48:15
Google I/O used to be about a whole lot of things. You used to come here at the Shoreline Amphitheater just minutes from Google's headquarters and see Google announce all sorts of things from Android to Chrome OS to Pixel devices and much more. But this year, there's only one thing on the menu, AI. I know people are seriously sick of AI and to be honest, I'm pretty sick of it, too. But despite the AI overload, there were quite a few cool things announced here and I'm going to show you five of them right now starting with a new thing coming soon to Android called Halo. So, to understand Android Halo, you first need to understand Gemini Spark. Spark allows you to run Gemini agents on a
virtual machine 24 hours a day, 7 days a week. Let's say you are really into Sprite and every day you want a report delivered to you that tells you if there's been any news on Sprite over the past 24 hours. How would you do that with Gemini today? You could do it, but you would need to have a browser window open on your laptop at all times for it to work. But with Gemini Spark, you offload that agent to a virtual machine so you get that Sprite report daily without needing to keep your laptop powered on at all times. That's a simple example, but the sky's the limit for what agents could do for you in Spark. You could even have one agent interacting with another agent
to perform truly complex tasks. But if you do that, how will you monitor them easily? That's what Halo does. On your Android phone, Halo will be a little Gemini icon that pops up in your status bar. You'll be able to interact with it and see what your Gemini agents are up to. There might even be a message from an agent that you need to respond to. So, when will we be able to do this? Spark is coming to trusted testers now and will be available in version for Google AI Ultra subscribers in the US next week. If you're not an Ultra subscriber because you think it's too expensive, stay tuned because I have
good news on that front for you. Anyway, Google Spark is also going to see a wider rollout to everyone in the coming weeks. Meanwhile, Android Halo is coming later this year with Android 17. Don't be surprised if we see it as a Pixel 11 exclusive, at least for a little while, but I could be wrong on that. Speaking of the Pixel 11, Android Halo could even have more advanced capabilities on devices that support Gemini intelligence. There's a lot of controversy going on right now about what current devices will support Gemini intelligence and what that means for devices that don't. Stay tuned for more on that coming soon, but for now, let's talk about something everyone watching this uses, YouTube. So, if you go up to
the search bar on YouTube, it's pretty simplistic. You type in the keywords and it finds videos that match those keywords. But, YouTube is going to get AI-powered search through a new feature called Ask YouTube. If you've ever used things like Ask Maps, this is very similar. Instead of surfacing a list of videos that match the keywords you're looking for, Gemini can really dive deep into the info it thinks you want and provide not only multiple videos on that topic, but specific timestamps to the things you want to know about. It organizes it nicely for you and even allows you to continue the search through more prompts. This makes a YouTube search more like a conversation and less like a simple show me videos
about this one thing. Ask YouTube is coming to the United States later this summer. There's no word yet on a wider rollout, unfortunately. Another Google product every single person watching this uses is Google Search. Google is changing this in a pretty significant way by bringing the power of AI mode right into the regular search bar. Honestly, the day Google announced AI mode, I knew this was coming and I'm super glad to see it. Now, you won't need to do this yourself. You won't need to say, "Is this a question for a regular search or something for AI mode?" Now, you just start typing into the same Google search box you've been using for 25 years. The box will expand
to accommodate your long prompts and give you the ability to skip right into AI mode. Of course, if you don't want to use AI mode, and I'm sure a lot of you don't, you can and do a typical search. You'll still have AI overviews there, which you can't avoid, but I want to make sure it's clear that you don't need to use AI mode. But, Google really wants you to use it. So much that now, within an AI overview, you can quickly transition to AI mode if you need to. If your search is very complex, Gemini will even help explain it by dynamically creating graphics, tools, and even short video clips. Once again, you don't need to tell search to do this. It just does
it once it realizes this will be the best way to get you the information you're looking for. The new AI search box is starting to roll out today in all countries and languages where AI mode is available. And the ability to create those graphics is coming soon for global audiences. Okay, so this next thing is something a lot of you watching this have done. Using Nano Banana in Gemini to create an image. But, what happens when the image is created and you like most of it, but not all of it? Well, you need to then completely remake that image from scratch again and hope Gemini gets it right. With Google Picks, you'll actually be able to edit that generated image in the same way you would do so in
something like Photoshop or Illustrator. You can select items and change that individual item, highlighting text and change it manually, and more. Eventually, you'll be able to resize images and then move them around the canvas, but that's not coming until later this year. When you do each of these actions, Gemini takes note of them in a running list, almost like a list of comments from one designer to another. When you're ready, Gemini will recreate the image based on those comments, bringing you the exact image you want without all the guesswork or needless re-prompting. Google Picks is launching
this week to a limited group of trusted testers. This summer, it will roll out globally to Google AI Pro and Ultra subscribers. Speaking of being a Gemini subscriber, Google is changing how that works. There is still a completely free tier, so don't worry about that. But, if you're willing to pay, you can unlock new features in Gemini through the Plus and Pro plans, both of which still cost the same as they did before, which is $8 per month or $20 per month. If you wanted to go to the highest ultra tier though, you would have had needed to fork out $250 per month, which is just crazy talk for most people out there.
Now though, Google is introducing an ultra plan that is far less expensive at $100 each month. That's still a lot of money, don't get me wrong, but it's definitely a lot better than $250. And if $250 isn't crazy for you, then you'll also love to know that Google dropped the price of that top tier to just $200. That means you have more choices for your Gemini subscription, and even if you go all the way to the very top, you'll be paying less than you did previously, which is always good news. Obviously, a lot of people watching this will never pay for a Gemini subscription because they don't want or need it. That's very understandable. On that note, I want to close this video out by
talking about the worst thing announced at Google I/O, which is the new Gemini Omni within YouTube Shorts Remix. This allows you to feed a short you or someone else's created to Gemini and have it recreated in various styles. These two examples are the ones that Google used to demonstrate the feature, and I won't lie, all of them are awful. If this AI slop is what Google thinks would get us excited about this feature, I really don't know what the company was thinking. If you don't agree and think this is cool, remixing with Omni is rolling out today at no cost in YouTube Shorts Remix. So, yeah, go have fun with that. Let me know in the comments what your favorite Google I/O feature was for 2026.
