Google Sneaks a Microphone Into Its TV Box While Microsoft Teaches AI to Decide Faster
Just as the model arms race runs out of fresh ways to say "mine has more parameters," two companies reached for something more grounded: one is shoving a microphone into your living room, the other is teaching AI to think less.
Three key facts
Google's next-generation TV Streamer is in development, and the biggest change is that the device itself now has microphones. According to 9to5Google, the new unit has a microphone array cut into its slanted surface to pick up "Hey Google" commands directly, no longer relying entirely on the remote. Another new hardware capability is presence sensing: it lights up the Google Photos screensaver when someone enters the room and shows ambient info like the time and weather when you get close. The processor is faster, storage exceeds the current model's 32GB, and the remote gains two more shortcut buttons.
The same day, Microsoft launched a new AI model built for "making decisions quickly." According to The Verge, such models are meant to pick something fast and efficiently — like figuring out which AI model to use, or what's in an image being analyzed by AI. Microsoft claims its new tech beat rivals in tests measuring speed and accuracy. It isn't chasing "thinking deeply," it's chasing "answering fast."
Behind both is the same direction: pushing AI out of the chat box and into daily actions. Google welds a voice entry point directly into a living-room device; Microsoft stuffs a model into the "which path to pick" middle layer. Neither forces users to open a chat window anymore — both want AI quietly taking over choices in the background.
Wang Dou's Take
Start with Google's microphone. A few more pickup holes in the hardware lets the marketing copy say "more natural voice interaction," but the real experience is this: the living room now has something that is always listening. Apple's HomePod and Amazon's Echo have done this for years; Google just moved the trick from a speaker into a TV box — a box that sits on the TV stand, facing the sofa, more central than any speaker. The benefit is convenience; the cost is that the question "who's listening?" can never be dodged again.
Microsoft's ambition is craftier: it isn't fighting you for the chat box, it wants to be the middleman that "decides which AI to use." You ask one question, three models may run behind the scenes, and Microsoft picks the best fit fast. The clever part is that users can't tell which model answered — they only feel "this assistant is snappy." Once that middle layer becomes the default, Microsoft holds the gateway to every model without having to build the best one itself.
One is grabbing your ears, the other is grabbing your right to choose. Neither is cheap, and neither asks your permission.
Source: The Verge · 9to5Google