Meta is reducing its plan to track employee computer activity for AI training after receiving internal criticism. Employees can now pause data collection for up to 30 minutes at a time.

AI startup Shift offers free home cleaning services to collect footage of cleaners in action. The recorded data will be used to train future cleaning robots.

Norway's National Library is developing a large language model (LLM) for the Norwegian language using 2 petabytes of Huawei OceanStor Dorado flash storage. Marius Husnes, Head of IT Platform, stated that no commercial LLM provider is currently developing a local Norwegian language model.

AI trainers in Hollywood are now focusing on tasks such as assessing chatbot tone, identifying patterns in images, and annotating video content. Professionals from the television industry are shifting their skills to train AI systems for various applications.

Meta and CEO Mark Zuckerberg are facing a lawsuit from five publishers and author Scott Turow, who allege that the company illegally copied millions of copyrighted works to train its AI systems. The plaintiffs claim that Meta's actions were motivated by a desire to advance in the AI space, reflecting the company's motto of "move fast and break things."

Uber plans to equip its human drivers' vehicles with sensors to collect real-world data for autonomous vehicle companies and other AI model training. This initiative aims to create a sensor grid leveraging the extensive network of Uber drivers.

The PyPI package 'lightning', versions 2.6.2 and 2.6.3, was compromised in a supply chain attack, affecting users of the PyTorch Lightning AI training library. The malicious versions include a hidden _runtime directory containing obfuscated JavaScript that activates upon running pip install lightning.

A cooking project involving venison required learning to shoot and tracking progress through the adaptation of a 2012 OpenCV paper. Training a state-of-the-art computer vision model contributed to a longer preparation time for the dinner.

Atlassian will start collecting customer metadata and in-app content from Jira, Confluence, and other cloud products by default on August 17, 2026, to enhance its AI offerings, including Rovo and Rovo Dev. Metadata collection is mandatory for Free, Standard, and Premium tiers, while Enterprise customers have the option to opt out.

MegaTrain is a memory-centric system that enables the full precision training of large language models with over 100 billion parameters on a single GPU. It utilizes host memory to store parameters and optimizer states, treating GPUs as transient computation units.