AI company Anthropic recently temporarily banned Peter Steinberger, founder of the open-source AI assistant framework OpenClaw, from accessing its Claude platform. The incident stemmed from Anthropic's announcement last week to adjust its Claude Code subscription policy, explicitly stating that it will no longer cover the usage costs of third-party tools including OpenClaw, which some developers jokingly called the "Claw tax." Although Steinberger stated that he had complied with the new rules by switching to the API, his account was still suspended for "suspicious activity."
This ban highlights the growing tension between AI platforms and the third-party developer ecosystem. Anthropic is actively promoting its own agent product "Cowork" and has added new features such as remote control and task assignment. Steinberger hinted on social media that the platform first integrated popular features from open-source tools into its proprietary product, then used policy adjustments to sideline open-source projects. This controversy underscores the fundamental conflict between platform control and developer innovation freedom in the AI application layer.
