It's Monday. Here's what changed in AI this week, and what it means for your design work:

Writing this from Paphos, Cyprus. Locked in.

This week was all about content. I posted a series of Claude Code videos for designers and the response has been huge on Instagram, TikTok, everywhere. What I've noticed is the shift is really happening. A year ago, talking about AI as a designer got you weird looks. Now designers are actively asking to learn it. The demand is real and the perception has completely changed.

Just before the month ends, I'm working on something big for Stage the project management tool I built for designers. A bunch of new AI features are coming to help you streamline your client projects so you can focus on designing only. More on that soon.

Outside of work, I'm preparing for my first HYROX next month. That means a lot of running - which I don't love - but I'm going to crush it.

This might be the most significant AI safety move of the year - and it's not even a product launch.

Anthropic just published a 244-page system card for Claude Mythos Preview, their most powerful model ever. But here's the thing - they're not releasing it. Not to developers. Not to the public. Nobody.

Instead, 12 of the biggest tech companies - Apple, Amazon, Google, Microsoft, Nvidia - are using it exclusively for defensive cybersecurity under a program called Project Glasswing, backed by $100M in Anthropic credits.

The numbers speak for themselves. 93.9% on SWE-bench. 100% success rate on all 35 CTF security challenges. It autonomously discovered thousands of zero-day vulnerabilities, including a 27-year-old OpenBSD bug that no human had found.

This is going to have a massive impact on how we design using AI. When the model that writes the code, generates the assets, and builds the interfaces is this capable - everything changes. The bar for what AI can do in a design workflow just moved from "helpful assistant" to "autonomous problem solver."

Anthropic didn't just say "this model could be dangerous" - they built real safeguards before releasing it. First time we've seen a company write a system card for a model and then decide not to ship it. That tells you how powerful this thing really is.

For designers, pay attention. If AI can autonomously find security vulnerabilities in every major tech stack, imagine what the next public release will do for design and development. The tools you use, the platforms you build on - they're all about to level up.

You can't use Mythos - yet. But here's what you should do right now: start learning Claude Code. Claude Opus 4.6 is already incredibly powerful for designers, and tools like Codex are catching up fast. The designers who learn these tools now will be ready when Mythos eventually ships to the public. Don't wait - read the system card and start building with what's available today.

You describe what you want. Wayframe's AI agent writes production-ready React components and renders them live on a visual canvas. Not a mockup - a real, working app.

Now powered by Claude Sonnet 4.6, the agent understands design taste at a level that actually makes sense. Select any element, type what you want changed, and the agent rewrites the component in place. No context switching, no copy-paste.

I found a GitHub repo with the design systems of Stripe, Apple, Linear, Figma, SpaceX and 53 other leading companies. And you can use any of them to build your next website.

Before you go we’d love to know what you thought of today's newsletter to help us improve The Logiaweb experience for you.

💌 Know a designer who should be using AI smarter?
Forward them this email. Or just send them to logiaweb.net to join.

See you next Monday,

Say hi on X, Instagram or YouTube