Link Dump #262

Friday the 13th! Lucky Friday! In such case I'm sure today you will read something life changing :)

Link Dump #262

Friday the 13th! Lucky Friday! In such case I'm sure today you will read something life changing:

  1. Software Architecture
    1. AI readiness starts with data ownership
      Centralized Business intelligence and data teams often manage pipelines for data they didn't create, leading to downstream quality issues and severe access bottlenecks. Real AI readiness requires placing data ownership directly with the business domain teams who understand the context and business rules. Central tech teams shift from data gatekeepers to platform enablers.
    2. Why We Do What We Do
      Technologies and tools change constantly, but the foundational goals of platform engineering remain identical regardless of AI advancement. Infrastructure must be reproducible, visible, shift-left compliant, and explicitly owned. This article explains how preserving architectural context protects long-term platform resilience.
    3. 429s, Webhook Storms, and the Wrong Fix #PickOfTheWeek
      Facing a sudden storm of 429 errors from an upstream API, adding a rate limit to drop excess traffic stops the alerts and turns the monitoring dashboard green. While quietly discarding thousands of production tasks. Learn how combining debouncing, throttling, and concurrency caps protects upstream APIs without dropping data.
    4. Metastable Failures and Retry Storms: When Your Resilience Code Is the Outage
      Discover how unbudgeted retries fuel metastable failures and reduce useful server goodput to zero. Learn how token-bucket retry budgets, jittered backoff, and deadline-aware load shedding stop your resilience code from attacking your own infrastructure.
  2. Software Development
    1. Macro trends in the tech industry | November 2026 #PickOfTheWeek
      This practical deep dive into Chris Ford’s macro trends analysis explores how platform teams must adapt for the agentic era. The breakdown details methods for building gateway routing layers, instituting risk-tiered review WIP limits, sandboxing agent execution, and managing the harness as a platform product.
    2. The harness is a product #PickOfTheWeek
      If your engineering org treats AI agent setup as disposable boilerplate while chasing frontier model benchmarks, your execution layer is built on sand. The author demonstrates why the harness is the true operating layer for enterprise AI. Learn why defining product ownership, metrics-based evaluations, and clear operational boundaries is essential for scaling agentic systems.
    3. Don’t Snooze Your First Event Storming
      Still snoozing on your first Event Storming? Spot the situations from last week that call for one, find a wall and an ally, then book a 90-minute session.
    4. bliki: Generated Glossary #PickOfTheWeek
      Glossaries are not passive documentation; they are active inputs to Agentic Engineering. Beyond guiding initial code and document generation, glossaries should power automated sensors inside the agent harness. These deterministic checks verify that generated internal variables, API schemas, and customer-facing UI labels strictly adhere to the target bounded context's vocabulary.
  3. Leadership
    1. The Hidden Value of Feeling Bad
      A common leadership blind spot is rushing to soothe, silence, or explain away a team member's frustration or anxiety. True coaching requires holding space for discomfort so team members can diagnose the underlying signal and build emotional agility.
  4. Fun
    1. Concatenating jargon
    2. How to create your roadmap