AllLLMCoding Co-pilotPersonal AssistantsAI AgentsHealthcare AIEthics & RegulationResearch & PapersStartups & Funding
Company:AllOpenAIAnthropicGoogleMicrosoftAppleDeepSeekMistralxAIMetaOthers

Research & Papers

Latest AI research

Nous Research Proposes Lighthouse Attention: A Training-Only Selection-Based Hierarchical Attention That Delivers 1.4–1.7× Pretraining Speedup at Long Context
Research & Papers

Nous Research Proposes Lighthouse Attention: A Training-Only Selection-Based Hierarchical Attention That Delivers 1.4–1.7× Pretraining Speedup at Long Context

Nous Research introduces Lighthouse Attention, a hierarchical attention mechanism that reduces computational complexity...

MarkTechPost
New benchmark confirms AI video generators look stunning but still can't reason about the world
Research & Papers

New benchmark confirms AI video generators look stunning but still can't reason about the world

WorldReasonBench, a new benchmark, evaluates AI video generators on physical and logical plausibility rather than image...

The Decoder
Researchers train AI model that hits near-full performance with just 12.5 percent of its experts
Research & Papers

Researchers train AI model that hits near-full performance with just 12.5 percent of its experts

Researchers at Allen Institute for AI and UC Berkeley developed EMO, a mixture-of-experts model that organizes experts...

The Decoder
NVIDIA Introduces SANA-WM: A 2.6B-Parameter Open-Source World Model That Generates Minute-Scale 720p Video on a Single GPU
Research & Papers

NVIDIA Introduces SANA-WM: A 2.6B-Parameter Open-Source World Model That Generates Minute-Scale 720p Video on a Single GPU

NVIDIA introduces SANA-WM, a 2.6B-parameter open-source world model capable of generating 60-second 720p videos with...

MarkTechPost
AI research papers are getting better, and it’s a big problem for scientists
Research & Papers

AI research papers are getting better, and it’s a big problem for scientists

AI-generated research papers are improving in quality, creating a problem for scientists as these papers are being...

The Verge AI
A Coding Implementation to Master GPU Computing with CuPy, Custom CUDA Kernels, Streams, Sparse Matrices, and Profiling
Research & Papers

A Coding Implementation to Master GPU Computing with CuPy, Custom CUDA Kernels, Streams, Sparse Matrices, and Profiling

This tutorial explores CuPy as a GPU-accelerated alternative to NumPy for high-performance numerical computing in...

MarkTechPost
Two from MIT named 2026 Knight-Hennessy Scholars
Research & Papers

Two from MIT named 2026 Knight-Hennessy Scholars

The prestigious fellowship funds graduate studies at Stanford University.

MIT News AI
Nous Research Releases Token Superposition Training to Speed Up LLM Pre-Training by Up to 2.5x Across 270M to 10B Parameter Models
Research & Papers

Nous Research Releases Token Superposition Training to Speed Up LLM Pre-Training by Up to 2.5x Across 270M to 10B Parameter Models

Nous Research introduces Token Superposition Training (TST), a two-phase pre-training method that reduces wall-clock...

MarkTechPost
Bosch, Researchers Develop AI for Humanoid Dexterity
Research & Papers

Bosch, Researchers Develop AI for Humanoid Dexterity

Bosch and research teams have developed a "touch dreaming" AI system that significantly improves humanoid robot...

AI Business
Mira Murati’s Thinking Machines Lab Introduces Interaction Models: A Native Multimodal Architecture for Real-Time Human-AI Collaboration
Research & Papers

Mira Murati’s Thinking Machines Lab Introduces Interaction Models: A Native Multimodal Architecture for Real-Time Human-AI Collaboration

Thinking Machines Lab introduced TML-Interaction-Small, a 276B parameter multimodal model that processes audio, video,...

MarkTechPost
Universal AI is “a pathway to AI fluency that’s accessible and approachable to anyone, anywhere”
Research & Papers

Universal AI is “a pathway to AI fluency that’s accessible and approachable to anyone, anywhere”

MIT Open Learning has launched Universal AI, a new educational program designed to make AI accessible to learners...

MIT News AI
Q&A: Expanding MIT’s global reach through Universal Learning
Research & Papers

Q&A: Expanding MIT’s global reach through Universal Learning

Dimitris Bertsimas and Megan Mitchell discuss the motivation behind Universal Learning, and what sets the new MIT Open...

MIT News AI
Tilde Research Introduces Aurora: A Leverage-Aware Optimizer That Fixes a Hidden Neuron Death Problem in Muon
Research & Papers

Tilde Research Introduces Aurora: A Leverage-Aware Optimizer That Fixes a Hidden Neuron Death Problem in Muon

Tilde Research has introduced Aurora, a new neural network optimizer that fixes a critical flaw in the Muon optimizer...

MarkTechPost
What Parameter Golf taught us about AI-assisted research
Research & Papers

What Parameter Golf taught us about AI-assisted research

Parameter Golf was a large-scale research event with over 1,000 participants exploring AI-assisted machine learning...

OpenAI Blog
Meta and Stanford Researchers Propose Fast Byte Latent Transformer That Reduces Inference Memory Bandwidth by Over 50% Without Tokenization
Research & PapersMeta

Meta and Stanford Researchers Propose Fast Byte Latent Transformer That Reduces Inference Memory Bandwidth by Over 50% Without Tokenization

Meta FAIR and Stanford researchers have developed three inference methods for the Byte Latent Transformer that reduce...

MarkTechPost
Sakana AI and NVIDIA Introduce TwELL with CUDA Kernels for 20.5% Inference and 21.9% Training Speedup in LLMs
Research & Papers

Sakana AI and NVIDIA Introduce TwELL with CUDA Kernels for 20.5% Inference and 21.9% Training Speedup in LLMs

Sakana AI and NVIDIA researchers demonstrated that L1 regularization can achieve over 99% sparsity in LLM feedforward...

MarkTechPost
Best Vector Databases in 2026: Pricing, Scale Limits, and Architecture Tradeoffs Across Nine Leading Systems
Research & Papers

Best Vector Databases in 2026: Pricing, Scale Limits, and Architecture Tradeoffs Across Nine Leading Systems

A comprehensive guide comparing nine leading vector database systems used as core infrastructure for RAG and agentic AI...

MarkTechPost
NVIDIA AI Just Released cuda-oxide: An Experimental Rust-to-CUDA Compiler Backend that Compiles SIMT GPU Kernels Directly to PTX
Research & PapersGoogle

NVIDIA AI Just Released cuda-oxide: An Experimental Rust-to-CUDA Compiler Backend that Compiles SIMT GPU Kernels Directly to PTX

NVIDIA Labs released cuda-oxide v0.1.0, an experimental Rust compiler backend that enables developers to write GPU...

MarkTechPost
How to Build a Single-Cell RNA-seq Analysis Pipeline with Scanpy for PBMC Clustering, Annotation, and Trajectory Discovery
Research & Papers

How to Build a Single-Cell RNA-seq Analysis Pipeline with Scanpy for PBMC Clustering, Annotation, and Trajectory Discovery

This tutorial demonstrates an advanced single-cell RNA-seq analysis workflow using Scanpy on the PBMC-3k dataset,...

MarkTechPost