

At the end of October 2025, 1X Technologies' humanoid robot Neo swept through the tech world like a heatwave. This sleek robot, backed by OpenAI, is touted as the first truly home-friendly physical assistant.
Priced at around $20,000 or $499 per month for leasing, Neo can clean, carry items, and even learn new tasks through imitation. In just a few days, it became the internet's focal point — seemingly, a tireless family companion has finally arrived.
Yet, behind the cheers, a profound reflection on "autonomy" quietly unfolds. Remote control offers the illusion of convenience, but it exposes a core pain point in the AI industry: human operators still lurk in the shadows, and what happens to your privacy data?
As Curious CEO David Tomasian puts it:
"True autonomy is the only way machines can belong to us."
Neo's launch is indeed exhilarating: standing 5 feet 6 inches tall and weighing 66 pounds, it uses tendon-driven actuators mimicking human muscles, wrapped in a soft shell for safety. Hugging Face co-founder Thomas Wolf exclaimed on X that Neo has "advanced" his timeline for home robot adoption.
In demos, Neo waters plants, opens doors, washes clothes, and scrubs dishes, turning mundane chores into something poetic and efficient.
But this excitement was quickly doused by reality. The Wall Street Journal's hands-on report reveals that many of Neo's movements are still remotely controlled in real time by "experts" via VR.
This isn't sci-fi — it's the current state of AI, where remote piloting aids companies in training models through imitation and reinforcement learning, yet reduces the robot from "independent helper" to "human extension."
Tomasian sharply notes that under this model, your "private robot" isn't truly private: it not only observes your life but uploads data to the cloud, fueling the manufacturer's training.
When a robot can "see" your home layout, recognize your voice, and analyze your habits — yet remains tethered to the manufacturer's servers — who does it really belong to?
The wave of humanoid robots is flowing from factories to living rooms. Figure AI's Figure 02 and Tesla's Optimus aim to reshape industry, while Neo pushes the vision into consumer territory — not just productivity, but companionship itself.
This trend is especially urgent in elderly care. Pilot projects in Japan, Korea, and parts of Europe are testing robots for assisting daily activities, monitoring health, and providing emotional support. But Tomasian points out: "The difference between aid and true care lies in understanding context and emotion." If data isn't encrypted and stored locally, "the robot isn't yours—it's someone else's lens."
Privacy expert Kohei Kurihara disclosed on X that Neo users must sign a waiver allowing manufacturers access to certain operational data. This "tech-for-convenience" pact hides cracks in trust. A Medium article bluntly states that this $20,000 robot "needs a human babysitter", with complex tasks requiring an appointment for "expert mode," making users feel like they're renting a "surveilled puppet."
Tomasian emphasizes that for embodied intelligence to evolve like language models, three things are essential: on-device reasoning, multimodal understanding, and encrypted autonomy. AI must not just execute commands but comprehend "why" they are given, ensuring data sovereignty belongs to the user. True care reliability stems from security and privacy, not algorithmic complexity. In other words, autonomy isn't just a technical issue — it's a social contract: Machines should embody trust, not extend surveillance.
Neo's controversy reflects a deeper trend: "Autonomy" isn't confined to mechanical limbs — it's also about digital intelligence. Rather than teaching robots in your living room how to wash dishes, why not have agents on the network learn to "act on your behalf"?
AI Agents are the extension of this direction. They're not humanoid replicas but digital extensions of human will — capable of executing tasks, making decisions, and completing transactions on behalf of users, with data ownership retained by the individual.
IBM's "2025 AI Agent Report" states that Agentic AI promises an 8x productivity boost, hinging on autonomous reasoning combined with privacy protection.
This shift redefines "autonomy": no longer machines mimicking human limbs, but agents learning to represent human intent.
Amid this trend, the XWorld platform's explorations stand out. Since its 2023 launch, it has built a self-sustaining "agent economy" by combining AI training with token incentives: users can create, deploy, and monetize their own AI Agents. The integration of stablecoins makes settlements lower-friction, ensuring value flows under user control.
Today, XWorld boasts over 11 million downloads and 1 million monthly actives in its Telegram MiniApp ecosystem, with cumulative token trading volume exceeding $34.7 million.
Here, autonomy is no illusion — it's a reality co-built by users, developers, and agents: machines not only execute instructions but become "intelligence we own."
Neo reminds us: when "autonomy" becomes a selling point, oversight and trust must evolve in tandem.
The future shown by the AI Agent industry offers another possibility:
XWorld's experiment may provide the answer: when "agent autonomy" merges with "user ownership," machines finally begin to belong to us.
In the future, when robots no longer need human eyes, that may be humanity's true liberation.
🔗 Learn more and join XWorld
Website | Whitepaper | Twitter | Telegram | Youtube | Linktree



