🧠 Google Just Broke the Language Barrier

Fable 5 is Anthropic's newest model, and it's in a different league from anything they've publicly released before. It tops nearly every benchmark tested, with standout performance in software engineering, research, vision tasks, and complex knowledge work. The longer and harder the task, the bigger its edge over previous Claude models.

Here's what makes this launch unusual: Anthropic was genuinely concerned about releasing it. Fable 5's capabilities in areas like cybersecurity are strong enough that misuse could cause real damage. So they shipped it with a conservative safety layer — queries on a narrow set of sensitive topics get rerouted to Claude Opus 4.8 instead. It'll catch some harmless requests too, but Anthropic says that happens in under 5% of sessions on average.

For a small group of vetted cyber defenders and infrastructure providers, there's also Claude Mythos 5 — the same underlying model with some safeguards lifted, deployed through Project Glasswing in collaboration with the US government.

Pricing: $10 per million input tokens, $50 per million output tokens. Less than half the cost of Mythos Preview.

Google just shipped Gemini 3.5 Live Translate — a new audio model that translates spoken conversation in near real time across 70+ languages, without any manual language selection. You speak, it detects the language, and translated audio plays back within seconds, preserving your intonation, pace, and pitch.

What sets it apart from older translation tools is that it doesn't wait for you to finish speaking to start translating. It processes speech continuously, so conversations flow naturally instead of stopping and starting.

It's live now in the Google Translate app on Android and iOS. For Google Meet, it's rolling out in private preview for business Workspace accounts this month and expanding 5-language support to 70+ languages and 2,000+ language combinations. Developers can access it via Google AI Studio and the Gemini Live API today.

Amazon has added a new AI design feature to Alexa for Shopping. Type a prompt into the Amazon app or Amazon.com, and Alexa generates a custom graphic you can print on t-shirts, hoodies, tumblers, water bottles, and more — no design experience needed.

The feature runs on Amazon's Merch on Demand service, so once your design is ready, Amazon handles printing and ships it with Prime delivery. You can also share the design directly with friends or family so everyone can order the same thing.

It's live now in the Amazon Shopping app. Just tap the Alexa icon or search "customize" to get started.

If you drop a new file into a folder on your Mac, Apple Intelligence will now suggest a filename based on how you've named other files in that folder and the content of the document itself.

It's a small feature, but it quietly saves the mental overhead of staying consistent across hundreds of files. Best for anyone who manages a lot of documents and has ever spent 10 seconds wondering what to call something.

iOS 27 adds a feature where the Phone app can automatically surface relevant information from your other apps during a call.

Call an airline to change a flight, and your confirmation code from Mail appears on screen before you even ask. The app pulls context from across your device so you have exactly what you need, right when you need it.

This is coming with iOS 27 later this year — worth keeping an eye on.

If you use AI coding tools like Claude Code or Cursor to build web products, you've probably noticed the outputs can look generic. Uiverse Design fixes that.

It's a library of framework-agnostic CSS design systems with a DESIGN.md file that tells your coding agent exactly how to style your product — fonts, colors, components, the whole visual language. Drop it into your codebase and your AI builds consistent, polished interfaces automatically.

Here’s a quick overview on how to use this.