engblogs

Forward Deployed Engineering teams at Salesforce accelerate Agentforce AI deployments across complex enterprise architectures by resolving integration challenges, optimizing agent configurations, and leveraging AI-augmented productivity tools to reduce timelines from six months to three weeks.

Explains how ChatGPT's new group chats enable shared conversations for planning and collaboration across platforms, with privacy-preserving memory controls and invitation-based participation.

Philips is scaling AI literacy across 70,000 employees by combining leadership training, grassroots innovation, and responsible AI principles to reduce administrative burdens and enhance patient care.

ChatGPT now supports group chats, enabling collaborative conversations with friends, family, or coworkers powered by GPT-5.1 Auto, featuring shared control, personalized interactions, and enhanced social behaviors.

GPT-5.1 introduces adaptive reasoning, extended prompt caching, new developer tools, and enhanced coding capabilities for faster, more efficient, and reliable AI-powered workflows.

Training sparse neural networks with disentangled circuits enables more interpretable AI models by simplifying internal computations and isolating task-specific mechanisms.

Two Sigma's adoption of treating data as code integrates software engineering best practices like version control, CI/CD, and automated quality checks to build scalable, reliable, and maintainable data systems, transforming their data engineering approach and improving operational efficiency.

Spotify enhanced its Shuffle feature by introducing a 'Fewer Repeats' system that selects the freshest random playlist sequence to better align with user expectations of variety and surprise, while preserving true randomness.

Prime Data Centers partners with Lambda to deploy NVIDIA AI-optimized, high-density GPU infrastructure at the AI-ready LAX01 data center in Southern California, enabling superintelligence-scale AI workloads with scalable, efficient, and high-performance computing solutions.

Optimizing SaltStack triage at Cloudflare by decentralizing job caches, automating failure attribution, and enabling hierarchical self-service debugging to reduce release delays and operational toil.

Instacart revolutionizes query understanding by leveraging Large Language Models with domain-specific fine-tuning, context engineering, and production optimizations to enhance search accuracy, handle long-tail queries, and streamline system complexity.

SIMA 2 is an advanced AI agent powered by Gemini models that plays, reasons, learns, and collaborates with users in diverse virtual 3D gaming environments, demonstrating significant progress toward generalist embodied intelligence with self-improvement capabilities.