Anthropic just pulled a move that's got the AI agent community buzzing. Peter Steinberger, creator of the popular open-source tool OpenClaw, was temporarily banned from accessing Claude early Friday morning after posting a screenshot of his account suspension citing "suspicious activity." The ban lasted only a few hours — once his post went viral, Anthropic reinstated his access. An Anthropic engineer even jumped into the comments to clarify that the company has never banned anyone for using OpenClaw and offered to help sort things out.

The Timing Is... Something

Here's where it gets interesting. This ban came just one week after Anthropic announced that subscriptions to Claude would no longer cover "third-party harnesses including OpenClaw." Users now have to pay for that usage separately through Claude's API — essentially a "claw tax." Anthropic claims subscriptions weren't built to handle the usage patterns of claws, which can run continuous reasoning loops, automatically retry tasks, and tie into third-party tools. That's a lot of compute compared to simple prompts. Steinberger isn't buying the explanation. After Anthropic changed the pricing, he posted: "Funny how timings match up, first they copy some popular features into their closed harness, then they lock out open source." He's likely referring to Claude Dispatch — a feature that lets users remotely control agents and assign tasks — which rolled out a couple of weeks before Anthropic changed its OpenClaw pricing policy. Coincidence? In this industry, we don't believe in coincidences.

The OpenAI-Anthropic Dynamic

One commenter suggested some of this is on Steinberger for taking a job at OpenAI instead of Anthropic. His response? "One welcomed me, one sent legal threats." Ouch. When asked why he uses Claude at all despite working for OpenAI, he clarified that he only tests with it to ensure OpenClaw updates won't break things for Claude users — a totally reasonable workflow for maintaining cross-platform compatibility. He also responded "Working on that" when someone mentioned Anthropic's new pricing, which is a pretty strong hint about what his product strategy work at OpenAI actually entails.

Key Takeaways

  • Anthropic temporarily banned OpenClaw creator Peter Steinberger from Claude, reinstating him hours later after his suspension post went viral.
  • The ban came one week after Anthropic announced a new "claw tax" requiring separate API payments for third-party harness usage.
  • Steinberger now works at OpenAI and uses Claude only for cross-platform compatibility testing.

The Bottom Line

This whole thing smells like competitive lock-in. Anthropic builds a popular open-source tool into their ecosystem, watches thousands of developers adopt it, then suddenly decides they want a cut. The temporary ban on the creator — who just happens to work for their biggest rival — reads as either poorly-timed automated enforcement or something more intentional. Either way, it's a bad look for Anthropic and exactly the kind of behavior that makes developers paranoid about platform dependency. The open-source community should watch this closely.