Tech Times

Compile Once, Run Offline: New AI Method Matches 32B Models With a 23MB File

Local AI inference at 32B-parameter quality, no cloud API required: University of Waterloo researchers released PAW on July 2 ...

Waterloo's PAW compiles task specs into 23MB LoRA adapters a 600M-parameter model runs entirely offline.

Local AI inference at 32B-parameter quality, no cloud API required: University of Waterloo researchers released PAW on July 2, 2026, a system that compiles any natural-language task spec into a 23MB ...
InfoQ

Hardwood Promises High-Speed JVM Apache Parquet Processing with Zero Mandatory Dependencies

Hardwood, the project Gunnar Morling kick-started handling of Parquet files in Java, reached version 1. Its multi-threaded approach and zero mandatory external dependencies promise a simpler, more ...

Data protection rules slow LLM rollout in Europe, study says

A new GovAI study reveals that EU data protection rules are stalling AI adoption, leaving 11% of advanced LLM releases ...
The BMJ

Australia increases penalties for social media companies to enforce under 16s ban

Australia is stepping up its attempts to prevent children under 16 from accessing social media, a year on from becoming the first country in the world to enact a ban. In new moves, the Australian ...
eLife

TopoMetry systematically learns and evaluates the latent geometry of single-cell data

A framework for analyzing single-cell genomics data, in which geometrical properties are harnessed to obtain insights on cellular diversity, including precise clustering, clear visualizations, and ...