---
title: "SWE-Marathon: Evaluating Coding Agents at Billion-Token Scale - Rishi Desai, Abundant AI"
category: "resources"
sourceLabels:
  - "Public YouTube metadata"
  - "YouTube transcript"
videoId: "Rx8f05JI_WA"
last_enriched: "2026-07-15T23:40:57.578830+00:00"
---
# SWE-Marathon: Evaluating Coding Agents at Billion-Token Scale - Rishi Desai, Abundant AI

## What It Is
An official AI Engineer YouTube channel video with a cached transcript. It is retained as supporting context only because it has not been verified as an actual AI Engineer World's Fair San Francisco 2026 event recording.

## Source Classification
- Source role: supporting official-channel video context, not first-class event evidence.
- Channel/source: official AI Engineer YouTube channel AI Engineer cut video.
- Use: background, speaker, company, or historical AIE context only; do not use for World's Fair San Francisco 2026 session claims unless manually verified against the official event.

## Transcript Status
Cached transcript text is available at `raw/sources/youtube-transcripts/Rx8f05JI_WA.txt` (4,329 words).

## Topic Signals
- [[coding-agents|Coding Agents]], [[agent-evaluations|Agent Evaluations]], [[ai-sandboxes|AI Sandboxes]], [[inference-engineering|Inference Engineering]]
- Transcript keywords: `simple`, `noticed`, `used`, `getting`, `asked`, `probably`, `library`, `starting`

## Link
[YouTube](https://www.youtube.com/watch?v=Rx8f05JI_WA)

## Extracted Slides
- [[youtube-Rx8f05JI_WA-slides]]