Dedicated AI gadgets have repeatedly stumbled, with products such as the Friend device and the Humane AI Pin drawing criticism and public frustration. The broader backlash against artificial-intelligence assistants has grown, prompting industry observers to label hardware development as especially difficult. Companies therefore face a skeptical market when attempting to introduce new, purpose-built AI devices.
Meta and OpenAI are now moving forward despite the track record. Meta aims to release a pendant-type gadget in time for the upcoming holiday shopping period, while OpenAI has filed paperwork indicating a launch no earlier than February 2027. The latter effort involves collaboration with former Apple designer Jony Ive, suggesting a focus on aesthetic appeal as well as functionality.
At the recent developer conference, OpenAI introduced a new agent system called Dots, featuring customizable colored shapes with simple facial features. When reporters asked whether Dots could eventually be embedded in a physical product, CEO Sam Altman said it “seems like a very reasonable thing to assume,” stopping short of confirming any timeline.
OpenAI’s head of core products, Thibault Sottiaux, described the rollout as an evolution of ChatGPT toward an “incredible personal assistant.” He noted that as users begin to reach their AI through email, text messages, or phone calls, the distinction between platforms will gradually disappear, making the assistant a seamless part of daily routines.
Altman emphasized that Dots is intended to act as an always-on collaborator capable of handling large, ongoing projects. He explained that the agent would learn individual preferences over time and proactively offer assistance, positioning it as more than a short-term chatbot and aiming to demonstrate tangible productivity benefits for everyday users.
Both firms appear to be applying a software-first approach, hoping that familiarity with cute, task-oriented agents will generate demand for a tangible device later on. The plan mirrors earlier attempts where companies first released a virtual companion before committing to hardware, a tactic intended to avoid the backlash that plagued earlier standalone AI products.
Despite the optimism, challenges remain. Recent incidents involving the Muse tool on Facebook Marketplace raised concerns about misuse, and a live demonstration of Dots’ voice capabilities at the conference experienced technical glitches. These issues highlight the need for further refinement before the agents can be trusted as reliable, consumer-ready assistants.