Google's Universal Cart & AI That's Always Listening

This year AI is the one doing the browsing.

Google wants to be your one shopping cart to rule them all

Google's Universal Cart is now live in the US, accessible through its Gemini app and Search. The idea: you shop across multiple retailers, but instead of juggling separate checkouts at Nike, Target, and wherever else, everything lands in one Google-managed cart. This is Google using AI to insert itself as a permanent middleman between you and every purchase you make online, which is enormously valuable to them because it means they own the purchase data, not the retailer. This will scale. Google doesn't launch things like this without the infrastructure already in place, which is exactly what makes it worth paying attention to — it's one of the more significant power grabs in retail in years.

https://www.theverge.com/tech/1011797/google-ai-universal-cart-shopping-tool-launch

Amazon's always-on AI wearable just got more useful and more invasive

Amazon acquired Bee last year — Bee made a $50 clip-on device that listened to your conversations all day and gave you summaries and to-do lists from them. Now Amazon is relaunching the product as two new devices: the $74.99 Bee Frontier wearable and a credit-card-sized recorder that sticks to your phone. Both integrate with Alexa and, naturally, Amazon shopping. The pitch is that your AI assistant actually knows what's happening in your life because it was listening when your dentist mentioned a follow-up appointment or when your colleague asked for that report. The reality is that Amazon, a company that sells advertising and products, now has access to the raw audio of your daily life, and you are paying them for the privilege.

https://www.theverge.com/tech/1011475/amazon-bee-frontier-eclipse-price-specs

That adorable robot toy is sending your kid's secrets to China

A robot companion called Eiliko, made by a company called Energize Lab, has a cute digital face that reacts to your mood and carries on AI-powered conversations. Security researchers found that everything you say to it travels to servers in China, stored in a way that is poorly secured and potentially accessible to bad actors. This is the same category of problem as CloudPets in 2017, when a stuffed animal exposed millions of children's voice messages. The device looks harmless, the company is obscure, and they almost certainly won't answer honestly if you ask where the data goes — which means the only real move is not buying it at all. If someone on your gift list has an AI companion toy on their list, that's the question worth asking before you check out.

https://www.theverge.com/tech/1010727/eiliko-robot-hack-energize-lab

Figma is laying the groundwork for AI agents that design alongside you

Figma — the design tool that most product, marketing, and UX teams use to build interfaces — announced new features defining how AI agents (autonomous software programs that can take actions, not just answer questions) will work inside its platform. They're also releasing "code layers" in open beta, which bridges the gap between a visual design and the actual code a developer writes. What used to mean three rounds of redlines between a designer and an engineer — tweak the padding, export the updated spec, wait for the build — starts collapsing into a single handoff. Teams shipping features in two-week sprints may be doing it in one. The people who figure that out first won't need to hire for the role that used to sit in between.

https://techcrunch.com/2026/10/14/figma-announces-new-features-to-define-how-agents-work-on-its-platform/

A startup wants AI to design the chips that run AI

Ricursive Intelligence is building AI systems that design semiconductor chips — the physical hardware that makes all computing possible, including AI itself. A chip shortage is one of the main things slowing AI development right now, and anything that accelerates chip production accelerates everything else downstream. The chip design process today takes years and armies of specialized engineers; the premise here is that AI can do it faster and learn from each iteration to improve the next design. If this works even partially, the two concrete implications are a shorter timeline to the next generation of AI models and less leverage for the handful of countries and companies that currently control chip fabrication. Either of those would matter enormously. Both at once would be a different world.

https://techcrunch.com/2026/10/14/ricursive-intelligence-is-betting-on-ai-designed-chips-to-help-solve-the-compute-crisis/

The through-line today is listening: Google listening to your shopping habits, Amazon literally listening to your conversations, Eiliko listening to your children. At some point the question stops being whether you're comfortable with that and starts being whether you even know it's happening.

← More daily briefs