Uploaded July 2026 | Updated November 2026, 17 hours ago
Hundreds of traces and they all look fine? π€ Here's how to find the ones that are quietly failing.
Watch this video to learn:
How to spot silent agent failures β where the span errors but the model replies normally
Why a friendly, well-formatted reply can still be a bug
How to leave notes on high-signal traces before you commit to categories
Chapters
00:00 Finding the traces that matter
00:08 Opening PXI
00:24 Sending traces to the Phoenix team
00:43 Picking your model
01:06 Hundreds of traces, all look fine
02:04 Asking PXI to find silent failures
02:38 8 of 11 errored traces replied normally
03:15 A confirmed order that never happened
04:01 What a high-signal trace is
04:10 Leaving a note (notes vs. annotations)
04:52 When the database dies and the agent invents products
05:33 Notes now, annotations next
π Try Phoenix: arize.com/phoenix/?utm_source=youtube&utm_medium=video&utm_content=rnabors
π Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
#ArizeAI #LLMEvaluation #AgentObservability #Phoenix
Hundreds of traces and they all look fine? π€ Here's how to find the ones that are quietly failing.
Watch this video to learn:
How to spot silent agent failures β where the span errors but the model replies normally
Why a friendly, well-formatted reply can still be a bug
How to leave notes on high-signal traces before you commit to categories
Chapters
00:00 Finding the traces that matter
00:08 Opening PXI
00:24 Sending traces to the Phoenix team
00:43 Picking your model
01:06 Hundreds of traces, all look fine
02:04 Asking PXI to find silent failures
02:38 8 of 11 errored traces replied normally
03:15 A confirmed order that never happened
04:01 What a high-signal trace is
04:10 Leaving a note (notes vs. annotations)
04:52 When the database dies and the agent invents products
05:33 Notes now, annotations next
π Try Phoenix: arize.com/phoenix/?utm_source=youtube&utm_medium=video&utm_content=rnabors
π Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
#ArizeAI #LLMEvaluation #AgentObservability #Phoenix








![NVIDIAs Peter Belcak Explains Why SLMs are the Future of Agentic AI
In our most recent AI research paper community reading, we had the privilege of hosting Peter Belcak β an AI Researcher working on the reliability and efficiency of agentic systems at NVIDIA β who walked us through his new paper making the rounds in AI circles titled βSmall Language Models are the Future of Agentic AI.β
The paper posits that small language models (SLMs) are sufficiently powerful, inherently more suitable, and necessarily more economical for many invocations in agentic systems, and are therefore the future of agentic AI. The authorsβ argumentation is grounded in the current level of capabilities exhibited by SLMs, the common architectures of agentic systems, and the economy of LM deployment. The authors further argue that in situations where general-purpose conversational abilities are essential, heterogeneous agentic systems (i.e., agents invoking multiple different models) are the natural choice. They discuss the potential barriers for the adoption of SLMs in agentic systems and outline a general LLM-to-SLM agent conversion algorithm.
Learn more about the paper and get a TL;DR: https://arize.com/blog/nvidias-small-language-models-are-the-future-of-agentic-ai-paper/
Ask questions about the research: agents [at] nvidia.com NVIDIAs Peter Belcak Explains Why SLMs are the Future of Agentic AI](https://i.ytimg.com/vi/9xgRTznP21E/mqdefault.jpg)

