10 results on this page
OpenAI said it will enhance safety processes for paying users who access its most advanced AI models. The move responds to customers using AI for increasingly complex tasks and handling more sensitive information.
Meta is expanding Pocket, an experimental AI-powered app for creating and sharing interactive games, to users across the U.S. The app was previously tested quietly in Brazil.
A position paper analyzing 500 Hugging Face model cards argues that model cards alone do not provide enough information for governing open-weight foundation models. It proposes combining model cards with acceptable use policies and licenses to address safety, provenance, behavior, and enforcement gaps.
NVIDIA FLARE enables organizations to coordinate training for vision-language models while keeping sensitive image and text data at local institutions. The approach supports multimodal use cases such as visual question answering, captioning, and image-text reasoning without centralizing raw records.
Expedia Group is reorganizing its product and technology organization around AI, resulting in the departure of at least eight vice presidents and senior vice presidents. An internal memo also announced promotions for five leaders.
Google DeepMind researchers found that training AI agents through debate can reduce reward hacking when an LLM judge evaluates their work. On mathematics tasks, debate achieved higher sustained ground-truth accuracy than directly optimizing for LLM judge rewards.
A developer reports large prediction errors after deploying a MobileNetV3 model converted to TFLite in a Flutter camera application. The issue may involve YUV-to-RGB conversion, resizing, tensor layout, normalization, quantization, or camera orientation mismatches.
Big Tech companies are increasing local public-relations efforts to win support for the data centres needed to expand AI infrastructure. Campaigns reportedly include town halls, tax concessions, and sponsorship of community groups as opposition grows across the United States.
UC Berkeley mathematics professor Zvezdelina Stankova wrote an op-ed criticizing students who cannot perform middle school mathematics. The discussion shifted after an AI detector reportedly flagged the human-written article, highlighting concerns about detector reliability.
Vercel Functions now support Bun 1.4, a major rewrite from Zig to Rust with over 2,900 issues resolved and more than 1,500 additional Node.js compatibility tests passing. The upgrade is opt-in and includes breaking changes that developers should review before deployment.