Two Anthropic announcements, two outlets, zero overlap
Anthropic dropped two distinct announcements Thursday: a free security scanning tool for open-source projects called OSS Scanner, and a usage policy update barring users from being 'abusive or cruel' to Claude. The Verge covered the scanner. The NY Post covered the behavior policy. Neither covered both. The result is that readers of either outlet got half a product launch and a complete picture of neither.
Anthropic launches free AI security scans for open-source projects and new policy against abusing Claude AI
2 sources · hover a dot to see coverage
What happened
On Thursday, Anthropic made two separate announcements. First, it launched OSS Scanner, a free service offering open-source projects 'thorough, periodic security' vulnerability scans. Second, it updated its usage policy to prohibit users from engaging in 'abusive or cruel behavior' toward Claude. The behavior policy is notable because Anthropic is simultaneously arguing, in ongoing legal and public contexts, that Claude is not sentient. The company did not publicly explain what enforcement of the behavior policy looks like in practice. Two outlets covered the story. Each covered one announcement and not the other.
The Verge reported the scanner; the policy update does not appear
The Verge's coverage focuses entirely on OSS Scanner, describing it as opt-in and framing it as a security infrastructure play for the open-source community. There is no mention of the behavior policy update in the excerpt. For readers who follow AI safety and open-source tooling, the scanner is the more operationally significant announcement. But the behavior policy is the one that raises a genuine philosophical question Anthropic has not answered: what is the enforcement mechanism for a rule against being mean to software?
The NY Post led with the behavior ban and skipped the scanner entirely
The NY Post's piece opens with the Claude behavior policy and includes the editorial aside 'even though it's not, ya know, actually alive.' The framing is bemused, and the skepticism is fair given Anthropic's own legal positioning on Claude's non-sentience. The OSS Scanner announcement does not appear. The Post's coverage is the more entertaining read and also the less informative one on the question of what Anthropic actually shipped Thursday.
What one side told you that the other didn't
No single outlet reported both announcements Anthropic made Thursday.
Anthropic put out two distinct product and policy items in one day. The Verge took the technical one. The NY Post took the quirky one. Neither publication apparently noticed, or cared, that the other story existed. A reader who saw both headlines would think they were looking at different news days. The practical consequence: the behavior policy's enforcement gap (what actually happens when a user violates it?) went unasked in both pieces.
The behavior policy's enforcement mechanism went unreported in both outlets.
Anthropic's updated policy bars 'abusive or cruel behavior' toward Claude. Neither outlet asked what enforcement looks like: does Anthropic suspend accounts, flag sessions, or simply state a preference? The NY Post noted the philosophical tension with Claude's non-sentience but did not push further. The Verge did not cover the policy at all. A rule with no reported mechanism is an announcement, not a policy.
What to watch
If Anthropic publishes enforcement details for the behavior policy, or if open-source projects begin publicly opting into OSS Scanner, watch whether the outlet that missed each announcement circles back. The behavior policy in particular will attract follow-up if a high-profile account suspension can be attributed to it.
See how outlets across the political spectrum framed this differently — and what each side left out.