Project Glasswing

• Anthropic PBC kondigde dinsdag Project Glasswing aan, waarbij Amazon, Apple, Microsoft, Cisco en anderen toegang krijgen tot het nog niet uitgebrachte Mythos AI-model om fouten in producten te identificeren voorafgaand aan een bredere release. • Het initiatief is erop gericht potentiële cyberaanvallen vanuit geavanceerde AI-systemen te beperken door deelnemers hun bevindingen met sectorgenoten te laten delen. • Dit weerspiegelt de groeiende voorzichtigheid binnen de sector over de cyberrisico's van krachtige AI-modellen en markeert een proactieve samenwerking tussen toonaangevende techbedrijven.

• Anthropic announced Project Glasswing, a collaborative initiative involving Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks to secure critical software infrastructure. • The initiative provides select firms access to Claude Mythos2 preview model for defensive cybersecurity work, which has reportedly identified thousands of severe vulnerabilities in major operating systems and browsers. • The program aims to give defenders a durable advantage in the AI-driven cybersecurity era while controlling access to prevent misuse by attackers.

• Anthropic launched Project Glasswing, collaborating with Apple, Google, and Microsoft to deploy Claude Mythos Preview in identifying and patching critical software vulnerabilities preemptively. • The initiative unites rivals to leverage AI against cyber threats targeting US software infrastructure. • By fixing flaws before exploitation, the project aims to bolster national cybersecurity amid escalating AI-driven attacks.

• Anthropic announced Project Glasswing, a cybersecurity initiative bringing together AWS, Apple, Broadcom, Cisco, CrowdStrike, Google, Microsoft, Nvidia, and Palo Alto Networks to identify and fix software vulnerabilities using its Claude Mythos Preview model. • The Claude Mythos Preview model achieves 93.9% accuracy on SWE-bench Verified, significantly outperforming Opus 4.6's 80.8%, and demonstrated the ability to find thousands of hidden bugs in critical systems. • Anthropic is committing up to $100 million in usage credits for Project Glasswing, plus $4 million in direct donations to open-source security organizations to strengthen cybersecurity infrastructure.