News
AI News Brief — May 18
GitHub velocity is led by affaan-m/everything-claude-code; paper attention is clustering around RewardHarness: Self-Evolving Agentic Post-Training; social attention is tilting toward Elon Musk, after contract to acquire Cursor, “plans to train Grok V9 with Cursor data”; biggest mover: rtk-ai/rtk (+3). 10 repo signals, 10 paper picks, and 10 community items made today's cut.
GitHub velocity is led by affaan-m/everything-claude-code; paper attention is clustering around RewardHarness: Self-Evolving Agentic Post-Training; social attention is tilting toward Elon Musk, after contract to acquire Cursor, “plans to train Grok V9 with Cursor data”; biggest mover: rtk-ai/rtk (+3). 10 repo signals, 10 paper picks, and 10 community items made today's cut. This fallback brief keeps the page live when the optional Gemma pass times out.
Signal Board
Repositories and papers
Keep the full repo and paper scan above the fold, then read the day as one short brief below.
Top list
Repository Momentum
Fresh GitHub projects worth scanning before the feed turns over.
affaan-m/everything-claude-code
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond. Updated 1h ago. 1…
openai/codex
Lightweight coding agent that runs in your terminal. Updated 1h ago. 83253 stars, +800/7d, created 399d ago.
NousResearch/hermes-agent
The agent that grows with you. Updated 4h ago. 154334 stars, +800/7d, created 299d ago. Up 1 spots from the previous run.
langgenius/dify
Production-ready platform for agentic workflow development. Updated 1h ago. 141664 stars, +800/7d, created 1131d ago.
ggml-org/llama.cpp
LLM inference in C/C++. Updated 1h ago. 110585 stars, +800/7d, created 1164d ago. Down 3 spots from the previous run.
rasbt/LLMs-from-scratch
Implement a ChatGPT-like LLM in PyTorch from scratch, step by step. Updated 1h ago. 94977 stars, +800/7d, created 1029d ago.
rtk-ai/rtk
CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies. Updated 1h ago. 49205 stars, +800/7d, created 115d ago. Up 3 spots from…
screenpipe/screenpipe
YC (S26) | Give AI the ability to live your experience. Records everything you do, say, hear 24/7, local, private, secure. Updated <1h ago. 18732 stars, avg 26.9/day, created 697d ago.
Eventual-Inc/Daft
High-performance data engine for AI and multimodal workloads. Process images, audio, video, and structured data at any scale. Updated <1h ago. 5478 stars, avg 3.7/day, created 1483d ago.
MemPalace/mempalace
The best-benchmarked open-source AI memory system. And it's free. Updated 1h ago. 52360 stars, +603/7d, created 43d ago.
Top list
Fresh Papers
New research worth bookmarking for a deeper read.
RewardHarness: Self-Evolving Agentic Post-Training
RewardHarness is a self-evolving framework that improves image edit evaluation by iteratively developing tools and skills from limited human demonstrations, achieving superior performance c…
FutureSim: Replaying World Events to Evaluate Adaptive Agents
FutureSim enables evaluation of AI agents' long-term predictive capabilities by simulating chronological real-world event sequences, revealing significant gaps in current forecasting perfor…
Darwin Family: MRI-Trust-Weighted Evolutionary Merging for Training-Free Scaling of Language-Mo…
The Darwin Family framework enables training-free evolutionary merging of large language models through gradient-free weight-space recombination, achieving superior reasoning performance wi…
Adaptive Teacher Exposure for Self-Distillation in LLM Reasoning
Adaptive Teacher Exposure for Self-Distillation (ATESD) improves large language model reasoning by dynamically adjusting teacher exposure during training through a learnable policy controll…
ViMU: Benchmarking Video Metaphorical Understanding
Video understanding models lack capability to interpret implicit meanings and social contexts beyond literal visual comprehension, necessitating new benchmarking approaches. Surfaced via Hu…
MemEye: A Visual-Centric Evaluation Framework for Multimodal Agent Memory
MemEye framework evaluates multimodal agent memory by measuring visual evidence granularity and retrieval usage complexity across 8 life-scenario tasks. Surfaced via HF Papers 3d…
Causal Forcing++: Scalable Few-Step Autoregressive Diffusion Distillation for Real-Time Interac…
A novel causal consistency distillation method enables efficient frame-wise video generation with reduced latency and improved quality compared to existing chunk-wise approaches. Surfaced v…
Learning to Build the Environment: Self-Evolving Reasoning RL via Verifiable Environment Synthe…
Self-improving language models construct environments for training rather than generating data, utilizing stable solve-verify asymmetry to maintain informative rewards during learning. Surf…
Beyond Individual Intelligence: Surveying Collaboration, Failure Attribution, and Self-Evolutio…
Multi-agent systems face challenges in sustained coordination and error propagation, requiring integrated approaches that enable continuous diagnosis, reorganization, and behavioral refinem…
Warp-as-History: Generalizable Camera-Controlled Video Generation from One Training Video
A novel approach called Warp-as-History enables camera-controlled video generation by transforming camera-induced warps into pseudo-history representations, achieving zero-shot capability w…
Today in AI
The day in one pass
GitHub velocity is led by affaan-m/everything-claude-code; paper attention is clustering around RewardHarness: Self-Evolving Agentic Post-Training; social attention is tilting toward Elon Musk, after contract to acquire Cursor, “plans to train Grok V9 with Cursor data”; biggest mover: rtk-ai/rtk (+3). 10 repo signals, 10 paper picks, and 10 community items made today's cut.
The quickest scan starts with affaan-m/everything-claude-code, RewardHarness: Self-Evolving Agentic Post-Training, and The Uselessness of Lava Lamps: What Random Really Means, while 10 GitHub-led signals anchor the repo side of the brief.
Use the structured digest below when you need the full wire, raw links, and the longer tail of items that did not make the front narrative.
Archive
Recent issues
Generated from the ranked feed for May 18, 2026 as one single daily issue.