Alphabell.

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
AlphaEvolve
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

18 Nov 2026Full radar →

6 picked from 120 candidates · 30 sources read · 78 loops indexed

  1. 01
    Loop of the dayPaperarXiv18 Nov 2026Self-modifying agents

    Language Models Can Rewrite Their Own Weights

    The loop: A 27B language model diagnoses its own failures and directly edits its own weights without gradients or training. These edited weights are then used by the model to achieve higher accuracy on subsequent unseen problems.

    Direct weight editing by the model itself bypasses the need for external optimizers or training runs, opening a new and highly direct path to recursive self-improvement.

    loop fit 10/10Jiajun Fan, Neil He, Changyi Yang, Andy Mao, Xi Zhu, Chaoran Cheng, Kaiqu...via arXiv
    Language Models Can Rewrite Their Own Weights
  2. 02
    PaperarXiv18 Nov 2026Research automation

    RSI-Prime: Training Models to Train Models

    The loop: A GLM-5.3 model trains on a dataset of experimental trajectories to improve its ability to make high-quality model training decisions. This improved capability allows the model to autonomously conduct post-training and build better successor models.

    loop fit 9/10Dewu Zheng, Yang Ye, Hongyu Zhang, Jun Yu, Yanlin Wangvia arXiv
    RSI-Prime: Training Models to Train Models
  3. 03
    Articlestarsling.dev17 Nov 2026Tools and harness

    Launch HN: StarSling (YC P25), Self-improving CI for agent-volume repos

    The loop: StarSling AI agents analyze continuous integration workflows to generate and merge pull requests that resolve performance bottlenecks. These merged pull requests make the continuous integration harness faster and cheaper for all subsequent agent runs.

    loop fit 8/10yonasbvia Hacker News
    Launch HN: StarSling (YC P25), Self-improving CI for agent-volume repos
  4. 04
    Blog postTogether AI16 Nov 2026Kernels and compilers

    Last year we built megakernels. This year we deleted them: The road to Together Compiler

    The loop: AI agents rewrite and fuse GPU kernels for a model's forward pass to optimize execution. These optimized kernels improve the inference speed of the models, which allows the agents themselves to run faster.

    loop fit 7/10via Together AI
    Last year we built megakernels. This year we deleted them: The road to Together Compiler
  5. 05
    ProjectGitHub14 Nov 2026Tools and harness

    oriliz/agent-crop

    The loop: An AI programming assistant records its own local timing and performance data to audit its time usage. The assistant then uses this evidence to rewrite its own instructions and tools, improving its efficiency on future tasks.

    loop fit 7/10orilizvia GitHub
    oriliz/agent-crop
  6. 06
    PaperarXiv18 Nov 2026Training data

    From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop

    The loop: The RepoFactory pipeline uses an environment evolution loop to optimize synthetic training environments based on proxy feedback. These improved environments are then used to train a successor coding model that is better at long-horizon tasks.

    loop fit 9/10Yang Tian, Tianyu Zheng, Xiyuan Yang, Yaxin Du, Chengran Yang, Bo Zhao, Jian...via arXiv
    From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop
  7. 07
    PaperarXiv14 Nov 2026Kernels and compilers

    CompoundingKernel: Self-Improving from Zero to First

    The loop: CompoundingKernel agents search for and write optimized GPU kernels. These kernels improve the execution speed of the hardware running the agents, making the search process more efficient.

    loop fit 9/10Qian Qu, Hongyang Wang et al.via Hugging Face Papers
    CompoundingKernel: Self-Improving from Zero to First
  8. 08
    PaperarXiv12 Nov 2026Tools and harness

    Can You Trust Your Finance Agent? Drift, Diagnosis and Auditable Intervention on One Shipped Financial Research Agent

    The loop: An optimization loop reads agent failure traces and makes targeted changes to the agent's instructions. These changes improve the agent's performance and reduce its operating costs.

