Peter Steinberger, the innovative mind behind OpenClaw, recently shared on X that his access to Anthropic's Claude was temporarily restricted due to "suspicious" activity. This announcement stirred considerable attention online, prompting a swift response from Anthropic.
Fortunately, Steinberger's account was reinstated just hours later. The rapid turnaround followed a flurry of comments, including one from an Anthropic engineer who clarified that the company has never prohibited the use of OpenClaw and expressed willingness to assist Steinberger.
This incident occurred against the backdrop of a recent policy change from Anthropic, which stated that subscriptions to Claude would no longer encompass third-party tools like OpenClaw. Users are now required to pay separately for API access based on their usage. This shift has been informally dubbed a "claw tax" by the community.
Anthropic explained that the pricing adjustment was necessary due to the unique usage patterns associated with tools like OpenClaw, which can demand more computational resources compared to standard prompts or scripts. Steinberger, however, expressed skepticism about this rationale, suggesting that the timing of the ban coincided suspiciously with Anthropic's rollout of new features for their own agent, Cowork.
In a recent exchange, Steinberger defended his decision to utilize Claude for testing purposes, emphasizing his commitment to ensuring that updates to OpenClaw remain compatible with Claude users. He stated, "My work at the OpenClaw Foundation aims to enhance the experience for all model providers, while my role at OpenAI focuses on future product strategy."
Despite facing criticism for his transition to OpenAI, Steinberger remains dedicated to his dual roles, highlighting the ongoing popularity of Claude among OpenClaw users. His efforts to navigate this complex landscape reflect a broader trend in the technology sector, where collaboration and competition often intersect.
As the dynamics of AI tools evolve, these developments could significantly influence how third-party applications interact with major AI models, potentially fostering a more collaborative environment in the tech industry.
