Sam Altman and Jony Ive are facing major hurdles in their ambitious attempt to build a tangible AI device through their new company, IO. The project, pitched as a revolutionary "third core" device beyond smartphones and computers, is supposed to be a palm-sized, screenless gadget that listens to audio and observes visual cues from its environment to respond to user needs. It’s designed to sit on a desk or be carried around, but according to a Financial Times report, the team is struggling with fundamental challenges that could delay its launch. One of the biggest issues is compute power. Despite OpenAI’s success with ChatGPT, the company has faced ongoing limitations in scaling its AI infrastructure. As one anonymous source familiar with IO’s operations put it, “Amazon has the compute for an Alexa, so does Google, but OpenAI is struggling to get enough compute for ChatGPT, let alone an AI device—they need to fix that first.” Without sufficient processing power, the device can’t reliably run the complex AI models needed to interpret real-time inputs and deliver useful responses. Beyond hardware and performance, the device’s core functionality is proving difficult to perfect. It’s designed to be always listening, which raises serious privacy concerns. With a camera and microphone constantly active, the device could record everything and everyone in its vicinity—potentially creating a persistent surveillance risk, especially given OpenAI’s track record with user data. The challenge isn’t just capturing data, but knowing when to respond and when to stay silent. Getting this balance right is proving far harder than expected. Adding to the complexity, the device’s “personality” and user experience are still undefined. Unlike existing smart speakers that operate on clear commands, IO’s vision relies on contextual awareness and subtle cues. But if the system misinterprets signals or responds inappropriately, it could frustrate users or feel invasive. The broader landscape of AI gadgets isn’t helping. Humane, which launched the high-profile but underwhelming Ai Pin, has been dismantled and sold to HP. Rabbit’s R1, once seen as a bold alternative to smartphones, has struggled to gain traction despite recent updates like rabbitOS 2, which simplified its interface and added touch-based controls. Nearly 18 months after launch, it remains a niche product with limited impact. While IO has the advantage of resources and star power, the challenges it faces are not just technical—they’re existential. The real question isn’t whether it can build a device, but whether people want one that’s always listening, always watching, and always responding. In a world where privacy concerns are growing and AI adoption is still uneven, the answer may be no. For now, Altman and Ive appear to be learning the hard way that building the future isn’t just about vision—it’s about solving the messy, real-world problems that come with it.