Google unveils advanced Gemini AI models, 'Preferred sources' feature
Technology
Google just dropped some big updates to its Gemini AI and Search tools.
The new Gemini 2.5 text-to-speech models let developers tweak voices for tone and expressiveness, making things like virtual assistants and audiobooks sound way more natural.
Other details:
Gemini 2.5 (in preview as Flash TTS and Pro TTS) also helps keep voices consistent across multiple speakers, all accessible through Google AI Studio.
On top of that, Google's teaming up with media outlets to test out AI-powered article summaries and audio briefings that link back to original sources.
Plus, the new "Preferred Sources" feature lets you personalize your news feed by picking which sites you want to see more from—so your feed feels a bit more like yours.