The Research Desk.

The most upvoted and starred AI research crossing the community today.

Last Brew Time: Jun 20, 2026, 2:53 PM PT

X.com Research Buzz

Reinforcement Learning
Ponytail: Makes your AI agent think like the laziest senior dev in the room
X.com
17488

Ponytail: Makes your AI agent think like the laziest senior dev in the room

DietrichGebert

Computer Vision
Palmier: open-source AI-native video editor for Claude
X.com
14039

Palmier: open-source AI-native video editor for Claude

AlphaXiv Trending

AlphaXiv
290

GLM-5.2: Built for Long-Horizon Tasks

Z.ai

Z.ai

Computer Vision
You Don't Need Strong Assumptions: Visual Representation Learning via Temporal Differences
AlphaXiv
146

You Don't Need Strong Assumptions: Visual Representation Learning via Temporal Differences

Ninad Daithankar, Alexi Gladstone, Yann LeCun, Heng Ji

New York University

Looped World Models
AlphaXiv
108

Looped World Models

Hongyuan Adam Lu, Z.L. Victor Wei, Qun Zhang

NLP
VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models
AlphaXiv
102

VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models

Sen Xu, Shixi Liu, Wei Wang

NLP
Variable-Width Transformers
AlphaXiv
71

Variable-Width Transformers

Zhaofeng Wu, Oliver Sieberling, Shawn Tan, The paper introduces ">

MIT-IBM Watson AI Lab

HuggingFace Daily Papers

NLP
Context-Aware RL for Agentic and Multimodal LLMs
HuggingFace
9

Context-Aware RL for Agentic and Multimodal LLMs

Peiyang Xu, Bangzheng Li, Sijia Liu, Karthik R. Narasimhan, Pramod Viswanath

Princeton University

Freeing the Law with LOCUS: A Local Ordinance Corpus for the United States
HuggingFace
6

Freeing the Law with LOCUS: A Local Ordinance Corpus for the United States

Denis Peskoff, Joe Barrow, Christopher Vu, Diag Davenport

LOCUS

Reinforcement Learning
LedgerAgent: Structured State for Policy-Adherent Tool-Calling Agents
HuggingFace
6

LedgerAgent: Structured State for Policy-Adherent Tool-Calling Agents

Md Nayem Uddin, Amir Saeidi, Eduardo Blanco, Chitta Baral

Arizona State University

Reinforcement Learning
LegalHalluLens: Typed Hallucination Auditing and Calibrated Multi-Agent Debate for Trustworthy Legal AI
HuggingFace
3

LegalHalluLens: Typed Hallucination Auditing and Calibrated Multi-Agent Debate for Trustworthy Legal AI

Lalit Yadav, Akshaj Gurugubelli

Independent Research

Reinforcement Learning
Configurable Clinical Information Extraction with Agentic RAG: What Works, What Breaks, and Why
HuggingFace
3

Configurable Clinical Information Extraction with Agentic RAG: What Works, What Breaks, and Why

Osman Alperen Çinar-Koraş, Marie Bauer, Sameh Khattab, Merlin Engelke, Moon Kim

IKIM

Mrseongminkim
ReSyn: A Generalized Recursive Regular Expression Synthesis Framework
HuggingFace
1

ReSyn: A Generalized Recursive Regular Expression Synthesis Framework

Seongmin Kim, Hyunjoon Cheon, Su-Hyeon Kim, Yo-Sub Han, Sang-Ki Ko