… While there’s no certainty this tool works until Anthropic releases the software it uses to detect a watermark, understanding the basic SynthID-text approach underpinning Claude’s watermarking makes them fairly sure the method works, says Wayne Pan, chief technology and cofounder at Silicon Valley–… …
… In total, Carroll believes that the vulnerability he and Claude found would have provided access to the information of millions of customers, including names, emails, and mailing addresses—but not credit card details—as well as that of Front Gate’s staff. …
… Despite Anthropic’s efforts to control who can use Mythos Preview, a group of Discord users gained access to the tool through some straightforward relatively detective work: They examined data from a recent breach of Mercor, an AI training startup that works with developers, and “made an educated g… …