Testing AI agent reliability across the full coordination loop, including failure recovery, evidence preservation, and authority enforcement.
Designing stable AI agents: separate retries, reconciliation, and recovery to control interventions, not just model retries.
Enterprise AI is a distributed system requiring hybrid cloud and edge architecture, not a single platform.
Explores the operational costs and design considerations of scaling an agentic coding SDK for concurrent repository fixes, focusing on isolation, concurrency, and persistence.
Explores Azure Local as a distributed infrastructure platform using a power grid metaphor, covering architecture, governance, and operational benefits.
A guide to deciding whether to split a monolith into microservices, featuring five key questions to evaluate before making the transition.
Explains compensating transactions as a key resiliency pattern in distributed systems, using real-world payment reversal examples.
Explores how Domain-Specific Languages (DSLs) guide LLMs to generate reliable, intent-aligned code, using Tickloom as an example.
Explores a subtle failure in event-driven systems where messages are acknowledged but not processed, causing silent data loss.
Explores whether retries and timeouts should be handled by the application or service mesh, balancing infrastructure simplicity with business context.
Explores how retries, circuit breakers, and fallbacks can reduce system resilience if misapplied, using a message queue example.
Analysis of Coinbase's 10-hour outage due to single-AZ dependency and lack of automated failover, highlighting reliability lessons for tech companies.
Analysis of CockroachDB's scalable leader lease protocol for multi-consensus groups, addressing failover and liveness issues.
A monthly roundup of non-trendy software resources, including articles on TUI accessibility, CSS font fallbacks, and a proposal for a universal do-not-track flag.
Explains deterministic routing as a key technique for reducing consistency problems in distributed systems at scale.
Summary of the Antithesis BugBash Conference 2026, covering software verification, distributed systems, formal verification, and test coverage talks.
Review of SysMoBench, a benchmark evaluating AI's ability to write TLA+ formal specs for real-world distributed systems.
Explores idempotency in system design, its patterns, and critical role in reliable distributed systems, generative AI, and AI agents.
Interview with a Fly.io developer discussing her work on Corrosion, a distributed SQLite replication system, and other infrastructure projects.
Explains the difference between high availability and high resiliency in system design, and why both are crucial.
