Descriptions:
Michael Richman, an engineering leader at Bitly who co-wrote the company’s MCP server, presents Command+Control, a mobile-first system he built to solve what he names FOMAT — Fear of Missing Agent Time. The core problem: as AI coding agents take on longer-running tasks, developers have no reliable way to know when a session is blocked waiting for input, has gone off the rails, or has completed — especially when away from their dev machine. Richman notes that tasks currently ranging from 5 to 45 minutes will grow to hours or days, making terminal-based babysitting economically unsustainable.
Command+Control provides a unified dashboard across iOS, Android, and web that aggregates sessions from Claude Code, OpenAI Codex, GitHub Copilot, and Cursor into a single interface. Developers receive push notifications when agents complete tasks or get stuck, can respond to prompts directly from their phone, and can launch entirely new sessions remotely — including from a smartwatch. Live demos show a Claude Code session receiving a notification on an iPhone simulator, a mobile-initiated session resuming seamlessly in the terminal, and a session management view organizing sessions by notification priority, monitoring status, and recency.
Richman notes that Anthropic and Cursor both shipped features addressing the same problem in the days surrounding his talk, validating the underlying workflow gap. He frames Command+Control as one practical solution to a challenge the whole industry is actively converging on: as agentic coding matures from a terminal-centric workflow into something closer to asynchronous infrastructure management, the interface layer for human oversight needs to follow developers wherever they are.
📺 Source: AI Engineer · Published May 24, 2026
🏷️ Format: Showcase
