Alex Finn compares Hermes Agent with OpenClaw and says Hermes has become his primary tool because its updates have been more reliable, its model support arrives quickly and its roadmap feels clearer. Alex Finn still uses OpenClaw for shorter, more conversational responses, but prioritizes Hermes for repeatable work.
Alex Finn says he replaced Claude with GPT-5.6 inside Hermes Agent because the newer model performed better for his tasks, used computer and browser controlsComputer use lets an AI agent interact with software through interface controls such as screens, buttons, keyboards and pointers. more effectively, and worked through an existing subscription instead of a costly API workflow. He recommends the medium reasoning settingReasoning effort is the amount of internal computational work an AI model applies before producing an answer or action. as a balance between capability, speed and instruction following, with low reasoning as an option for tighter usage limits.
Alex Finn demonstrates three practical uses: selecting and managing local modelsA local model runs on computing infrastructure controlled directly by the user rather than only through a remote model service. for a home AI lab, operating Unity as an autonomous game-development environment, and monitoring retail pages for price changesWorkflow automation uses software or AI to complete a connected sequence of routine steps with less manual effort.. These examples support his broader point that model choiceModel selection chooses the AI model whose capability, quality, cost, speed, safety, and operating constraints best fit a task. matters most when it improves reliable tool use across a concrete agent workflow.
Watch on YouTube
