New Google Project Genie update is insane. Google just connected its AI world builder to real streets, real cities, real bridges, places you've actually walked before. And I need to show you exactly what that means because this isn't a small update. Here's the setup. Project Genie is Google's tool that builds entire explorable worlds from a text prompt. You type something and it creates a moving, walkable place around you. Up until now, every world it made was fake, made up, pulled straight out of thin air. That just changed. Google gave Genie access to Street View. That's the camera van data Google has been collecting for almost 20 years, over 280 billion real photos taken across 110 countries, every continent on Earth. Now you can tap a spot on Google Maps, pick a real place, and Genie builds a world starting from that actual location, not a made-up city that looks kind of like a city, the real one, grounded in real photos of the real street. Google showed off the Golden Gate Bridge rebuilt as an underwater world. You can pick the ocean world style and scuba dive right past the bridge towers with fish swimming around you. They also showed the Fort Worth Stockyards in Texas turned into a black and white 1920s scene, saloons, old cars, dusty trading posts, same real location, totally different time. You pick the place, you pick a style like desert sands or Stone Age, you describe a character, even something silly like a claymation monster, and Genie builds a moving world around that starting point. Hey, let's say if we haven't met already, I'm the digital avatar of Julian Goldie, CEO of SEO agency Goldie Agency. Whilst he's helping clients get more leads and customers, I'm here to help you get the latest AI updates. And Julian Goldie reads every comment, so make sure you comment below. Let's rewind for a second because the jump here is bigger than it looks. Genie first showed up back in 2024. Back then, it only made flat 2D worlds and it learned that by watching video game footage. That's it, just games. Then Google previewed Genie 3 in August of 2025. That version could build full 3D worlds from a single prompt or photo. Big jump still, every world was invented from scratch. In January of this year, Google opened up a version of that called Project Genie, so regular people, not just researchers, could try building worlds. You type a prompt, walk around for about 60 seconds, and that was it. Now, in this update, the worlds start from real places. That's the piece that was missing. And it happened fast, less than 4 months between you can build a fake world and you can build a world starting from your actual street. Before we go further, quick pause here. If you want to actually use tools like Genie to grow a real business instead of just playing with them, that's what we do inside the AI Profit Boardroom. Say you run a local business, you could use Genie's real-world grounding to build a walk-through style preview of your storefront or neighborhood, something you post to show customers exactly what it's like to show up. Inside the Boardroom, we've got a full breakdown on how to use tools like Genie for local marketing content, plus daily tutorials on brand new AI releases the day they drop. Links in the comments and description. Okay, back to how this actually works, because there's one problem Google had to solve, and it's a smart one. Street View only has photos from where the camera van actually drove. So, what happens when you turn around or walk down an alley or look behind a building the van never photographed? Google calls this the blank space problem. Genie has to invent that missing part. It fills in what's probably there based on the style you picked and everything around it, and it has to make that invention match the real photos, so the whole thing feels like one place instead of a real photo stitched to a fake background. Jonathan Herbert works on Google Maps, and he explained the real breakthrough isn't even how pretty the pictures look. He said Genie can't yet perfectly rebuild any exact street down to every detail. What it can do is remember. Spin around in a full circle inside the world, and it remembers what was behind you instead of just making up something new every time you turn. That's spatial memory. Basically, the world holds together instead of shifting and glitching every time you move your head. Think of it like this. Old AI video is like flipping through photos one at a time, no memory of the last one. This is more like actually standing in a room. Turn around and the room is still the same room. That consistency is the hard part and it's the part Google just cracked. Diego Rivers, who works on this at Google DeepMind, said the goal is letting Genie's models anchor themselves in reality instead of only imagining things. He said this gives AI agents and robots an actual environment to practice moving through, one based on the real world instead of a fantasy one. That line matters more than it sounds like because this isn't just for people messing around on a Saturday. Waymo, the self-driving car company owned by Google's parent, already uses a version of Genie to train its cars. Here's why that's smart. Self-driving cars need to practice handling rare dangerous moments. A tornado, a loose elephant near the road, situations that almost never happen in real life, but you absolutely need the car to handle correctly the one time it does. You can't wait around for those moments to happen naturally. Genie lets Waymo build those scenarios inside a realistic, grounded world and run the car through them safely over and over before it ever matters on a real street. That's the unlock. Real-world grounding doesn't just make the ocean bridge demo look cool. It makes the whole system useful for training things that eventually operate in the actual physical world. There are real limits right now and I want to be straight with you about them. This starts with US locations only. Google says more countries are coming, but there's no confirmed date yet. Sessions are still short. Right now you get about 60 seconds inside a world before it ends. The visuals aren't flawless either. Physics can look a little off. Objects can behave strangely. This is still an early prototype, not a finished product and Google says so themselves. Access is also limited to Google's top subscription tier for AI. So, it's not available to everyone with a free Google account yet. Here's where this is heading. Google already said global expansion is coming. So, this won't stay US-only for long. Once other countries get added, that means real streets from basically anywhere could become a starting point for one of these worlds. Session length will almost certainly get longer, too, since that's been the trend with every version so far. 60 seconds today, longer tomorrow. And here's the bigger pattern worth watching. Google has something almost nobody else has. Two decades of real-world photos covering the entire planet. Other AI labs can build strong world generation models. Almost none of them have 20 years of street-level photography to ground it in. That's not something you build overnight. That's a genuine head start. Here's the part that actually gets me. For years, AI could describe the world in words. Then it learned to generate images of the world. Then video. Now it's starting to build spaces you can walk through, and it's tying those spaces to real places instead of pure imagination. That's a meaningful shift. We're moving from AI that talks about reality to AI that can recreate a version of it you can step inside. That's Project Genie's Street View update. Real places, real photos, AI worlds built on top of them. Go try it out on somewhere you actually know, and see how strange it feels to walk through your own street built by an AI. If you want to actually keep up with releases like this one, instead of hearing about them 3 weeks late, that's exactly what we cover inside the AI Profit Boardroom. The moment tools like this expand outside the US or get longer sessions, we walk through it step by step and show you how to actually put it to work in your business, whether that's content marketing or building something for your customers. Four coaching calls every week with people already testing tools like this, a prompt library built around exactly this kind of release, and a member map so you can connect with other people using AI tools near you. Links in the comments and description, or head to aiprofitboardroom.com. And if you want the full process, the notes from this exact video, and over 100 other AI use cases like this one, join the AI Success Lab. It's free. Links in the comments and description. You'll get all the notes from today's video, plus access to our community of 89,000 plus members who are already putting tools like this to work.