Glasswing: Anthropicโs Curse and Cure
Anthropic announced its new cybersecurity initiative, Glasswing. The findings from their new model, Mythos Preview, are at the center of the project. Anthropic says the model performs extremely well in cybersecurity but is too risky for public access right now. According to the company, Mythos Preview found serious vulnerabilities in widely used operating systems and browsers. For now, access will be limited to Glasswing partners such as AWS, Apple, Google, Nvidia, and Microsoft. Anthropic also indicated that it is committing $100 million usage credits for Glasswing and will donate $4 million to open-source cybersecurity initiatives.
So, Mythos Preview appears strong enough that it will stay closed until key issues are patched and Anthropic finds a safer rollout path. It is notable to see an AI company take responsibility for potential risks and act before larger problems happen.
OpenAIโs Surprise Media Move
OpenAI announced that it acquired TBPN, a daily tech talk show with a loyal audience. The company says the goal is to increase global engagement around AI and contribute more directly to the conversation.
Tthis move can also be read as a strategic step before a possible IPO: tighter control over AI messaging that strongly shapes public demand and perception and a better chance to steer that narrative in its favor.
SpaceX Prepares IPO Filing
SpaceX is preparing to file for an IPO at a $1.75 trillion valuation. If it goes public at this level, it would become one of the most valuable companies in the U.S. on day one.
After its merger with xAI, SpaceX is no longer just a rocket company. Starlink revenues, xAI, and plans for space-based data centers are now part of this valuation. In other words, investors would not only be paying for today’s rocket and internet business, but also for future promises that are not guaranteed yet.
