AI makes AI better.

A daily hub for work where AI is part of the loop that improves AI, each entry with its loop spelled out. Read more →

1/4

A loop in motion

Project · Google DeepMind

Gemini models propose code changes

From the labalphaBell Qwen Open27B post-training research, with capability tradeoffs measured.Explore the model → From the labModel ObservatoryMeasured Flint and Qwen traces with published decoder maps.Open the lab tool → From the labMiniMax Video BudgetVisualize sampled frames, two-frame groups and vision placeholders.Try the inspector → From the labMOSS Blind Layout CheckCompare exact text and placement in two local images.Try the blind check → From the labAgent Request Evidence LadderSee what a trace proves about an agent's request and its outcome.Explore the cases →

Today's radar

12 Nov 2026Full radar →

7 picked from 17 candidates · 31 sources read · 76 loops indexed

01
Loop of the dayPaperarXiv11 Nov 2026Research automation
Autonomous Optimization of Numerical Algorithms by an AI Agent: Application to Atomistic Simulation
The loop: An AI agent rewrites and redesigns numerical algorithms against a pre-specified metric in a guarded loop. The resulting efficiency gains improve the underlying mathematical tools the agent uses for optimization.
Optimizing the fundamental numerical algorithms that underpin scientific computing is a powerful niche loop, as these same algorithms are often the bottleneck in the agent's own search and evaluation processes.
loop fit 9/10Jeheon Woovia arXiv
02
LeaderboardEpoch AI12 Nov 2026AI R&D evals
New best on PostTrainBench: Fable 5.1 (Max) · Claude Code (44.6%)
The loop: AI agents perform post-training on a base model to improve its benchmark average. The resulting model is evaluated on a benchmark specifically designed to measure AI driven engineering tasks.
loop fit 10/10via Epoch AI Benchmarking Hub (CC BY)
03
PaperarXiv4 Nov 2026Training data
Micius-Flash: Lightweight Agentic Intelligence for Engineering
The loop: Micius-Flash generates synthetic data and reasoning sequences for hardware coding and semiconductor design. These outputs are used in a multi-teacher on-policy distillation process to train a unified successor model.
loop fit 9/10Hongcheng Gao, Weiheng Li et al.via Hugging Face Papers
04
Forum postLessWrong11 Nov 2026Interpretability and oversight
The Chain Lens [interim]
The loop: The chain lens extracts multi-token concepts from a model's latent representations using contrastive decoding. This surfaces specific computational intermediates that can be used to oversee and refine the model's reasoning.
loop fit 6/10vgelvia LessWrong
05
PaperarXiv8 Nov 2026Tools and harness
InfraRSI: Harnessing Recursive Self-Improvement for Training Infrastructure
The loop: The InfraRSI agent co-optimizes launch knobs and code patches for distributed LLM training. This improves the underlying training infrastructure the agent relies on.
loop fit 9/10InfraRSI Team, Yuchuan Tian et al.via arXiv
06
Forum postLessWrong11 Nov 2026Research automation
How I'm thinking about different types of automated alignment
The loop: An AI system automates empirical alignment experiments, producing data and insights that are used to make the next generation of AI systems more aligned.
loop fit 6/10owencbvia LessWrong
07
PaperarXiv9 Nov 2026Kernels and compilers
Correct First, Fast Second: What a Local 7B Code Model Can and Cannot Do for GPU Kernel Optimization
The loop: A multi-agent loop generates, evaluates, and refines custom CUDA kernels to replace PyTorch operators. These kernels improve the execution speed of the model on its own hardware.
loop fit 8/10Jui-Chien Tsou, Po-Yi Wuvia arXiv
08
PaperarXiv7 Nov 2026Training data
No Second Run: Does RSI Stop at Frontier Pre-training?
The loop: The system decomposes frontier pre-training into smaller loops that repair broken steps, which improves the data pipelines and infrastructure, making the resulting foundation model better.
loop fit 9/10Xiaojie Cai, Jiayi Chen et al.via arXiv