Seeding Liquidity for Agent Platforms

This is the core risk in any new platform business, because supply and demand have to be created at the same time. For /dev/agents, developers only spend time building agent apps if there is real distribution, and users only come back if the system already does useful jobs out of the box. That means the company cannot rely on openness alone, it has to seed both sides with its own apps, tight distribution, and clear early use cases.

• The closest pattern is app marketplaces like Shopify and HubSpot. Developers build once there because merchants are already present and discovery is built in. HubSpot still requires active installs for listing and has imposed install caps before listing, which shows how much distribution rules shape developer behavior on a platform.
  5 hubspot 6 hubspot 7 hubspot
• On /dev/agents, the cold start is harder because the app is not just a plug in, it often needs to orchestrate across other apps and operating system actions. That raises setup friction for users and platform risk on iOS, where Apple bars apps that depend on another app to function and rejects alternate home screen environments.
  1 sacra 8 apple
• The practical way through this is to start with a few repeatable jobs where the agent is obviously better than today’s phone workflow, then let outside developers extend those workflows. That is how a platform turns from a demo into a habit, first by solving a narrow daily task, then by widening the catalog around it.
  1 sacra 2 sacra 4 sacra

Going forward, the winners in agent operating systems will be the ones that manufacture their own early liquidity. /dev/agents is likely to move from broad platform ambition toward a more controlled playbook, with first party experiences creating user demand first, then third party developers filling in the long tail once install base and distribution are real.