Key Takeaways
- Friend 2.0 adds a built‑in speaker and voice capability, turning a text‑only pendant into an audible AI companion that requires a $10/month subscription for long‑term memory.
- Avi Schiffmann frames the device as a “confidant, friend, God”‑like presence, deliberately distancing it from productivity tools or romantic AI, which mirrors the ambiguous role OpenAI reportedly envisions for its screen‑less speaker.
- The shift from silent text replies to spoken responses raises privacy concerns, as audible interactions are inherently public and can unsettle bystanders.
- Friend’s high‑profile New York subway campaign and subsequent backlash illustrate the societal friction that always‑listening wearables can provoke, a lesson likely relevant to OpenAI’s forthcoming device.
- While Humane’s AI Pin failed commercially, Friend remains the only company currently testing a screen‑less, voice‑enabled AI companion in the real world, offering a preview of what OpenAI’s joint venture with Jony Ive’s io might deliver.
Introduction and Context
Screenless devices are emerging as the next frontier in artificial‑intelligence hardware, promising interactions that bypass traditional displays in favor of audio, haptics, or ambient cues. Rumors have swirled that OpenAI’s first foray into this space could be a screen‑free smart speaker developed alongside Jony Ive’s hardware startup, io, designed to anticipate user needs and act as a humanlike companion. Although concrete details remain scarce, a smaller player—Friend, the AI pendant startup founded by Avi Schiffmann—has recently offered a tangible glimpse of what such a voice‑driven companion might feel like in everyday use.
Friend 2.0 Launch and Features
On Thursday, July 30, Friend relaunched its pendant with a feature it previously lacked: a built‑in speaker that enables the device to talk back directly. The original $99 model could only convey responses via text messages sent to a companion app; Friend 2.0, now priced at $249, adds audible output, allowing the pendant to speak aloud. As Engadget reported, a $10 monthly subscription is now required for the pendant to retain memory of conversations beyond 30 days; without it, the device “forgets what it discussed with its owner roughly a month after the conversation happened.” This shift transforms the pendant from a passive notification tool into an active conversational partner that can interject in real time.
The Significance of Voice Interaction
The move from text‑only replies to spoken responses is far more than a minor feature upgrade; it fundamentally alters the nature of the user‑device relationship. As Schiffmann explained in a recent post on X, “I am interested in this kind of relationship in attempting to offer, some kind of confidant, friend, God, not really sure what it is,” a sentiment captured by TechCrunch. A pendant that types out responses in an app is something a user checks intermittently; a pendant that talks back out loud becomes a continual presence, audible to the wearer and to anyone nearby. This public dimension introduces new social dynamics, turning private reflections into potentially overheard dialogues.
Schiffmann’s Vision and Ambiguity
Schiffmann deliberately avoids positioning Friend as a productivity assistant or a romantic AI, instead describing it as an open‑ended confidant whose exact role remains fluid. That ambiguity mirrors the functional description Bloomberg gave for OpenAI’s rumored device: “designed to become more personalized over time, learning a user’s habits and anticipating needs rather than simply answering direct requests.” Both companies aim to create an AI presence that evolves with the user, offering companionship that feels intuitive rather than transactional. By refusing to pin the device down to a specific use case, Schiffmann hopes to foster a bond that feels more like a friend or mentor than a tool.
Comparisons to OpenAI’s Rumored Device
Bloomberg’s July report suggested that OpenAI’s collaboration with io will yield a screen‑free smart speaker built to anticipate a user’s needs and function as a “humanlike companion.” The described capabilities—personalized learning, habit anticipation, and proactive assistance—parallel Friend’s own emphasis on a “randomized personality” and persistent memory enabled by the subscription model. While OpenAI’s device remains speculative, Friend’s real‑world deployment provides a concrete test case for how users might react to an ever‑listening, voice‑enabled AI that speaks aloud in public settings.
Public Reception and Privacy Concerns
Friend’s initial launch a year ago attracted significant attention, not all of it positive. Schiffmann revealed that the company’s original marketing campaign, which ran in New York’s subway system and cost “more than $1 million,” sparked public debate and even vandalism from commuters uneasy about an always‑listening pendant. The backlash has lingered, and the recent price increase to $249 coupled with a mandatory subscription for memory retention invites further scrutiny. As The Next Web observed, “a pendant that silently texts is private, while one that audibly replies is not,” highlighting the tension between utility and privacy when AI companions become audible in shared experiences that bystanders can overhear.
Lessons from Humane’s AI Pin and Market Outlook
Friend is not the first company to test a screen‑less wearable AI. Humane’s AI Pin, marketed as a smartphone alternative, folded within a year of launch after sales fell short of expectations. Friend’s continued presence—despite its controversies—makes it the sole company currently delivering a voice‑enabled, screen‑less AI companion to consumers. This position offers valuable insights for OpenAI and io: the importance of clear privacy safeguards, the need to manage public perception of always‑listening devices, and the potential value of a subscription model that balances ongoing service costs with user expectations for memory and personalization.
Conclusion and Future Implications
Friend 2.0’s evolution from a text‑based pendant to a speaking AI companion illustrates both the promise and the pitfalls of screen‑less, voice‑first hardware. By adding audible responses, the device moves closer to the kind of ambient, anticipatory companion that OpenAI reportedly seeks to build with Jony Ive’s io. Yet the journey also underscores societal hurdles: users must weigh the comfort of a conversational AI against the discomfort of exposing private dialogues to the surrounding world. As the market watches Friend’s reception and anticipates OpenAI’s eventual unveiling, the lessons learned—regarding pricing, subscription models, privacy transparency, and public etiquette—will likely shape the next generation of screenless AI hardware, determining whether such devices become trusted confidants or fleeting novelties.
Friend’s Talking Pendant Gives OpenAI’s Screenless Future a Voice