The Research Desk.

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

Last Brew Time: Apr 27, 2026, 9:26 AM PT

X.com Research Buzz

Reasoning
The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
X.com
6258

The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity

Parshin Shojaee, Iman Mirzadeh, Keivan Alizadeh, Maxwell Horton, Samy Bengio, Mehrdad Farajtabar

NLP
Recursive Language Models
X.com
2723

Recursive Language Models

Alex L. Zhang, Tim Kraska, Omar Khattab

AlphaXiv Trending

Computer Vision
Image Generators are Generalist Vision Learners
AlphaXiv
200

Image Generators are Generalist Vision Learners

Valentin Gabeur, Shangbang Long, Songyou Peng

Efficiency
DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
AlphaXiv
193

DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence

DeepSeek-AI

DeepSeek-AI

Machine Learning
There Will Be a Scientific Theory of Deep Learning
AlphaXiv
94

There Will Be a Scientific Theory of Deep Learning

Jamie Simon, Daniel Kunin, Alexander Atanasov

Scaling Self-Play with Self-Guidance
AlphaXiv
73

Scaling Self-Play with Self-Guidance

Luke Bailey, Kaiyue Wen, Kefan Dong

NLP
Hyperloop Transformers
AlphaXiv
71

Hyperloop Transformers

Abbas Zeitoun, Lucas Torroba-Hennigen, Yoon Kim

Context Unrolling in Omni Models
AlphaXiv
43

Context Unrolling in Omni Models

Ceyuan Yang, Zhijie Lin, Yang Zhao

HuggingFace Daily Papers

Computer Vision
Video Analysis and Generation via a Semantic Progress Function
HuggingFace
32

Video Analysis and Generation via a Semantic Progress Function

Gal Metzer, Sagi Polaczek, Ali Mahdavi-Amiri, Raja Giryes, Daniel Cohen-Or

Reasoning
Contexts are Never Long Enough: Structured Reasoning for Scalable Question Answering over Long Document Sets
HuggingFace
7

Contexts are Never Long Enough: Structured Reasoning for Scalable Question Answering over Long Document Sets

Harshit Joshi, Priyank Shethia, Jadelynn Dao, Monica S. Lam

Computer Vision
Building a Precise Video Language with Human-AI Oversight
HuggingFace
3

Building a Precise Video Language with Human-AI Oversight

Zhiqiu Lin, Chancharik Mitra, Siyuan Cen, Isaac Li, Yuhan Huang

Reinforcement Learning
Memanto: Typed Semantic Memory with Information-Theoretic Retrieval for Long-Horizon Agents
HuggingFace
1

Memanto: Typed Semantic Memory with Information-Theoretic Retrieval for Long-Horizon Agents

Seyed Moein Abtahi, Rasa Rahnema, Hetkumar Patel, Neel Patel, Majid Fekri

Retrieval
DiagramBank: A Large-scale Dataset of Diagram Design Exemplars with Paper Metadata for Retrieval-Augmented Generation
HuggingFace
1

DiagramBank: A Large-scale Dataset of Diagram Design Exemplars with Paper Metadata for Retrieval-Augmented Generation

Tingwen Zhang, Ling Yue, Zhen Xu, Shaowu Pan

Machine Learning
EmbodiedMidtrain: Bridging the Gap between Vision-Language Models and Vision-Language-Action Models via Mid-training
HuggingFace
1

EmbodiedMidtrain: Bridging the Gap between Vision-Language Models and Vision-Language-Action Models via Mid-training

Yiyang Du, Zhanqiu Guo, Xin Ye, Liu Ren, Chenyan Xiong