I’ve been playing around with Gemini 3 for several days now, and my obsession with it is no less than when I first got into CryptoKitties back in December 2017, which kickstarted my Web3 journey. Gemini 3 absolutely nails consistency and precision. Before this, I used GPT and Grok to try and make some backyard renovations—like adding some rock gardens, stones, and replacing the ground under the pavilion with bluestone. But those tools couldn’t execute the instructions accurately and would mess up other parts of the image. Gemini 3, on the other hand, is spot-on. Whatever you point at, it delivers. In the future, if you want to redesign your home, you won’t even need to hire a designer. This is just one small example of its practical use in daily life. This week, my workplace hosted a big event, and we needed a special webpage to feature it on our homepage. The designer spent two days on it, but I wasn’t satisfied. So, I just used Gemini to whip one up myself and handed it over to the operations team to use. Cases like this, I think, will become more and more common in both work and life. With the launch of the Gemini series, I feel like my AI journey has truly begun. Overall, Gemini 3 is flawless when it comes to images, web pages, and even simple functional systems (like the file management system I built). But I do have one big question while using it: When I’m making games, as soon as it involves characters, it always gives me these rough pixelated figures, and they’re stuck at 64x64 resolution. How is everyone solving this? That said, have you seen the Heroes of Might and Magic remake someone did with Gemini 3 in their video? It’s absolutely amazing~