Daily AI Briefing

Anthropic CEO Dario Amodei has firmly refused the Pentagon's demands for unrestricted access to the company's AI systems, stating he "cannot in good conscience accede" to the military's requests. This sets up a major confrontation as the Pentagon deadline looms. The move highlights the growing tension between AI safety priorities and military adoption.

DeepMind released Nano Banana 2, combining Pro-level capabilities with Flash speed for image generation. Also announced: Gemini 3.1 Pro for complex reasoning tasks, Lyria 3 music generation model integrated into the Gemini app, and Gemini 3 Deep Think mode for advanced scientific research.

OpenAI published its February 2026 Threat Report examining how malicious actors combine AI models with websites and social platforms. The report highlights emerging attack vectors and detection challenges.

OpenAI Codex now integrates with Figma for seamless code-to-design workflows.

Mistral AI inks global consulting partnership with Accenture, joining OpenAI and Anthropic as clients.

Launching Ada, an email-based digital twin that manages schedules and extracts answers from knowledge bases.

Dorsey halved Block's workforce to ~4,000, signaling more "efficiency" ahead for tech.

Zuckerberg spotted at Prada Fashion Week, fueling speculation about luxury-brand AI glasses collaboration.