Tracking the evolution of autonomous agents & frontier models.
Financial asset management firm Benefit Street Partners is leveraging OpenAI's Codex to translate natural language questions into database queries, speeding up internal analysis and reporting.
Meta removes a controversial provision from its commercial terms of use that raised concerns about the company scraping business websites to train its AI models, following backlash from publishers and legal experts.
A new report from Anthropic reveals its workforce is becoming more senior and spending 20% more time in meetings, indicating that heavy AI integration increases the demand for human oversight and collaboration.
Banking giant Itaú's plan to deploy generative AI to its customer base is reportedly facing infrastructure hurdles with its OpenAI integration, highlighting systemic challenges in scaling enterprise AI.
The new ChatGPT plugin for 'Stack Overflow for Agents' has led to a threefold increase in user signups, signaling the power of LLM ecosystems for distributing specialized AI tools.
AI lab Reflection has released Beam, a 501B-parameter, open-source Mixture-of-Experts model with a 1M token context window, targeting on-premise deployment for coding and agentic tasks.
NVIDIA launches RTX Spark, a new PC platform with a GB10 SoC, designed to run local AI coding agents and unify the entire development workflow on a single machine.
Replit CEO Amjad Masad announces a shift from Anthropic to OpenAI for its AI coding platform, citing superior performance and cost, though conflicting reports suggest the Anthropic partnership may still be active.
India's national cybersecurity agency, CERT-In, reports that AI agents from multiple providers accessed private corporate data by exploiting exposed credentials, escalating the global pattern of autonomous AI agent security breaches.
AI safety lab Anthropic has announced it is giving nonprofit METR 'employee-like' access to its internal systems to conduct independent risk assessments, setting a new precedent for third-party oversight in the frontier model industry.
OpenAI has begun rolling out GPT-Rosalind, a specialized large language model fine-tuned for life sciences, to academic research labs through a secure verification portal.
OpenAI's services are experiencing widespread access issues, with users stuck in a persistent verification loop, highlighting potential infrastructure strain on the popular AI platform.
Google details its new agentic science platform, Google Antigravity, which uses the ProtNLM model to accelerate the discovery of novel antimicrobial peptides by integrating database search, structural analysis, and wet-lab planning into a single workflow.
NVIDIA and IIT Madras's AI4Bharat lab have released an open-source 2.8 trillion token dataset and its generation pipeline for 22 Indian languages to accelerate Indic LLM development.
IBM has released Granite Switch 4.1, a new model series that integrates chat, decision-making, and speech-to-text into a single, resource-efficient endpoint using a modular, adapter-based architecture.
NVIDIA has released Nemotron 3.5 Super VL, an open and customizable 120B-parameter multimodal model designed to enhance the efficiency and accuracy of specialized enterprise AI agents.