Computing Power and Personality Design Are the Biggest Challenges

The ‘screenless AI device,’ originally scheduled for release next year, has encountered obstacles due to shortages in computing power, software bottlenecks, and challenges in defining the ‘AI personality.’ OpenAI is still struggling to meet the computational demands of ChatGPT, let alone support another ‘always-on’ consumer-grade AI device.

OpenAI CEO Sam Altman and former Apple chief designer, “Father of the iPhone” Jony Ive, are attempting to develop an artificial intelligence device that “redefines human-computer interaction.” However, this ambitious plan is currently facing multiple bottlenecks.

On October 5, the Financial Times of the UK reported, citing sources familiar with the matter, that the launch of this “screenless AI device,” originally planned for next year, has been hindered by a shortage of computing power, software bottlenecks, and difficulties in defining the “AI personality.”

A person close to Ive stated: “Computational capacity is another key factor causing delays. Amazon has the computing power needed for Alexa, and Google has it (for its Home devices), but OpenAI is already struggling to meet the computational demands of ChatGPT, let alone an AI device—they need to address this issue first.”

Initial Hardware Specifications Emerge

In May this year, OpenAI acquired io, a subsidiary of LoveFrom, the design firm founded by Ive, for $6.5 billion, marking the formal collaboration between the AI pioneer and the design maestro.

It is reported that the two envision creating a palm-sized, screenless device capable of perceiving the world through cameras and microphones and interacting with users via natural language—a kind of “next-generation AI companion.”

The device is said to be “always on,” continuously collecting users’ visual and auditory information to build personalized “AI memory.” It can either be placed on a desk or carried around, aiming to transcend voice assistants like Echo or Siri and become a true “AI friend” that understands users deeply.

According to a previous article from WallstreetCN, OpenAI has reached a hardware manufacturing agreement with Luxshare Precision Industry, a domestic consumer electronics manufacturing giant, with plans to launch its first AI device product by the end of 2026 or early 2027.

Shortage of Computing Power; Difficulty Defining AI “Personality”

The report cited informed sources who indicated that one of the biggest challenges for this project lies in computing power. Unlike Amazon (Alexa) and Google (Home), which possess vast cloud computing resources, OpenAI is currently struggling to meet the computational needs of ChatGPT, let alone support a consumer-grade AI device designed to run “around the clock.”

“Computing power is one of the biggest factors causing delays,” a person close to Ive said. “Without sufficient computational infrastructure, no device can even be discussed.”

Aside from computing power, another major challenge is how to define the ‘personality’ of AI. The OpenAI team hopes that this AI assistant will be as natural and friendly as a friend, without coming across as ‘strange’ or ‘overly intrusive.’

“We don’t want it to become your ‘weird AI girlfriend,'” an informed source half-jokingly remarked. “It should be like Siri, but smarter and warmer.”

However, striking the right balance between ‘smart’ and ‘quiet’ has become a difficult task. If the AI talks too much, it could become irritating; if it’s too aloof, it may seem uninteresting. As one participant put it: “The personality of the model is the hardest part to balance—it shouldn’t be too ingratiating or too blunt, helpful but not caught in a self-referential loop.”

Lessons from the Past: Humane and Friend’s ‘AI Disillusionment’

OpenAI and Ive’s project is not the first of its kind. In the past year, several companies have attempted to create ‘AI companion’ devices, with most ending in failure.

Humane’s AI pin, backed by Altman, was met with lukewarm market reception due to performance and interaction issues; another product called Friend was criticized for being ‘too talkative’ and having a ‘strange personality,’ described as ‘creepy.’

These failed cases have made OpenAI more cautious: ensuring that AI remains ‘always on’ without becoming a ‘nuisance’ has become a key focus.

Despite these challenges, Altman’s hardware ambitions remain undeterred. OpenAI’s valuation has soared to $500 billion this year, surpassing Elon Musk’s SpaceX. To match this valuation, Altman urgently needs to demonstrate that OpenAI can go beyond being a ‘software company’ and build a complete AI ecosystem.

Not only is OpenAI aggressively recruiting former Apple hardware talent, but it has also poached multiple engineers who worked on VR and smart glasses projects at Meta. Industry insiders believe that these moves indicate Altman is attempting to replicate Apple’s ‘hardware-software integration’ approach.

Editor/Lambor