Nathaniel Whittemore recommends trying ambient voice interaction, teaching AI by letting it observe a workflow and turning recurring writing preferences into reusable skillsAn AI agent skill is a reusable package of instructions, resources, and tool guidance for performing a bounded kind of work.. He also highlights visual design workspaces that let people compare alternatives and make targeted changesA visual design workspace presents AI-generated alternatives and editable elements in one interface so people can compare and refine them directly. without repeatedly prompting an entire interface from scratch.
Nathaniel Whittemore treats skill design as an ongoing discipline and points toward shared team agents as the next step beyond personal assistants. In that model, agents live in collaborative spaces with task-specific context, permissions and toolsAn AI agent workspace is a shared environment where people and AI agents communicate, coordinate tasks, exchange files, and observe work state., while computer-use agents can handle repeatable information-heavy work on their own virtual machines.
Nathaniel Whittemore closes with local AI as an increasingly capable option and a set of short conversational prompts that can trigger planning, correction, simulation and memory updatesPrompt engineering is the practice of designing and refining instructions, context, examples, and constraints to obtain useful AI outputs.. His broader advice is to reserve time for deliberate experimentation because small changes in how people interact with AI can materially change the value they get from familiar tools.
Watch on YouTube
