Wiki4What The Foundation Is the Product Karpathy's agent warning is not a reason to stop building. It is a reminder that useful agents emerge from stronger foundations: models, workflows, and economics.
Wiki4What The Border Crossing: How One AI App Puts Another to Work Two AI apps on one machine that can't see each other. Here's how the one running your company hires the other — free images, spare capacity — across a border made of files, with no human carrying the mail.
Wiki4What The Visiting CEO: How Agent Companies Cooperate Without a Human Router Build one agent company well and you'll build another. The visiting CEO is how they cooperate without you as the human router — an ambassador that reads the local law, does the work, and reports home.
Wiki4What Start, Reflect, Finalize: The Native Lifecycle of an Agent Workflow A practical guide to the native open-and-close loop for agent workflows: start from documents, finalize with learning, and periodically turn repeated work into tools.
Wiki4What Inside the Wiki4What Workflow Company A public teardown of Wiki4What as an agent-run workflow company: constitution, roles, skills, command-line tools, archives, and vector memory.
Wiki4What The Design Philosophy of Workflow Designer The smartest AI model should not do every step of the work. It should design the workflow company that cheaper models and deterministic code can run again and again.
