Anthropic Briefly Suspends Peter Steinberger's Access to Claude

Peter Steinberger, the creator of OpenClaw, announced early Friday morning on X that his access to Claude, the AI model from Anthropic, had been temporarily suspended. He shared a photo of the message received from Anthropic, indicating that his account had been suspended for "suspicious activity." This post quickly went viral, sparking numerous comments, some of which veered into conspiracy theories due to Steinberger's current employment at OpenAI.

A few hours after posting his message, Steinberger informed that his account had been restored. Among the hundreds of comments, an engineer from Anthropic clarified that the company had never prohibited the use of OpenClaw and offered his assistance to Steinberger. However, it remains uncertain whether this intervention directly led to the reactivation of the account. We have reached out to Anthropic for clarification on this matter.

This suspension comes shortly after Anthropic announced that subscriptions to Claude would no longer include "third-party harnesses," such as OpenClaw. Users are now required to pay separately for usage via the Claude API. Steinberger stated that he complied with this new rule and was using the Anthropic API, yet he was still banned. Anthropic justifies this change by stating that current subscriptions are not suited to the "usage models" of claws, which can be more computationally intensive. Claws can execute continuous reasoning loops, automatically repeat or retry tasks, and connect to numerous other third-party tools.

Steinberger expressed skepticism regarding this explanation, noting that the pricing changes occurred shortly after Anthropic integrated features similar to those of OpenClaw into its own agent, Cowork. Among these features, Claude Dispatch allows for remote control of agents and task assignment, and was launched shortly before the pricing policy changes. Steinberger posted: "It's funny how the timings align; first they copy some popular features into their closed harness, then they exclude open source."

Steinberger's frustration with Anthropic was again evident on Friday. One person implied that it was partly his fault for taking a job at OpenAI instead of Anthropic, posting: "You had a choice, but you went to the wrong place." To which Steinberger replied: "One welcomed me, the other sent legal threats."

When several people asked him why he was using Claude instead of his employer's models, he explained that he was using it solely for testing purposes, to ensure that updates to OpenClaw would not disrupt Claude users. He clarified: "We need to separate two things. My work at the OpenClaw Foundation where we want to ensure OpenClaw works perfectly for all model providers, and my job at OpenAI to help them with their future product strategy."

Many also pointed out that the need to test Claude arises from the fact that this model remains a popular choice for OpenClaw users compared to ChatGPT. Steinberger has heard feedback on Claude's popularity and responded: "I'm working on it." This gives a hint about what his role at OpenAI might entail.