Daily intelligence on AI, technology, and the decisions that follow.
Written for people who make decisions — not people who make predictions.
Get The Signal →Free to start. No spam. Unsubscribe anytime.
Monday, March 30, 2026 — Evening Edition 🔄 Microsoft Ships Copilot Cowork: Multi-Agent Orchestration Inside M365 Agents & Autonomy Microsoft launched Copilot Cowork in its Frontier tier — a supervisor model that dispatches specialized AI agents across Excel,
Read the briefing →Anthropic warns government that Mythos elevates cyberattack risk. Amazon + OpenAI launch stateful runtime on Bedrock. RSAC 2026 marks $3.6B in agentic security funding.
Anthropic eyes October IPO. SoftBank's $40B bridge loan. The triple public market wave approaching $1.4T combined valuation.
A leaked model, an identity crisis at RSAC, and the enterprise security playbook rewritten in real time.
Every major frontier lab now holds a DoD contract. One hardware chief already walked. The governance framework that was supposed to handle this doesn’t exist.
Bezos wants $100B to automate factories. The White House wants one AI law for all 50 states. Meta's agent leaked data for two hours.
Cursor builds its own model. OpenAI buys its toolchain. 45K layoffs, 20% owned by AI. Three signals pointing the same direction.
Meta's ML agent runs in production, Jensen redefines compensation with token budgets, and Microsoft codifies Zero Trust for AI.
OpenAI goes classified. Atlassian cuts 10% citing AI. A CEO puts a number on what college grads are about to face.
NVIDIA doubles its revenue forecast to $1T, ships the first commercial humanoid robot model, and open-sources world simulators for surgical AI.
Dorsey's 40% cut under scrutiny, NVIDIA builds the first CPU for agents, and inference cost drops 10x.
Morgan Stanley puts numbers on the deflationary shock. Meta funds its AI bet by cutting 20% of its people. And agents now have their own social network.