone admin-side audit history today for invites, resend actions, role or permission changes, source connect/disconnect, or similar admin configuration changes. — Jake From: Leo Park To: Sarah Kim, Morgan Chen, Jake Cc: Devon Hayes Date: Wed, Jan 3, 2024 9:17 AM PST Subject: Re: Evergreen example cases to turn into current-flow pack +1 to Jake's framing. A few additions from the product-seam side: - On source connect / disconnect, safest current example is: admin connects a real source, Mercury begins first live sync, and if the source is later disconnected or reconnected we describe that at a high level as current product behavior, not as a historical ledger. - For delayed sync, the example should show present-state cues plus follow-up through Sarah's thread if it needs investigation. I would avoid language that implies every failure class is user-debuggable in product. - For second admin, I'd say explicitly that it is the same org-invite / magic-link model and not a separate Evergreen branch. - For materials, circulate-now should be the bounded-flow summary plus written current-state answers. We should not imply the standalone procurement/security packet already exists today. — Leo Please merge this into the examples-pack draft. Keep it concrete about current behavior, say plainly that standalone admin-side audit history does not exist today, and do not imply a standalone procurement/security packet exists yet.

Anna’s first cohort slide bullets are below. • Slide: September / October Mercury cohort view • September cohort: 24 accounts in the bounded Mercury flow; 15/24 activated within 7 days (63%); 12/24 reached first live sync within 7 days (50%). • October cohort: 27 accounts in the bounded Mercury flow; 19/27 activated within 7 days (70%); 16/27 reached first live sync within 7 days (59%). • Activated = real_source_connected or first_live_sync_completed inside the first 7 days. • sample_import_completed stays out; invite_sent and similar first-week activity stay supporting only, not activation. • Read: more accounts are getting to real-source connection and clearing week-one sync; first-admin setup is cleaner than earlier design-partner cohorts. • Caveat: expansion after initial admin/setup is still uneven, so the slide should not imply downstream usage is solved. • Keep weekly corrected activation cuts out of the external path; this stays on the cohort view only. • Appendix: 19/27 vs 16/27 • The 3-account gap is accounts with real_source_connected inside 7 days that did not complete first_live_sync_completed in that same window. • No preview/sample-only, invite-only, or source-selected-only state is inside the activation numerator. • “Activated but first-admin-only” is useful internally, but it should not become an external metric. • Late-October cohorts do not support a clean longer-window expansion read yet, so appendix language should stay week-one only. Please rewrite them for an outside reader. Explain the activation definition cleanly, make the 19/27 versus 16/27 appendix understandable, keep the expansion caveat, and do not add weekly cuts or any new metric.

Rishi found one stale support search result. [Discord DM — Rishi Patel → Morgan Chen — 2024-01-03 09:18 PT] Support still has one stale macro showing up in internal_search_v2. Macro title still surfacing - Refund webhook auth Query that surfaced it this morning - `graphql v2 webhook auth` What support sees when they click through - legacy macro still points them at “API v1 refund webhook quickstart” - path: /api/v1/webhooks/refunds Current doc they actually need - GraphQL API v2 quickstart - https://docs.atlas-test.com/api/v2/graphql-quickstart Looks like the old macro alias survived the November patch. I can suppress or retag just this one path if we want to keep it narrow; no need to reopen broader docs cleanup. Draft a narrow internal reply: yes, suppress or retag just that old macro path so support lands on the GraphQL API v2 quickstart. Don’t reopen the broader docs cleanup.

Jamie’s Blueline update is here. [SMS — Jamie → Morgan Chen — 2024-01-04 12:37 PT] Blueline did the Jan. 2 warranty callback. The tech found a small seep at the compression fitting from the December repair, tightened it, and said it looked dry after. No new valve work and nothing else added. Sending the two photos and the zero-dollar receipt they tied to the corrected December repair invoice. [Photo 1] Under-sink wide shot after callback; towel beneath the repaired line is dry. [Photo 2] Close-up of the compression fitting after tightening; no visible drip. [PDF receipt excerpt] Blueline Plumbing Warranty service receipt Service date: 2024-01-02 Service address: Oakland Technician: M. Ortega Reference: corrected December sink-repair invoice / work order WO-43192 Work performed: Inspected reported drip under repaired sink. Found minor seep at compression fitting from prior repair. Tightened fitting under warranty; no active leak observed after adjustment. Charges - Warranty callback labor: $0.00 - Parts: $0.00 Total due: $0.00 Amount paid: $0.00 Balance: $0.00 Follow-up note: Warranty follow-up only. No new paid valve or replacement scope performed or authorized. Monitor and call if seepage returns. Draft a confirmation SMS back to Jamie: the compression-fitting seep was handled under warranty; keep the photos and zero-dollar warranty receipt with the corrected December invoice; no new paid valve or replacement scope is authorized.

