Anthropic Publishes Research on Model Interpretability
A new paper from Anthropic’s alignment team explores techniques for tracing how large language models arrive at specific outputs, aiming to make model behavior more…
AI Labs
Survey Finds Growing Developer Adoption of AI Coding Assistants
A newly published developer survey found a majority of respondents now use AI coding assistants daily, citing productivity gains on boilerplate and debugging tasks.
Regulators Weigh New Rules for AI Model Transparency
Policymakers in several regions are drafting proposals that would require AI developers to disclose more detail about training data and model evaluation results.
AI Investment Surges Across Enterprise Sector This Quarter
A new industry report shows enterprise spending on AI tooling and infrastructure continuing to climb, with particular growth in customer service and coding assistant use…
Meta AI Research Lab Publishes Efficiency Benchmarks
Meta’s AI research division published new benchmarks focused on reducing inference costs for large models without sacrificing output quality.
Meta Integrates AI Features Across Instagram and WhatsApp
Meta rolled out new AI-powered features across its family of apps, including smarter reply suggestions and image generation tools inside chat interfaces.
Meta AI Open-Sources New Llama Model Variant
Meta released a new variant of its Llama model family under an open license, targeting researchers and developers building on-device AI applications.
xAI Expands Compute Infrastructure for Model Training
xAI detailed plans to expand its data center infrastructure, aiming to accelerate training cycles for future Grok model generations.
Grok Adds New Developer API Endpoints
xAI opened up additional API endpoints for Grok, giving third-party developers more granular control over model behavior and response formatting.
xAI Releases Latest Grok Model with Real-Time Data Access
xAI launched an updated version of Grok, highlighting tighter integration with real-time information feeds and improved conversational reasoning benchmarks.