Google I/O 2025: Three highlights that Apple should pay attention to

Similar to Microsoft's Build Keynote, Google's kickoff was a kind of Ai-Palooza for his annual developer event. If you are not finished with AI, it was probably a sleeping body, and I suggest you skip to the next post. However, I am not sure whether it is a good long -term plan.

From striking moon shots to real, implementable and useful functions that customers can use almost immediately, there was a bit of everything. Inclusion of things, from which Apple users will also benefit greatly.

Long -standing Google I/O observers know that the event marks the annual ritual of Google that can withstand as many projects as possible within the limits of an average human bubble. (To be fair, it has gotten much better in recent years.)

This time Google gave an insight into the way things could look like what things could look like if Apple hadn't missed several boats on this technology in the course of about two hours of AI and 30 mentions of agents (or “Agentic”).

Of course, many of Google's offers would make no sense if they came from Cupertino. As interesting as Google Beam may be, it would probably not work as a Facetime -Ultra device. On the other hand, some of these announcements would be quite fantastic if they had been made by Apple:

1) Gemini in chrome

This was probably the lowest fruit of all Google offers, since all other AI products that were already available. But it may be one that will be the most effective over time.

By embedding Gemini directly into the browser, Google transforms it into a easily accessible smart assistant, which understands both the content of the page and the user's context (calendar, etc.). Voice support is a welcome plus, especially for users who rely on failure functions.

I know that it sounds clichéd, but this will probably unlock an interesting new way of surfing on the web. At least until the doj Google asks to tear off Chrome.

2) Project Astra Update

Last year we introduced the Astra project, our research prototype of a universal AI assistant. Since then we have improved the memory, computer control and improved voice output and are working on bringing these new functions for Gemini live and other products.

To be honest, the comparison between this video and the entire decades of Siri blunder speaks for itself. Last year Siri seemed hopelessly back. In view of the latest reports, it seems further behind the competition, since the competition actually progresses in the past twelve months.

It is very fair to mention that the Astra project is still considered an experiment. It will probably take a while for it to become an official, widely accessible function. But still I mean … only … yes.

3) stitch

As with Microsoft-centered announcements that I commented here, Google I/O also had a share of tools and improvements in the quality of life for developers.

One of them was Stitch, a profoked platform from Gemini 2.5, which allegedly creates complex user interfaces with frontend code (which obviously needs human optimizations) from image input and user requirements.

This is important because the consistently beautiful but practical user interfaces have never been a strong suit from LLMS. If Google actually delivers this promise, this could push again to use the long -horned goal of Apple, Swift, Xcode and its entire development platform: reducing the entry barriers so that more people can express themselves through code.

And you? Has anything secured in Google I/O? Let us know in the comments.