Apr. 10 at 7:55 PM
Treasury Secretary Scott Bessent & Fed Reserve Chair Jerome Powell called the CEOs of every major systemically important US bank to Treasury HQ on Tuesday to warn them about cybersecurity risks from Anthropic’s Claude Mythos model
Mythos found thousands of zero-day vulnerabilities across every major OS & web browser, including a 27-yr-old bug in OpenBSD that survived 5M passes by automated testing tools.
That’s a problem for Wall Street, where a single breach at a systemically important bank could cascade thru the financial system.
Anthropic launched Project Glasswing this week, giving roughly 40 companies access to Mythos for defensive security work.
13 are confirmed - Anthropic is keeping the full list & the model itself restricted because Mythos is reportedly "too dangerous for public release"
Alphabet
Amazon
Anthropic
Apache Software Foundation
$AAPL
Broadcom
$CSCO
$CRWD
JP Morgan Chase
Linux Foundation
$MSFT
Nvidia
$PANW
Anthropic aims to secure critical infrastructure before offensive actors develop or acquire equivalent AI capabilities:
- Primary goal - fix existing products
- Secondary goal - scan & harden new code before it hits production
- Third goal - share findings & best practices
Note: Anthropic plans public reporting within ~90 days