AISystem

MASS

Sakana AI and Berkeley approach that uses a model to solve, optimize and judge runs before fine-tuning on selected results.

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to October 12, 2026 (theinnermostloop.substack.com)
Source linked by the October 12 issue (pub.sakana.ai)

Sakana AI and Berkeley approach that uses a model to solve, optimize and judge runs before fine-tuning on selected results. This description follows the October 12 newsletter and its linked source.

Visit resource ↗

Development timeline

Newest reporting dates first. Reporting dates may differ from event dates.

2026-10-12: MASS uses one model to solve, improve and judge its work

Knowledge graphopen in browser →

Developed by

Cite or correct this record

Citation formatCitation text

MASS
Innermost Loop KB — secondary record
Reporting date: unknown
KB permalink: https://nicholsn.github.io/innermost-loop-kb/systems/mass
Attributed sources:
- Welcome to October 12, 2026 — https://theinnermostloop.substack.com/p/welcome-to-october-12-2026
- Source linked by the October 12 issue — https://pub.sakana.ai/mass/
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.