More visits had long waits. What changed?In a fictional outpatient clinic, 28% of visits had a wait longer than 30 minutes before a scheduling change. Afterward, 52% did. The clinic saw more walk-ins in the second period. Within each visit t3 min read·doesithold.hashnode.dev5d ago
An eval split can leak even when every case ID is uniqueYou turn a support conversation into two evaluation cases. One asks the original question. The other paraphrases it. If one lands in the tuning set and the other in the holdout, unique case IDs give y2 min read·doesithold.hashnode.dev6d ago
The cheaper model call made each successful task cost moreA candidate model is cheaper per request. In a small replay of the same tasks, it completes fewer of them. The invoice still drops, so you have to divide by the work that actually got done before deci3 min read·doesithold.hashnode.dev6d ago
A coding agent turned the test green by deleting the assertionA coding agent tells you it fixed a duplicate-charge bug. The test passes. In the diff, the line that checked the number of charges is gone. Here’s a small, fictional example you can run locally. It c3 min read·doesithold.hashnode.dev6d ago
One passing test can hide shared state in PythonReproduce a Python mutable-default bug, fix object ownership and test independent calls plus an explicit empty list. Runnable examples included.4 min read·doesithold.hashnode.devNov 1