    loop fit 9/10Wenhong Zhu, Xuefeng Li et al.via Hugging Face Papers
    Can You Trust Your Finance Agent? Drift, Diagnosis and Auditable Intervention on One Shipped Financial Research Agent
AlphaEvolve
ProjectGoogle DeepMind
AlphaEvolve
  1. Gemini models propose code changes
  2. Evaluators keep best scoring versions
  3. Finds better Gemini training kernels
  4. Lowers cost to train next Gemini models
  5. repeat
Kernels and compilersGoogle DeepMind · 2025
Darwin Gödel Machine
PaperarXiv
Darwin Gödel Machine
  1. Coding agent edits its own code
  2. Scored child agents enter the archive
  3. Archived agents make later self modifications
  4. repeat
Self-modifying agentsUBC, Vector Institute, Sakana AI · 2025
AlphaChip
ProjectGoogle DeepMind
AlphaChip
  1. RL agent lays out TPU blocks
  2. Better layouts make AI hardware cheaper
  3. Agent pre-trains on earlier chip generations
  4. repeat
Hardware and chipsGoogle DeepMind, Google Research · 2020
autoresearch
Mini-projectGitHub
autoresearch
  1. Coding agent edits LLM training script
  2. Runs a 5-minute training job
  3. Keeps change if validation improves
  4. Changes transferred to larger models
  5. repeat
Research automationAndrej Karpathy · 2026
ScholarEvolve
PaperarXiv
ScholarEvolve
  1. Research agent writes harness mutation blueprints
  2. Coding agent implements harness module changes
  3. Winning harness becomes the new champion
  4. New champion harness runs further generations
  5. repeat
Tools and harnessYang et al. (UC Santa Barbara, Microsoft) · 2026
Nemotron-4 340B
PaperarXiv
Nemotron-4 340B
  1. Aligned model generates training data
  2. Reward model judges and filters it
  3. Data trains next intermediate instruct model
  4. Next model generates better alignment data
  5. repeat
Training dataNVIDIA · 2024
Self-Rewarding LMs
PaperarXiv
Self-Rewarding LMs
  1. Model generates candidate responses to prompts
  2. Model scores them itself as a judge
  3. Scores become preference pairs for DPO
  4. Trained model becomes next generator and judge
  5. repeat
Self-reward and self-playMeta, NYU (Yuan et al.) · 2024
Claude Code
ProjectAnthropic
Claude Code
  1. Claude Code writes its own code changes
  2. Changes become next Claude Code versions
  3. Release becomes harness for next development round
  4. repeat
Self-modifying agentsAnthropic · 2025
KernelEvolve
ProjectarXiv
KernelEvolve
  1. LLM agents write and benchmark kernels
  2. Kernels speed up AI model training
  3. Session data post-trains smaller kernel models
  4. repeat
Kernels and compilersMeta · 2025
PostTrainBench
BenchmarkarXiv
PostTrainBench
  1. Coding agent writes fine-tuning code
  2. Agent post-trains a base LLM
  3. Trained model scored on target benchmark
  4. One turn of model training another
  5. repeat
AI R&D evalsELLIS Institute Tubingen, MPI-IS, University of Tubingen, Thoughtful Lab (Rank et al.) · 2026
Self-training learned optimizers
FoundationarXiv
Self-training learned optimizers
  1. Learned optimizers update other learned optimizers
  2. Training keeps best performing optimizer parameters
  3. Better optimizers train each other faster
  4. repeat
Architecture and optimizer searchMetz et al. (Google Research, Brain Team) · 2021
CoT monitoring
PaperarXiv
CoT monitoring
  1. LLM reads reasoning model thoughts
  2. LLM flags reward hacking behavior
  3. Signal feeds back into training reward
  4. Trains more aligned capable agents
  5. repeat
Interpretability and oversightOpenAI (Baker et al.) · 2025

Support for loop research

How to apply →

Have a loop worth building?

Tell us what the AI improves and how the improvement comes back. Applications for grants, fellowships and compute all start on the contact page.

Get in touch