Tagged
Performance
Stories filed under Performance — seasonal recipes, chef's notes, and updates from The Cafe at The Inn in Benicia.
AI AgentsPerformance
The cost of a chatty tool call
Every round trip is latency the user feels. Batch what you can.
Parallel AI·
EngineeringPerformance
Rate limits are a design constraint, not an error case
Backoff belongs in the architecture, not in a try/except at the edge.
Parallel AI·
Performance
Prompt caching changes what is expensive
A long stable prefix is cheap. A shuffled one is not.
Parallel AI·
Come hungry. Leave happy.
Breakfast, lunch, and Friday night dinners — all made fresh, all served with care. Order ahead or stop by 145 East D Street.