AI makes AI better.

6 picked from 127 candidates · 31 sources read · 59 loops indexed

1. 01
   Loop of the dayPaperarXiv8 Oct 2026Research automation

   FreeEvolve: Learning to Evolve Beyond Fixed Loops

   The loop: The FREEEVOLVE agent automates the design of workflows and evaluation loops for language model agents. It improves its own evolution skill through meta-evolution by scoring each candidate skill on the fresh target agent it produces.

   FreeEvolve takes a significant step by automating not just the agent's prompts or skills, but the entire evolutionary search loop used to discover them.

   loop fit 10/10Lecheng Kong, Like Hui, Nikos Kanakaris, Prithwish Jana, Sahika Genc,...via arXiv
   
2. 02
   PaperarXiv6 Oct 2026Tools and harness

   PhysEvo: Astra Can Act, Let It

   The loop: A meta-agent diagnoses failures in robot manipulation tasks to revise its tools and skills. It also improves its own diagnostic tools, ensuring that retained revisions support both later action and subsequent self-improvement.

   loop fit 9/10Wenqing Tian, Zeyu Zhang et al.via arXiv
   
3. 03
   Blog postEpoch AI7 Oct 2026AI R&D evals

   EBR-bench update

   The loop: EBR-bench measures the ability of models to learn from their own experience. By evaluating how well models adapt, the benchmark provides a metric that tracks and guides the development of self-improving AI systems.

   loop fit 8/10via Epoch AI
   
4. 04
   ProjectGitHub24 Jul 2026Self-modifying agents

   Birfy/agentdescent

   The loop: The agentdescent framework treats agent parameters like skills and harnesses as optimizable components. This allows an optimizer to compute diffs as gradients and update the agents, enabling them to evolve their own components to improve performance.

   loop fit 6/10Birfyvia GitHub
   
5. 05
   PaperarXiv7 Oct 2026Training data

   FrogNano: Training a 4B Coding Agent via Online Task Synthesis

   The loop: An online task synthesis pipeline creates coding tasks calibrated to a 4B agent's current learnability frontier. The agent trains on these synthetic tasks via reinforcement learning, which shifts its frontier and prompts the pipeline to generate harder tasks for the next round.

   loop fit 9/10Minseon Kim, Zhengyan Shi, Emiliano Penaloza, Christopher Cui, Roger Creus...via arXiv
   
6. 06
   PaperarXivTools and harness

   AutoRef: Harness Optimization for Agentic Multi-Reference Image Generation

   The loop: AutoRef automates the optimization of evaluation harnesses for image generation agents, improving the accuracy and efficiency of model development.

   loop fit 9/10via Awesome-AI4AI
   
7. 07
   PaperarXiv6 Oct 2026Tools and harness

   RRSI: Regularized Recursive Self-Improvement of Agent Harnesses

   The loop: An LLM agent system iteratively proposes and selects edits to its own harness using a regularized proposer and critic. The resulting improved harness directly upgrades the agent's operating environment, making it better at solving tasks and further improving its harness.

   loop fit 9/10Peng Xia, Rujun Han, Zifeng Wang, Yanfei Chen, Yufan Zhuang, Yoonho Lee,...via arXiv
   
8. 08
   PaperarXiv5 Oct 2026Tools and harness

   Second-Order Problem Solving for Recursive Self-Improvement in Formal Verification

   The loop: An agent framework monitors its own execution traces to diagnose structural failures and systematically edit its workflow. These edits improve the agent's workflow for formal verification, making it more effective in subsequent rounds of recursive self-improvement.

   loop fit 9/10Yuxuan Jiang, Aditya Vempaty et al.via arXiv