It was risky but today’s Google I/O keynote burst into life when a live demo showed what we can expect when we get our hands on Android XR Glasses. The event was almost entirely about AI and search so it was no wonder most of the latest Android news was relegated to a separate presentation last week. Most, but not all. The live demo of glasses was worth the wait. There was so much to absorb from the presentation, it will take weeks to analyse. Let’s get started.
ANDROID XR
We got a sneak peek at Android XR at a recent TED Talk but this live demo was next level at Google I/O. We saw a live feed from backstage with the glasses using Gemini to interact with Nishtha Bhatia. Nishtha’s glasses streamed what she was seeing, plus her interactions with AI. It remembered the coffee brand she was drinking and took a photo of the crowd while on stage.
The biggest risk was a live translation demo. Android XR boss Shahram Izadi faced Nishtha speaking in Farsi while Nishtha replied in Hindi. The conversation was translated into English at the same time. It started well but the second statement from Shahram did not compute and the demo ended early. It was good to see Google take a punt and it actually worked well enough to show this is viable technology. Live demos add drama to keynotes and the rewards are huge it the tech works.
Google stressed these are still prototypes and developers won’t get to start working with them until later this year. I also picked up Shahram saying the in-lens display is optional. So there’s a lot more work to be done but these are the glasses I’ve been waiting for – no doubt. If you want to look stylish, you’ll be able to choose frames made by Gentle Monster or Warby Parker.
AI MODE SEARCH
Google introduced a new search tab called AI MODE, it’s available now but only in the United States. It follows growth in the US and India from AI Overviews, the little summary you see at the top of search results. Google claims it’s one of the most successful launches in search in the past decade.
“Our Gemini models are helping to make Google Search more intelligent, agentic, and personalised,” said Sundar Pichai, Google CEO.
“Since launching at I/O last year, they’ve (AI Overviews) scaled up to over 1.5 billion users every month month, in more than 200 countries and territories. As people us AI Overviews, we see they are happier with their results and they search more often.”
Google Lens has grown 65% year-over-year, with more than 100 billion visual search this year alone. Google presented stats that show users are making longer queries and more of them. AI is clearly having an impact when it comes to search. Accuracy is the key and Google is satisfied that it’s doing a good job in that regard.
The user interface will change to suit the results. It may present more text, more graphics, maps and/or images depending on your query, You can ask follow up questions in a conversational style and that’s how it pulls you in. Before you know it, you’ve spent ten minutes chatting with Google search.
GEMINI
Gemini is going to be everywhere. It’s also going to be proactive. It will know what to do without you saying a thing. Personal context means AI gets to know you across all of your Google apps. J.A.R.V.I.S. from Iron Man is a step closer.
More Gemini developments:
- New capabilities to 2.5 Pro and 2.5 Flash: native audio output for a more natural conversational experience, advanced security safeguards, and Project Mariner’s computer use capabilities. 2.5 Pro will get even better with Deep Think, an experimental, enhanced reasoning mode for highly-complex math and coding.
- Thought summaries in the Gemini API and in Vertex AI for more transparency, extending thinking budgets to 2.5 Pro for more control, and adding support for MCP tools in the Gemini API and SDK for access to more open source tools.
- 2.5 Flash is now available to everyone in the Gemini app, and we’ll make our updated version generally available in Google AI Studio for developers and in Vertex AI for enterprises in early June, with 2.5 Pro soon after.
GOOGLE BEAM
It’s not a hologram but Google Beam is a video communications device designed to make calls 3D from any perspective. Using its AI volumetric video model, Google transforms standard 2D video streams into realistic 3D experiences, allowing you “to connect in a more natural and intuitive way” – sort of like being there I guess.
“An array of six cameras captures you from different angles and with AI, we can merge these video streams together and render you on a 3D light field display,” explained Pichai.
It performs millimetre accurate headtracking at 60 frames per second. Google Beam is clearly aimed at enterprise and HP will be revealing more in coming weeks. The first devices will be available later this year.
FLOW
This is where the magic is made. Flow is a new AI filmmaking tool Google says is “built with and for creatives” and it has massive potential. Using Veo, Imagen, and Gemini, you can produce consistent clips to make a movie. All the special effects and world building can be based on images you upload plus your imagination via text prompts.
Currently, you need to add sound effects and dialogue in a video editor but soon you’ll be able use AI to generate that as well. Veo 3 lets you add sound effects, ambient noise, and dialogue to your creations – generating all audio natively. But it’s locked up in a US-based plan so the rest of us need to keep doing it the old way.
Although this technology is a long way from Hollywood quality it will, at some point, become the way we make movies. Google is saying all the right things now by engaging filmmakers as it develops AI video. How long that lasts is the big question.
SHOPPING – TRY ON FEATURE
The other live demo that wowed the crowd involved a Try On feature for clothes. Vidhya Srinivasan, VP/GM, Advertising & Commerce at Google, showed how AI Mode uses its knowledge of 50 billion product listings to take the pain out of shopping.
In this example Vidhya used the try-on feature to virtually wear a dress she liked. Google made a special model to work out how clothing would fit on a person. By uploading a photo of herself, the try-on feature perfectly matched the dress to her body. AI is working out a lot of physics here. It is not a generic match, it breaks down the fundamentals of your body shape and the dynamics of clothing, in this case a dress.
From there it’s a trivial process to purchase the item.
SO MUCH STUFF
There is so much more to go through from Google I/O. Over the coming days I’ll be looking a big deeper into the Keynote and other aspects from the developers conference.
So keep coming back for more.
