--- a/eng-update-draft.md
+++ b/eng-update-draft.md
@@ -1,46 +1,32 @@
-# Engineering Status Update — Week of April 7
+# eng update — week of april 7
 
-## Platform Infrastructure
+## infra
 
-This week marked a pivotal moment in our infrastructure modernization journey. The team successfully completed the migration of our primary database cluster from PostgreSQL 14 to PostgreSQL 16, ensuring enhanced performance and long-term maintainability. This initiative underscores our commitment to maintaining a robust and scalable foundation for our growing platform.
+postgres 16 migration is done. 47 schema definitions updated, zero downtime — honestly didn't expect that to go as smoothly as it did. shoutout to Dana for figuring out the parallel batch approach for schema diffs, that's what made the zero-downtime thing possible.
 
-The migration encompassed a comprehensive update of 47 schema definitions, highlighting the depth and complexity of this undertaking. Zero downtime was achieved during the transition, reflecting the team's meticulous planning and execution.
+## api performance
 
-## API Performance
+- /search p99 latency: 340ms → ~180ms after the pgbouncer rollout. not bad for a config change
+- connection overhead down ~40% across the board
+- redis caching layer is live for product catalog
 
-Our API team has made significant strides in optimizing endpoint response times. Key achievements include:
+## auth rewrite
 
-- Reduced p99 latency from 340ms to 180ms on the /search endpoint, showcasing the team's dedication to delivering a seamless user experience
-- Implemented connection pooling via PgBouncer, contributing to a 40% reduction in database connection overhead
-- Deployed Redis caching layer for the product catalog service, enhancing its ability to handle increased traffic volumes
+94% test coverage, token refresh works, session management is way cleaner than the old spaghetti. rate limiting on login is in. on track for staged rollout april 14 — 10% first, then 25%, then 100% over two weeks assuming nothing catches fire.
 
-These improvements represent a crucial step forward in ensuring our platform can scale effectively to meet growing demand.
+## mobile
 
-## Auth Service Rewrite
+react native upgrade to 0.76 is done (was on 0.73). startup times are better and android crash rate dropped from 2.1% to 0.8%.
 
-The authentication service rewrite stands as a testament to our engineering team's technical expertise. The new service, built on OAuth 2.1 with PKCE, boasts a comprehensive test suite encompassing 94% code coverage. This groundbreaking effort sets the stage for future security enhancements and ensures compliance with industry best practices.
+v4.2 going to testflight by wednesday.
 
-Key milestones achieved:
-- Token refresh flow fully implemented and validated
-- Session management refactored, fostering a more maintainable codebase
-- Rate limiting added to login endpoints, symbolizing our proactive approach to security
+## what's next
 
-The auth rewrite is on track for staged rollout starting April 14, marking a key turning point in our platform's security posture.
+- auth rollout starts april 14
+- search team kicking off semantic search prototype with embeddings
+- hiring: two backend roles close friday, three strong candidates in final round
 
-## Mobile App
+## risks
 
-The mobile team has been instrumental in driving forward our cross-platform strategy. React Native upgrade from 0.73 to 0.76 is complete, nestled within a broader effort to modernize our mobile stack. The upgrade contributes to improved startup times and enhanced developer experience.
-
-Crash rate has been reduced from 2.1% to 0.8% on Android, underscoring the team's commitment to quality and reliability.
-
-## What's Next
-
-- Auth staged rollout begins April 14 (10% → 25% → 100% over two weeks)
-- Search team starting work on semantic search prototype using embeddings
-- Mobile releasing v4.2 to TestFlight by Wednesday
-- Hiring: two backend roles close Friday, three strong candidates in final round
-
-## Risks
-
-- The evolving landscape of third-party API dependencies presents ongoing challenges. Stripe webhook reliability issues persist — we're implementing retry logic as a vital safeguard
-- CI pipeline is experiencing intermittent failures on the arm64 runners, reflecting broader infrastructure complexity
+- stripe webhooks are still flaky. adding retry logic but it's not a real fix, just a band-aid until stripe gets their act together
+- arm64 CI runners keep randomly dying. I have a workaround (pin to x64) but it's slow and I hate it