It’s time for Google IO 2025 and of course it’s AI all the way down. Here’s today’s AI news. Gemini is getting properly integrated into Chrome Google want to turn your browser into a full on AI assistant that can answer all kinds of questions for you. This will include understanding your personal history and connecting to other Google apps.
[00:19.3]
Gemini Live will get a similar treatment. Rather than being constrained to its own environment, it will interact with your whole Google suite. Google Meet is getting an amazing new Live translation mode. Speak naturally in your own language and your words are translated in near real time while preserving your tone, voice and expression.
[00:38.3]
It’s available in English and Spanish with more languages coming soon. On the AI model front, Google are introducing a new ‘Deep Think’ Mode to Gemini 2.5 Pro, the enhanced reasoning mode. You uses the latest cutting edge research in reasoning including parallel thinking techniques and Google claimed to have seen incredible performances from it.
[01:00.5]
Google also announced a Gemini Diffusion model. This state of the art text diffusion model doesn’t predict text directly. It generates outputs by refining noise step by step. This helps it excel at coding and math where it can iterate over solutions at incredible speed.
[01:17.3]
As anticipated, the new video model Veo 3 was announced. Of course it boasts a massive step up in video quality over Veo 2, but it also integrates generative sounds. This means that with a single prompt you can create talking characters and add sound effects to your videos for seamless audio visual creativity Hand in hand with Veo 3 was the announcement of a new video storyboarding and editing tool called ‘Flow’.
[01:39.0]
With Flow, you can combine assets together to generate the perfect clip and then roll straight into the next scene while maintaining character and visual consistency from the previous clip. It looks incredible and we think it was the star of the show. Also on the generative audio front, Google announced that they were extending access to their music AI sandbox powered by the latest song creation model Lyria 2.
[02:01.5]
This might easily rival the likes of Suno if Google can get the tools right. Rolling out Today in the US Google unveiled a new tool that lets you try on clothes virtually. You can upload a picture of yourself and then see yourself in a range of outfits. It will be interesting to see if this remains as an experiment or if it really catches on.
[02:20.5]
On the hardware front, there were several big announcements. Android XR is a mixed reality system that’s been designed to operate a range of wearable devices from lightweight glasses to full on VR headsets. That’s right, Google glasses are back. We featured a preview of this a few weeks back the new glasses are designed to be an extension of your phone and integrate AI directly into your audio and visual senses.
[02:46.7]
On the VR side, Google previewed Samsung’s Project Muhan VR headset. This is the first Android XR device and it will be available later this year. It will be interesting to see how they stack up against the Meta Quest offerings.
[03:01.7]
And finally, there was a crazy looking hardware project called Google Beam that can transform 2D video streams into a realistic 3D experience. This multiple camera streaming device boasts near perfect head tracking down to the millimetre and can run at 60 frames per second.
[03:17.5]
It’s the ultimate 3D Zoom experience.