This is a structure designed from scratch 160 days ago, a rendered model that existed only in software. And this is a real humanoid robot that my two partners and I completed 160 days later. It started as an somewhat unrealistic idea. Over time, this idea gradually took shape, with the robot’s structure and systems continuously refined and evolved, until it finally appeared before us in physical form. At the beginning, the question friends asked most often was, “What’s the point of building this thing?” Of course, the underlying implication might have been, “Can you really pull it off?” The starting point for answering this is quite simple. In recent years, we’ve witnessed the entire journey of AI—from its initial release, through its development, to becoming a mainstream focus. During this process, I increasingly felt that we are undergoing an industrial revolution unique to our era. Meanwhile, alongside AI’s advancement, Tesla released Optimus, and Unitree showcased its G1 and H2 robots performing on the Spring Festival Gala. As you can see, numerous humanoid robots are truly entering reality. So I began asking myself: In this age of intelligence, as an ordinary person, am I destined to remain merely a spectator? How should we engage with this era? Could I also find my own entry point to create something with my own hands, rather than just witnessing it happen? Once we truly got started, we realized we might have underestimated the term “embodied humanoid.” It turned out to be slightly harder than imagined. Yes, slightly harder. There were moments when we stayed up until dawn, looking at parts scattered all over the floor, genuinely questioning whether this endeavor was worth it. After all, if fixing bugs consistently generates new ones, your hair starts falling out way too fast. Design, assembly, debugging, failure—the cycle repeated endlessly. We stumbled into countless pitfalls. Due to space limitations, we plan to share these experiences with you in detail later. But here’s the good news: Finally, one day, it started resembling a human. With its legs functional, it could stand. With its arms operational, it could wave. With its “brain” activated, it could speak. “I am DreamBot, an embodied prototype robot.” Only when the complete torso was installed did it cease to be just a pile of parts and truly stand before us as a humanoid robot. This experience made me realize that many things are not as unreachable as they seem—even this one. Currently, our latest progress includes upgrading the full-body structure and achieving control across all joints. In simulation, it can now deploy policies and walk stably. Next, we will continue debugging to enable smooth, steady walking, while documenting its progress along the way. If you’ve watched this far, perhaps you’re as curious as I am about its future development. Consider following us, and let’s complete this journey together. Our goal moving forward is to lose fewer items. Fewer, not none. Well, maybe not selectively losing them either. [Note: This line can be moved to the end credits.] Moving forward, we’ll aim to lose fewer items—fewer, not zero.