Testing Distributed Systems
Testing distributed systems is hard due to scale and changing the system's environment can affect how the system performs. Sampling and observability offer ways forward.
Testing distributed systems is hard due to scale and changing the system's environment can affect how the system performs. Sampling and observability offer ways forward.
Humans do not merely recognize patterns. They construct progressively richer representations from their observations. This article proposes a typed model of cognition in which learning is the creation of higher-order representational objects and...
Recent reports of AI going rogue are misleading and sensationalist. The AI is not rogue: agentic systems pursue objectives literally, making explicit boundaries, constraints, and escalation paths as important as the goals themselves.
Estimates are poor tools for managing uncertainty and humans are poor at generating them.
The AI race may not be won by the smartest model, but by whoever can embed enough intelligence into the physical economy at the greatest scale.
LLMs excel at generating new artefacts but fail when asked to interpret and safely update existing ones. Their value lies in text creation tasks (research, summary, explanation, deep dives), not system modification.
Agentic systems generate far more code but also around 1.2 production defects per day, requiring extra fix capacity to stay stable.
Explaining a problem to your manager is only half a solution. How to complete the circle, identify solutions, and show cost, benefits and delivery timelines.
LLM-centred agentic systems will always make mistakes, just like human engineers. Human-authored tests are all that can stop an agentic quality drop.
A walk‑through of how gossip communication alone produces a single global order without leaders, coordination, or messages in a Byzantine faulty distributed system.