GuideAI in production
How to evaluate incident-response agents
Build repeatable cases and check recovery independently, with a runnable Python grader.
Articles
Guides, news and explanations about AI and reliable systems.
GuideAI in production
Supervising coding agents with Herdr
How Herdr keeps coding sessions available while you detach, and what changes when its server restarts.
CommentaryAI in production
When agents can change production
The permissions, state and recovery paths an agent needs when its tool calls can change a running service.
ExplainerAI in production
The operational cost of AI tokens
Account for context limits, retries, latency, and the work behind a useful result.
GuideAI in production
Comparing incident sources with NotebookLM
Conflicting restart instructions show where a source-backed notebook helps: finding the relevant passages, dates and exceptions for the runbook owner.
ExplainerPlatform engineering
GPU capacity depends on power and cooling
Power and cooling affect the work a GPU pool can sustain. Device signals, request completions and queue growth help distinguish a physical limit from a workload change.
GuideAI in production
Evaluating Claude Opus 4.6 for incident work
A rollback exception tests whether Opus 4.6 preserves decisive evidence across a long session, and whether the resulting answer arrives in time to help.
GuideAI in production
Recovering an interrupted agent skill
A timed-out scaling request needs a durable record and destination-state checks. Reusable operational skills need to recover unfinished work as well as start it.
ExplainerAI in production
How an AIOps finding becomes an agent action
A rollback proposal crosses from recorded evidence into a changing system. The handoff carries source references, current preconditions and unfinished actions.
CommentaryObservability
What AIOps alert correlation can hide
Grouping can shorten the incident queue while hiding a second failure. Following a replay past the first mitigation reveals whether that failure stays visible.
GuideIncident response
Keeping incident state consistent across PagerDuty, Jira and Slack
A reference design for duplicate deliveries and partial API success, with durable incident identifiers and summaries drawn from the resulting record.
ExplainerPlatform engineering
AI storage and recovery in NetApp and NVIDIA architectures
Model loading and checkpoint recovery stress storage differently. Reference configurations help explain which workload and failure conditions their results cover.