Model choice is becoming commoditized. Microsoft, OpenAI, and Meta are all converging on the same battleground: execution, security, and control over how agents operate in production.
OpenAI lands a record $110B round, Anthropic clashes with the Department of War, Google launches Gemini 3.1 Flash-Lite, and agent infrastructure speeds up.
Open-standard momentum accelerates: AAIF scales to 146 members, Atlassian connects Jira and Rovo to third-party tools via MCP, and Anthropic sets defense boundaries for Claude.
This week in MCP: Slack’s agent-native servers, Google’s Managed MCP for Spanner, Cloud SQL and more, Gemini 3.1 Pro and Deep Think, Claude Sonnet 4.6, Microsoft’s Security Dashboard for AI,…
This week’s MCP Weekly: EU probes Meta’s WhatsApp AI lockout, OpenAI ships Frontier and tests ChatGPT ads, Anthropic launches Claude Opus 4.6, plus hosted MCP.
From Xcode to Snowflake, agents are going first-party. This week's digest breaks down Codex on Mac, GitHub Agent HQ, OpenAI’s $200M Snowflake partnership, and why Moltbook put agent security back…