Writing · Nº 041065 essays
The writing
Ideas from inside the work. Explore technology, entrepreneurship, and the practice of building things that last.
36 essays found · Page 1 of 3
Clear filtersEngineeringWhen Scraping Breaks at Scale, and How to Design Around ItSmall scrapers work until they do not. The failures at scale are predictable, and you can design for them up front instead of firefighting later.EngineeringHow to Set a Latency Budget for AI FeaturesA latency budget forces you to decide how slow an AI feature is allowed to be before you build it. How to set one, spend it across the pipeline, and design around it.EngineeringHow to Keep AI-Generated Code MaintainableKeeping AI-generated code maintainable comes down to one thing: a spec that keeps every feature consistent. Here is how to stop AI output from rotting over time.EngineeringVisual Regression Testing Keeps Shared Components SafeOne change to a shared component can silently break the look of every app that uses it. Visual regression testing is how you catch it. Here is how to set it up.EngineeringHow to Design a Review Queue for AI OutputA human review queue is how AI-native products handle low-confidence output at scale. How to design one that routes the right cases, ranks them, and closes the loop.EngineeringHow to Make AI Agents Recover From FailureAI agents fail. Timeouts, bad output, dead APIs. Here is how to design agent failure recovery with retries, fallbacks, and escalation so a job never just dies.EngineeringAdopting an AI Dev Platform Mid-ProjectBringing an AI dev platform into a project already underway is different from starting fresh. Here is how to adopt one without derailing the work in flight.EngineeringHow to Design a Product Around a ModelProduct design when the model is the core mechanism: shape the workflow around what it does reliably, design for non-deterministic output, and build the recovery paths.EngineeringSelf Host vs Managed Cloud: How to DecideA grounded decision guide for owning your stack versus renting it: cost, control, lock-in, on-call burden, and when managed services become a liability.EngineeringFall Back to Deterministic When the AI Is UnsureThe strongest AI-native systems keep a deterministic path underneath the model. When to route to rules instead of the model, and why a hybrid beats pure AI in production.EngineeringHow to Connect AI Agents to Your ToolsConnecting AI agents to your tools is where most automation stalls. Here is how to wire agents into your stack with scoped access, real connectors, and control.EngineeringHow to Move an Existing App Onto Your Shared FoundationBuilding a shared foundation is easy on new apps. Retrofitting an old one is the hard part. Here is how to migrate a live app onto the foundation without a rewrite.