Engineering
Can Codex Handle Real-World Data Analysis?
Researchers tested OpenAI’s Codex on real-world data analysis tasks spanning business and scientific domains. The findings show Codex performs well…
Engineering
How to visualize any Hugging Face model
Embedl introduced hfviewer.com, a tool that enables visualization of Hugging Face deep learning models through three workflows: URL replacement, a…
Engineering
After the party comes the free lunch: regularizing ColBERT models to enhance pooling capabilities and reduce index footprint
This article from LightOn AI researchers describes a Straight-Through Estimator-based regularization technique that improves hierarchical pooling in ColBERT-style retrieval models.…
Engineering
Atom2.7m: Representation-Level Specialization for Arithmetic-Aware Small Language Models
This article introduces Atom2.7m, a 2.74-million-parameter language model that reaches 69.24% accuracy on arithmetic benchmarks through an arithmetic-aware numeric representation…
Engineering
Distillation in 2026 (so far): which frontier models use it and how
Published July 8, 2026, this article surveys how frontier AI labs implement knowledge distillation through three approaches: training smaller student…
Engineering
PyTorch Team Uses LLMs to Speed Up Autotuning for Helion Kernels
PyTorch engineers described an LLM-guided autotuning method for Helion kernels that matches the performance of the existing Likelihood-Free Bayesian Optimization…
Engineering
Simon Willison: Delegate Coding Work to Lower-Power Models Through Subagents
Developer Simon Willison described a technique for using coding agents more efficiently by letting a top-tier model exercise judgment about…
Engineering
Hugging Face Brings Native vLLM Speed to the Transformers Backend
Hugging Face detailed improvements to the transformers library’s vLLM backend that let it match the throughput of hand-written native vLLM…
Engineering
PyTorch 2.13 Released With FlexAttention on Apple Silicon and a New CuTeDSL Inductor Backend
PyTorch 2.13 was released on July 8, 2026, bringing FlexAttention support to Apple Silicon (MPS) with up to 12x speedups…
Engineering
Hugging Face and SkyPilot Enable Zero-Egress Multi-Cloud Storage for AI Training Jobs
Hugging Face and SkyPilot announced an integration allowing users to mount Hugging Face Buckets and Hub repositories as storage backends…
Engineering
Hugging Face Overhauls Its Kernels Project With Signed, Verifiable Custom Kernels
Hugging Face rolled out major updates to its Kernels project, which standardizes distribution of custom compute kernels, adding kernels as…
Engineering
LeRobot v0.6.0 Adds World Model Policies and New Vision-Language-Action Models
Hugging Face released LeRobot v0.6.0, introducing three world model policies — VLA-JEPA, LingBot-VA, and FastWAM — that learn to imagine…