Peter Steinberger, the innovative creator of the AI agent OpenClaw, faced a torrent of social media notifications this past Saturday. The news that Anthropic had announced Claude Code (Claude Pro\/Max) accounts could no longer be utilized in OpenClaw sent ripples through the community. While many users of the platform expressed their discontent, others recognized the strategic reasoning behind Anthropic’s move. This shift raises intriguing questions about the landscape of AI regulation and competition.<\/p>\n
OpenClaw is an AI agent designed to seamlessly operate your machine and manage applications based on user requests. Its functionality becomes strikingly powerful when paired with quality models like Claude Opus 4.6 or Claude Sonnet 4.6. However, Anthropic’s latest announcement emphasizes that OpenClaw consumes tokens at an unsustainable rate, especially for its Pro and Max users\u2014plans initially meant for programming purposes, not general AI tasks.<\/p>\n
Anthropic has clarified that while it’s still possible to use Claude with OpenClaw, it now comes with a cost. Users must implement Anthropic’s API to access these models\u2014a significant change from the previous flexible usage policy. This shift can be likened to buying a transport pass that only allows limited use outside typical commuting times, causing frustration among loyal users.<\/p>\n
The decision has provoked backlash from users who feel that it undermines the utility of OpenClaw. Boris Cherny, a significant figure at Claude Code, acknowledged the backlash on social media, explaining that a commitment to optimizing subscription models for a larger customer base can sometimes mean making difficult decisions. The reality is that heavy usage of Claude via OpenClaw may not be economically feasible for Anthropic in the long run.<\/p>\n
The timing of this announcement is particularly noteworthy. It follows quickly on the heels of Anthropic’s release of new features\u2014Claude Cowork, Dispatch, and Remote Control\u2014that align closely with OpenClaw’s functionalities. Steinberger has pointed out the irony of Anthropic releasing these features only to restrict OpenClaw access shortly afterward. This leads to the suspicion that Anthropic is setting up a proprietary ecosystem reminiscent of established players like Nintendo and Apple.<\/p>\n
In the broader context of technology, Anthropic’s strategy mirrors the walled garden approach, where successful features from competing products are absorbed, while access to external innovations is curtailed. Though Anthropic is still in the early stages of this transition, the parallels with companies that enforce strict access controls in their ecosystems become apparent.<\/p>\n
While Nintendo and Apple have built their empires on distinctive intellectual properties such as Mario or the iPhone, Anthropic’s bouquet of offerings does not yet boast a comparable legacy. However, the growing control over API access, coupled with a pay-per-use model, continues to shape their emerging ecosystem. Comparatively, other companies, particularly in the rapidly evolving Chinese AI startup scene, remain more lenient, allowing users to access their models sans stringent limitations.<\/p>\n
As the dust settles on Anthropic’s recent decision, it’s clear that the industry is at a crossroads. With growing functionalities and tightening access, Anthropic appears to be laying the groundwork for an exclusive ecosystem in the realm of AI. How this will affect competition, user independence, and future innovations remains to be seen, but one thing is certain: the AI landscape is changing, and users will need to adapt.<\/p>\n