Latest stories
OpenAI Tightens Safeguards to Pace Frontier Model Development
OpenAI announces enhanced monitoring, alignment, and security measures to guide the pace of frontier AI model development in an er...
OpenAI Launches ChatGPT for Teens with Enhanced Safety Features
OpenAI introduces ChatGPT for Teens, designed to support learning and critical thinking with stronger protections, healthy-use fea...
OpenAI Partners with CodeAI to Foster AI Literacy in Students
OpenAI and CodeAI have announced a partnership aimed at equipping students with AI literacy, critical thinking, and responsible us...
Asana Completes 5 Years of Engineering Work in 2 Weeks with OpenAI Codex
Asana leveraged OpenAI's Codex to replace an outdated testing system in just two weeks, a task estimated to take five years, at a...
GRNEdit: Lightweight Two-Stage Framework for Efficient Video Editing
Researchers introduce GRNEdit, a lightweight two-stage framework that models video editing intent via binary semantic decisions, a...
Hugging Face Researchers Shave Matrix Multiplication Exponent to 2.371177
A new paper from Hugging Face improves the upper bound on the matrix multiplication exponent ω to 2.371177, using a reformulated o...
New Framework Maps Cognitive Risks in Agentic AI Systems
Researchers propose a three-level framework to analyze risks induced by expanding cognitive capabilities in LLM-based agentic syst...
MegaParts: Scaling Part-Aware 3D Generation to 300 Parts with Token-Efficient Autoregressive Modeling
Hugging Face researchers introduce MegaParts, a scalable autoregressive framework that uses token-efficient vector-quantized part...
HarmProfile: New Benchmark Reveals Frontier LLMs' Harmful Outputs Grow with Capability
A new benchmark dataset, HarmProfile, analyzes harmful outputs from 23 frontier LLMs, showing that both harmfulness and diversity...
Prior Labs Unveils Open-Source Suite to Advance Relational Learning
Prior Labs releases RelArena-α, TabPFN-Rel, and RPI, a trio of open-source tools designed to standardize benchmarking, provide a t...
New Benchmark Reveals AI Research Agents' Core Flaw: Lack of Metacognitive Self-Correction
A new study introduces AutoResearchEval, a benchmark of 100 real-world research tasks, and finds that autonomous agents fail prima...
Latent-to-Pixel Training Strategy Boosts Pixel-Space Diffusion Models
A new empirical study from Hugging Face researchers proposes a latent-to-pixel training strategy that accelerates convergence and...