Concept detail
Observability
43 tagged videos · 5 connection videos · degree 0
Publication timeline
- 2014-121 video
- 2019-041 video
- 2023-051 video
- 2025-071 video
- 2026-031 video
- 2026-041 video
- 2026-052 videos
- 2026-073 videos
- 2026-089 videos
- 2026-098 videos
- 2026-1018 videos
Source videos
Observed neighbors
context transparency
Hidden context mutations are hard to observe; building custom TUI enables visibility.
1 source videos
Building pi in a World of Slop — Mario ZechnerConversation-level benchmarking
Both replace component-local metrics with end-to-end, user-observable outcomes (bookings, first audible audio) as the source of truth.
1 source videos
Latency Is a Budget. Humanlike Is the Goal. — Jesse Hall, LiveKitSignal layer framework
The signal layer framework requires detecting and eliminating distortion between intent and customer; an AI-native observability platform is a natural instrument to measure that distortion.
1 source videos
The Signal Layer: What to Build When Anything Can Be Built — Lena Hall, Akamaitraining data pipeline
Trace and tool-call logging is repurposed from an ops function into the data source for error analysis and fine-tuning.
1 source videos
Trading Desks to Clinical Trials: Parallels in Applied Vertical AI — Ayush Bhardwaj, Allos AITrust
Observability provides the audit trail that makes an agent's actions inspectable, which is required for enterprises to trust autonomous behavior.
1 source videos
Inside Anthropic's 200-Person Platform Team | Katelyn Lesse and Angela Jiang