Keep pulling the thread on Sherry Davidoff.
The price of Brent crude rose above $100 per barrel after Iran seized container ships.
The US Department of Justice is moving to reclassify marijuana as a less restricted substance, causing cannabis stocks to soar.
Anthropic developed an AI model named Mythos that it deemed too dangerous for public release.
According to a Bloomberg report, a small group of unapproved users gained access to Anthropic's Mythos model for two weeks.
Anthropic's Mythos model can detect vulnerabilities in software and write exploit code, even without access to the source code.
Iran attacked three ships in the Strait of Hormuz and the Gulf of Oman.
On April 22nd, the S&P 500 and Nasdaq reached record closing highs, partly driven by news of a ceasefire extension.
The security breach of Anthropic's Mythos model occurred on the same day the company launched Project Glasswing, a controlled rollout to select companies.
Anthropic is investigating the security breach of its Mythos model but has not yet found evidence of malicious use.
The preview version of Anthropic's Mythos model has been released to over 40 technology companies.
The emergence of AI tools like Mythos will require a shift to more continuous software updates, rather than monthly or periodic patches.
The US government may need to establish standards for disclosure and tracking of AI models that possess advanced hacking capabilities.