Markdown source

How long can your skills be before your agent forgets what you told it?

Conference Context

Session Description

A year ago, frontier models lost the thread somewhere around 200 simultaneous instructions, so skills files had to stay short and lean on sub-skills and subagents. We re-ran IFScale on the 2026 frontier and found the ceiling has moved by an order of magnitude: closer to 2,000 instructions, up to 5,000 on the strongest models. The more interesting story is how models fail at the new frontier: DeepSeek quietly drops instructions, Opus refuses outright when innocuous words trip a safety classifier, Gemini burns its whole budget on reasoning and emits nothing, and GPT-5.5 stops to tell you your request was unreasonable. The capacity problem is largely solved; verification is wide open. We'll show the data, the failure modes, and what it costs to find out. You’ll come out with hard data on the ceiling for complex instructions to LLMs

Media Evidence

Ship Real Agents: Hands-On Evals for Agentic Applications — Laurie Voss, Arize (speaker-match related prior/adjacent AI Engineer video; captions: English auto-captions).

Evidence Graph

This evidence graph is generated from currently linked source material: official schedule text, related video pages, cached transcripts, visible slide text, dense/reconstructed slide pages, and AI slide-classification audits.

Media Signals

Agent Reading Notes

Use these signals to refine the synopsis, topic links, people/company context, and method notes. If a source is a related external video rather than an exact official recording, keep it framed as supporting evidence.

Transcript Status

Related video transcript availability: English auto-captions. Treat this as supporting context, not a recording of this exact scheduled session unless later confirmed. Cached at raw/sources/youtube-transcripts/Xfl50508LZM.txt (22,591 words).

People

Supporting Slides

Slide Evidence

Synthesis

Synthesized Breakdown

Hi everybody. Uh my name's Laurie Voss. I am head of developer experience at Arize AI. Uh in a former life, I co-founded npm Inc.

Speaker And Company Context

Topics Covered

Derived Links And Source Material

Novel Concepts / Clever Methods

Evidence Boundary

This synthesis uses the official schedule plus cached video transcripts. Official AI Engineer World's Fair San Francisco 2026 livestreams and cut videos are primary event video sources for transcript/slide evidence; external, historical, or speaker-matched videos remain supporting context unless manually verified as exact official event recordings.