The development of the AI device by OpenAI and Jony Ive is reportedly facing technical issues. What's happening?

OpenAI and Jony Ive are encountering significant technical hurdles in the development of their anticipated AI device. According to a report by the Financial Times, the team is working on a palm-sized, display-less device that should react to audio and visual cues from its surroundings, but they are running into problems that could delay its market launch, originally planned for 2026.

Technical problems hinder development

The project originated after OpenAI acquired io (founded by former Apple chief designer Jony Ive and OpenAI CEO Sam Altman) for $6.5 billion in May of this year. According to Financial Times sources close to the development, the team is grappling with three main challenges that complicate the product’s completion.

The first problematic area is defining the AI assistant’s “personality”. Developers are striving to create a balanced personality that is helpful but not intrusive. One source described the target persona as “a friend who is a computer, but not your weird AI girlfriend… like Siri, but better.” The team is figuring out how to ensure the assistant knows when to intervene and when to end a conversation.

The second challenge involves privacy concerns associated with the concept of an always-on device. Unlike typical smart speakers, this device will not wait for an activation word but will constantly perceive its surroundings, raising user privacy concerns.

The third and perhaps most serious problem is ensuring sufficient computational power to run OpenAI models on a consumer device. “Amazon has the computational power for Alexa, Google [for Gemini, editor’s note] also, but OpenAI is already struggling with a lack of computational power for ChatGPT, let alone for an AI device — they need to solve that first,” one source close to Jony Ive told the Financial Times.

What will the device look like and how will it function?

According to available information, it will be a palm-sized device without a display, which users will control using a microphone, camera, and speaker. Primarily, it should be placed on a desk or workspace, but it will be compact enough to be carried around.

A key feature is its ability to react to visual and auditory cues from its surroundings and respond to user requests. The device should gather information throughout the day, which will serve to create the assistant’s “memory,” allowing it to provide more relevant answers and better anticipate user needs.

It is speculated that OpenAI is collaborating with Chinese manufacturer Luxshare for hardware production, although final assembly may take place outside of China. The company is also actively recruiting former Apple and Meta employees with extensive hardware development experience.

Lessons from the failures of Humane AI Pin and Rabbit R1

The new device from OpenAI and Jony Ive plans to enter the market after two similar products – the Humane AI Pin and Rabbit R1 – experienced significant failures. The Humane AI Pin (sold for 699 dollars, or about 14,500 Kč at the current exchange rate) was recently discontinued entirely. The device struggled with overheating, an unusual user interface, and poor artificial intelligence.

The main reason for the failure of these devices, according to experts, was the ignorance of the ubiquity of mobile phones. “In a world where smartphones dominate our lives, why would consumers pay a premium price for an inferior, redundant device?” CNET quoted Francisco Geronimo, VP of data analysis at IDC.

Unlike unsuccessful competitors, more successful products like Meta Ray-Ban glasses function as an accessory to phones, not as their replacement. As of this February, 2 million units have already been sold, with AI features presented as a useful bonus rather than the main selling point.

Will OpenAI and Jony Ive succeed where others have failed?

It remains to be seen whether OpenAI and Jony Ive will manage to create a device that functions as an accessory with truly useful features. Given the fiasco surrounding the two aforementioned “AI gadgets,” it should be a product that can satisfy not only tech enthusiasts but also ordinary users with everyday problems – otherwise, its chances of success will likely dramatically decrease.

Do you think a standalone AI device makes sense? What should it be able to do?

Sources: Financial Times, TechCrunch, Android Authority