A key distinction in Hermes: sub-agents are copies of the main agent used to parallelize tasks with the *same* skill set (like coding multiple app features). Profiles are distinct agents with unique skills, better for multi-step workflows requiring different capabilities (e.g., research then writing).

The most sophisticated loops don't execute all work in a single thread. Instead, a primary agent identifies sub-tasks and instantiates new, specialized "sub-agents" to handle them autonomously. This creates a powerful, scalable hierarchy of automation.

The most dramatic productivity gains come not from a single AI assistant, but from a human operator orchestrating multiple specialized agents concurrently. This model involves setting up 5-15 agents with specific roles and controlled tool access to perform complex tasks in parallel.

The workflow of a "100x engineer" involves managing multiple AI coding agents simultaneously, with each agent working independently on tasks. The engineer's role shifts from writing code to orchestrating these agents, rotating attention between them like a conductor directing an orchestra.

Complex AI development uses a pool of specialized agents. Like ants building a hill, some are workers, some are managers, and some review and discard bad code. This collaborative, layered system produces emergent results without a single orchestrator.

Agents in Buzz don't alter local files. They create separate Git work trees to build and test features in parallel, allowing for safe, simultaneous software development. Agents can even push to their own hosted repositories, creating a self-contained ecosystem.

The agent development process can be significantly sped up by running multiple tasks concurrently. While one agent is engineering a prompt, other processes can be simultaneously scraping websites for a RAG database and conducting deep research on separate platforms. This parallel workflow is key to building complex systems quickly.

As AI generates more code, the bottleneck is no longer writing but managing parallel streams of work from AI agents. This shift is making single-threaded editing tools like Cursor obsolete in favor of multi-agent management platforms like Superset, which orchestrate cloned codebases for each agent.

By deploying multiple AI agents that work in parallel, a developer measured 48 "agent-hours" of productive work completed in a single 24-hour day. This illustrates a fundamental shift from sequential human work to parallelized AI execution, effectively compressing project timelines.