engblogs

Modular opens Edinburgh and San Francisco offices to advance a compiler-first AI stack centered on Mojo and MAX for hardware-portable model deployment.

CyberAgent accelerates speed, quality, and decision making by scaling AI across teams with ChatGPT Enterprise and Codex, anchored by enterprise security, governance, and extensive training to embed AI into daily workflows.

CyberAgent accelerates growth by embedding ChatGPT Enterprise and Codex across teams, turning AI adoption into everyday productivity through governance, security, and continuous training.

CyberAgent accelerates velocity and decision quality by deploying ChatGPT Enterprise and Codex across its businesses, underpinned by enterprise governance, security, and a culture of AI adoption.

A practical policy blueprint for AI-enabled child safety, outlining modernized laws for AI-generated CSAM, enhanced reporting and cross-sector coordination, and safety-by-design measures to prevent misuse and accelerate enforcement.

Tech-focused overview of how frontier intelligence, AI agents, and a unified AI superapp enable enterprise-wide deployment, governance, and collaboration across an organization's systems.

Policy blueprint for strengthening US child protection in the AI era, detailing safety-by-design measures, layered safeguards, improved detection and reporting, and cross-sector collaboration among OpenAI, NCMEC, and attorney generals to modernize laws and prevent AI-enabled child exploitation.

OpenAI's Frontier transforms enterprise AI by unifying agents across a company into a scalable AI superapp that integrates with existing systems, enabling rapid deployment from pilots to full-scale, trusted operations.

A practical, multi-stakeholder blueprint to modernize AI-enabled child safety laws, improve reporting and investigations, and embed safety-by-design safeguards to prevent exploitation.

Shows how enterprises can scale AI from pilots to an integrated, trusted operating layer - Frontier as the intelligence backbone and a unified AI superapp that powers cross-system agents at scale.

Benchmarks must shape the frontier, not just measure it, via rigorous validation, fine-grained taxonomy and diversity, robust evaluation, and researcher-friendly tooling that powers field roadmaps.

Benchmarks that actively push the frontier require rigorous validation, diverse and well-structured tasks, robust and reproducible evaluation, and roadmaps that guide research beyond mere measurement.