Alphabell.

Radar archive

Every daily edition.

DateTop itemItems
16 Nov 2026ModHarness-RL: Recasting Harnesses as Learnable and Evolving Programs for Agentic RL6
15 Nov 2026MLIRBench: A Benchmark and Dataset for Compiler and Hardware-Aware Reasoning5
14 Nov 2026Experience-Handoff Reinforcement Learning for Recursive Self-Improvement6
13 Nov 2026Singularity-Bench: Can Agents Build Agents?7
12 Nov 2026New best on PostTrainBench: Fable 5.1 (Max) · Claude Code (44.6%)7
11 Nov 2026No Second Run: Does RSI Stop at Frontier Pre-training?7
10 Nov 2026Where is all the AI safety research uplift?7
9 Nov 2026Ajeya: AI is dramatically accelerating AI research right now12
8 Nov 2026ServerBench: How can your coding agents design specialized LLM serving systems?12
7 Nov 2026Independent Assessment of AI R&D Acceleration at Anthropic in October 202612
6 Nov 2026When Can Self-Play Fix a False Belief? Answer Keys, Pass Bands and Replacement Rules in an Execution-Grounded Loop7
5 Nov 2026EvoOptiGraph: Weakness-Driven Coevolution via Graph-Based Structural Generation for Optimization Modeling9
4 Nov 2026Think Less, Locally: Local Preference Learning for Reasoning Compression8
3 Nov 2026Exo: An Architecture for Self-Improving Systems at Runtime11
2 Nov 2026RoboArchitect: Automating the Design of Robotic Agents with Persistent Memory5
1 Nov 2026Search-Guided Reasoning Trace Curation for Verilog Generation8
31 Oct 2026InfraDev: Can Coding Agents Develop AI Infrastructure, and How Can We Tell?8
30 Oct 2026A-RSI: Adaptive Task Refresh for Zero-Data Multimodal Self-Improvement6
29 Oct 2026Build, Don't Play: Self-Improving LLM Research Agents Develop a Hearthstone Battlegrounds AI That Rivals Top Players12
28 Oct 2026How Much Does the Hint Matter? Scope Conditions for Hindsight Self-Distillation in Financial Reasoning12
27 Oct 2026Infinity Machine: a self-improving AI agent for research8
26 Oct 2026RSI-Opt: Recursive Self-Improvement Optimizer Agents For Model Training7
25 Oct 2026Agentic Knowledge-Model Co-Evolution via Consistency Regularization7
24 Oct 2026VeriEvolve: Evaluator-First Evolution for Tool-Using Agents8
23 Oct 2026SplatLab: Autonomous Scientific Discovery for 3DGS Research8
22 Oct 2026Recursive Self-Improvement on a Budget: Scoring Code Edits Before They Run7
21 Oct 2026Training the Whole Agent, Not Just the Model7
20 Oct 2026VibeTraining-Bench: Data-Trainer-Harness Co-Design for Iterative System Improvement8
19 Oct 2026Self-Evolving Coding Agents via Harness-Policy Optimization4
18 Oct 2026AdaRSI: On-Demand Recursive Self-Improvement for Embodied Agents6
17 Oct 2026AutoEnvScaling: Automating the Data Flywheel with Terminal Agents6
16 Oct 2026A Self-Improving Agent that Grows its Own DSL: Oracle-Free Operator Induction for Exact Question Answering9
15 Oct 2026CARD: Closing the Teacher-Student Loop for Adversarial Robustness Distillation6
14 Oct 2026Cortex: Measurement-Gated Evolution and Pruning of Coding-Agent Skills and Rules8
13 Oct 2026Sage6
12 Oct 2026Recursive Self-Improvement through Collective Intelligence6
11 Oct 2026Recurrent Self-Improvement: Dynamic Cross-Loop On-Policy Distillation for Looped Language Models6
10 Oct 2026AdvSim2Real: Training Web Agents Against Adaptive Prompt Injection in a Web World Model5
9 Oct 2026The Red Queen Gödel Machine: Co-Evolving Agents and Their Evaluators9
8 Oct 2026FreeEvolve: Learning to Evolve Beyond Fixed Loops6
7 Oct 2026RRSI: Regularized Recursive Self-Improvement of Agent Harnesses5
6 Oct 2026SEIS: Self-Evolving Inference Systems6
5 Oct 2026Recursive Self-Improvement in Unified Multimodal Models7
4 Oct 2026Rep2Skill: Representation-Guided Skill Self-Evolution for LLM Agents7
3 Oct 2026MILO: Automated Harness Discovery via Orchestrated Multi-Agent Evolution8
2 Oct 2026AREX-2: Advancing Self-Improving Agents through Long-Horizon Reflective Tasks8