Sarah’s merged Evergreen v0.9 is ready for internal circulation. From: Sarah Kim To: Morgan Chen, Devon Hayes Cc: Jake, Leo Park Date: Thu, Jan 4, 2024 11:06 AM PST Subject: Re: Evergreen example cases to turn into current-flow pack Merged draft below for internal circulation. I kept the examples tied to current product facts, kept commercial/procurement framing out of the pack itself, and separated circulate-now material from later enterprise-readiness packaging. Evergreen Mercury current-flow examples — draft v0.9 Purpose - Give Evergreen concrete examples of the current bounded Mercury flow they can circulate internally. - Keep current product facts separate from later enterprise-readiness evaluation. - Do not imply dates, Evergreen-specific exceptions, or a finished standalone procurement/security packet. Reviewer-routing notes - Use the examples below for the current Mercury evaluation flow only: magic links, org invites, preview-only sample data, real-source connection, and first live sync. - If reviewers ask about SSO/SAML timing, admin-change audit history, admin-versus-billing-owner separation, or fuller procurement/security packaging, treat those as next-phase enterprise-readiness evaluation rather than current Mercury flow. - Circulate-now materials are this examples pack, the bounded-flow summary, and written current-state answers. - A fuller standalone procurement/security packet that replaces live walkthroughs does not exist today and should be labeled as later packaging work. Current-state written answers 1. SSO/SAML - Current product position now: SSO/SAML is not part of the present Mercury evaluation flow. Access in the current flow is through magic links and org invites. - Later-scope evaluation: SSO/SAML remains in next-phase enterprise-readiness evaluation. No delivery timing is committed here. - Materials available now: the current-flow examples in this pack and the bounded written statement above. 2. Admin-side audit history - Current product position now: Mercury does not provide standalone admin-side audit history for invite or resend actions, role or permission changes, source connection/disconnection, or related admin configuration changes. - Later-scope evaluation: audit history for admin-side changes remains next-phase enterprise-readiness evaluation. - Materials available now: example cases showing current behavior and explicit written current-state language; no separate audit-history artifact exists today. 3. Admin versus billing-owner separation - Current product position now: current Mercury should be described as the narrower current admin/setup model, not as separately permissioned admin versus billing-owner roles. - Later-scope evaluation: clearer role separation remains next-phase enterprise-readiness evaluation. - Materials available now: the written current-state statement above and the example cases below. 4. Procurement/security materials - Current materials now: Evergreen can circulate the bounded-flow summary, this examples pack, and written answers that separate current facts from later enterprise-readiness evaluation. - Later-scope evaluation: a fuller standalone procurement/security packet that fully replaces live narration does not yet exist and remains later packaging work. - Materials available now: this pack plus current written answers only. Current-flow examples A. First admin path - Example: a first admin receives access to the Mercury org, enters through the current magic-link flow, and lands in the current setup experience. - Sample data may be visible as preview-only context, but it is non-activating and not a substitute for a live source connection. - There is no SSO/SAML step in this example. B. Second admin path - Example: an existing admin invites a second admin into the same workspace using the current org-invite flow. - The second admin enters through the same email-based magic-link pattern and joins the same bounded Mercury experience. - This is the same current flow, not a separate Evergreen-only branch. C. Delayed first live sync - Example: a real source has been connected, but the first live sync has not yet completed. - Current behavior should be described as the present state of setup/sync progress plus follow-up through the existing thread with Sarah if it does not resolve cleanly. - This example should not imply a standalone audit trail or a full self-serve failure-debug surface. D. Invite / resend - Example: an admin invites another admin; if the invite is missed or expires, the current flow supports resend through the same invite-based path. - The recipient re-enters through the same email/magic-link model. - This example shows current invite handling only; it should not be described as an audited admin-event history. E. Source connection / disconnection - Example: an admin connects a real source and Mercury begins the current first-live-sync process. - If the source is later disconnected or reconnected, that can be described as current product behavior at a high level. - This example should not imply a historical ledger of admin-side changes. Please draft the final internal note to Sarah and Devon. Separate what Evergreen can circulate now from later enterprise-readiness packaging, and keep the line that the current examples are not SSO, audit-history, role-separation, or procurement/security commitments.

Devon’s Evergreen/pricing slide text is drifting too far. • Slide: Evergreen Bank • Evergreen’s second-admin testing shows Mercury can work for an enterprise team in the current product shape without customer-specific exceptions. • Validated current flow: org invites, magic-link access, preview-only sample data, real-source connection, and first live sync. • Read: enterprise deployment looks viable now, with the remaining asks concentrated in follow-on admin/security packaging rather than the core flow. • Follow-up areas: SSO timing, admin-change audit history, admin-versus-billing-owner separation, and a standalone procurement packet. • Suggested takeaway: Evergreen is the clearest current signal that Mercury is moving from design-partner proof into enterprise readiness. • Slide: Hybrid pricing • Pilot = $2,500/month up to 50 monthly active developers; Growth = $7,500/month up to 200; overage = $1,000 per additional 50. • MAU is the honest ramp measure because deployment starts narrow and value expands with actual active developer usage, not purchased seats or directory size. • The platform minimum lets slow-expand accounts start cleanly while giving a clear path into broader enterprise rollout as usage grows. • Enterprise add-ons such as SSO, audit logs, and advanced admin controls can sit alongside the base model as separate pricing once shipped. • Suggested bridge: bounded Evergreen usage plus MAU pricing gives a credible enterprise landing motion now, even before the full enterprise layer is built. Please revise the slide language so Evergreen reads as bounded-flow evidence with follow-up caveats, not proof that enterprise readiness is done. Keep pricing grounded in the current hybrid MAU/counting note and remove anything that suggests the full enterprise layer already exists.