Post by Supercharge

This Week in AI – Week [15] Welcome back! The theme this week is AI breaking out of the lab and doing some extremely heavy lifting in the real world. Anthropic is teaming up with tech giants to deploy "Claude Mythos" (an unreleased model capable of autonomously hunting zero-day vulnerabilities) to proactively secure global software infrastructure before malicious actors can exploit it. Meanwhile, Sam Altman's prediction of the hyper-efficient, AI-driven company has officially arrived: a two-person telehealth startup just casually hit $1.8 billion in sales. Top it off with a massive wave of agentic coding updates and OpenAI actively proposing a new "AI tax" to fund human safety nets, and it's clear the sci-fi future is already our present reality. The highlights: 🛡️ Anthropic launched Project Glasswing 💰 A two-person telehealth startup reached $1.8 billion in sales 💾 Milla Jovovich open-sourced MemPalace, a hyper-efficient AI memory tool 🧠 Anthropic discovered distinct emotion-like states inside Claude 4.5 ⚖️ OpenAI proposed a tax on advanced AI systems to fund social safety nets Click to read the detailed bulletin with all the links and get notified by subscribing to our AI news! Edited by Adam Kovacs | Reviewed by Adam Almasi