Observability in Systems
Giovanni and Phillip discuss the parallels between observability in large language models and conventional systems. They emphasize how upstream decisions can lead to user experience issues, even when the components themselves are functioning correctly. By sharing a real-world example, they illustrate the importance of proper instrumentation and tracing to uncover hidden latency problems, ultimately improving user satisfaction.In this clip
From this podcast
Software Engineering Radio - the podcast for professional software developers
SE Radio 610: Phillip Carter on Observability for Large Language Models
Related Questions