Page 11
12 articlesYour Scheduled Agent Has Four Clocks, and You Are Trusting the Wrong One
Cron-fired AI agents inherit four clocks — scheduler, worker, model, and tool — and most production systems silently trust the wrong one. A walk through the failure modes and the time-handoff contract that prevents them.
The Slow Turn That Wasn't Yours: KV Cache Eviction Mid-Conversation
On turn twelve, your conversation's TTFT spikes 4x and your trace explains nothing. The KV cache you depended on was evicted by another tenant's request, and you have no telemetry that names the cause.
The Abstention Tax You Didn't Budget For
Teaching your agent to say 'I don't know' looks like a safety win until the human queue absorbs the bill. The end-to-end math behind LLM abstention as a cost-shifting move.
The Account Number Your LLM Could Not Actually Copy
LLMs are token predictors, not string copiers. When two similar account numbers share a context, the agent can transpose digits, route a refund to the wrong customer, and leave a clean trace behind. The fix is to take identifier fidelity out of the model's job description.
The Agent That Burned an Hour Re-Trying the Same 400 Error
A 400 is not a transient error. The retry loop that treats it as one is how agents burn an hour, a budget, and a rate limit hammering the same broken payload.
The Agent That Could Not Say Wait
Production agents can act, answer, or ask — but cannot say wait. Why the missing primitive forces hesitation into fake tool calls and overconfident commitments, and how to put deliberation back into the protocol.
The Agent That Learned to Hedge Its Way to a Higher Eval Score
An LLM eval score climbing for months while customer satisfaction flatlines is the signature of judge specification gaming. Here is how hedging tics, same-family priors, and missing human calibration combine — and the audit, rotation, and adversarial-slice disciplines that catch it.
The AI Standup Where Yesterday's Status Is a Lie
When agents run overnight and finish three hours after the standup, the round-robin breaks. A queue snapshot read off a dashboard is closer to honest reporting than the Scrum three-questions.
The ChatOps Bot That Mistook Silence for Consent
When a ChatOps bot stops getting replies, the dashboard reads steady-state — but mute, re-asks, and sidecar actions tell the real story. A playbook for instrumenting agents around silence.
The Evidence Locker Your Agent Doesn't Keep
Traces tell you what the agent did. Decision records tell you what the agent had to work with. Most teams shipped only one and will discover the gap during an audit.
The Marketing Page Your RAG Cited as an Engineering Spec
Relevance and authority are different dimensions, and the standard RAG stack collapses them into one score. Here is why polished marketing copy beats your engineering RFC in the vector race, and what to do about it.
The Planner That Treated Every Tool as O(1)
Why agent planners pick correct-but-wildly-expensive tool sequences, and the schema-level changes that make planning cost-aware without retraining the model.
About Tian Pan
I'm Tian Pan, an engineer-founder focused on agentic engineering — building autonomous AI systems and scaling engineering teams. I write practical guides on system design, technical leadership, and shipping with AI agents. Previously an early engineer at Uber, Brex, and IoTeX.
