My 2025 AI Engineer Setup: Portable, Productive and Fun (zackproser.com)

A veteran multi‑monitor enthusiast describes abandoning his elaborate Terminal Velocity rig for a highly portable, single‑laptop setup built around an M3 Pro 14" MacBook Pro, Sony WH‑1000XM5 headphones and five core tools: Wispr Flow (system‑wide voice input), Cursor (AI pair programmer + background agents), Warp (context‑aware terminal), Claude Desktop (writing/biometrics‑aware assistant) and ChatGPT Desktop (planning, image gen, advanced voice). Window management is handled by Rectangle, which recreates tmux‑style layouts on a single screen. Wispr Flow boosts his effective typing speed (claimed 90 → 175 WPM) and lets him speak intent directly into any macOS text field, while Cursor and Warp offload scaffolding, debugging and contextual command work to AI—eliminating the need to scatter docs and shells across multiple monitors. The significance for AI/ML practitioners is threefold: first, voice‑first, agentized tools are maturing enough to change developer ergonomics and workflows; second, tightly integrated desktop AIs (including biometric context from devices like Oura) enable smarter, more proactive assistants; and third, portability and a single‑viewport model simplify context sharing (screenshots pasted into Claude) and support new practices like “walking meetings.” This signals a shift in tooling priorities for the community—invest in robust multimodal input, context‑aware terminals and agent orchestration rather than larger visual real estate—and highlights UX, latency and privacy as key areas for future tool development.