Clankers with claws

David Heinemeier Hansson’s bet is that agents will end up using the interfaces built for people, rather than needing purpose-built ones. His analogy:

if I was going to skate to where the puck is going to be, it’d be a world where agents, like self-driving cars, don’t need special equipment, like LIDAR or MCPs, to interact with the environment

To test it he gave an agent called Kef no skills, MCPs or APIs, and ran it on a Proxmox virtual machine isolated from his own logins and data. From simple prompts it signed up for HEY email, created a Fizzy account, built a board of five business ideas with images it found on the web, and accepted a Basecamp invitation that arrived by email. He reports zero corrections. He ran it first on Claude Opus 4.5, then repeated most of it on the Chinese open-weight model Kimi K2.5.

What makes this different from a chatbot demo, in his telling, is that the agent is no longer stuck in a prompt-and-response cycle: it can check its own inbox and notifications and carry on from what it finds. That is what turns a sequence of separate tasks into one continuous piece of work.

He does not push the conclusion too far. The accommodations built for agents, MCPs and CLIs and APIs, still have a place for a while yet, because working cold through web interfaces is slower and burns more tokens. The idea worth keeping is that the accommodations are a stopgap, not the destination.