Outerloop 0.4.0: autonomous research agents on Slurm, −8.8% on gpt-speedrun in 6 days
mengyer · x · 2026-10-21
Outerloop 0.4.0, the cloud release, runs autonomous research agents on a cloud Slurm cluster (tested on Google Cloud), with research taste as a configurable guidance file. Workflow: agents propose hypotheses and code changes → experiments run in parallel on your cluster → a benchmark scores candidates against base at the same seed → two reviewers verify before a PR opens; every attempt, negatives included, is recorded. On gpt-speedrun, 4 agents spent 770 GPU-hours over six days with 82 attempts, cutting steps from 9472 to 8640 (−8.8%) with zero human edits.
More from coding & agent
- Claude Code Pushes Urgent Security Update Across All Installations — umeshonai · 2026-10-23
- Claude cut a 2.6-hour video export to 20 minutes — but only after being told to rebuild the server — byhewar · 2026-10-23
- 3 flight-booking rules my AI agent checks that I'd never do by hand — Efistoffeles · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23