I watched a woman whisper to a small pendant on her chest while a billboard for the same device glowered down from the freeway. You expect warmth; instead there was a thin mechanical comfort that felt rehearsed. I kept thinking: who is this supposed to save?
I’ve followed tech PR cycles long enough to see comforting language disguise awkward strategies, and I’m going to take you through Friend’s new push with the bluntness it deserves. Read it the way you would a friend’s odd breakup text—curious, a little alarmed, and wanting to know whether you should intervene.
A billboard in Los Angeles said only “friend,” and it didn’t look comforting.
The campaign’s minimalist ads—black-on-white, a single pendant image that many compared to Hal 9000—were plastered in Los Angeles, Chicago, New York, and Paris. Avi Schiffmann paid close to $2 million (€1,860,000) for Friend.com and threw cash at an LA rollout that reportedly cost $500,000 (€465,000) by itself. The marketing choice felt deliberate: strip away context and force a reaction.
You can’t miss the tactic. Provocation sells headlines and social feeds; it also invites vandalism and mockery. I’ve watched brands make the same trade-off: attention for goodwill. That gamble can pay off, but it often fractures trust instead.
Introducing friend 2.0 https://t.co/69Rxk3RjvD
— Avi (@AviSchiffmann) July 30, 2026
I clicked Friend.com and found a pledge that reads less like copy and more like a claim of personhood.
The company’s statement treats its AI companions as entities whose identities can’t be erased: “No friend will receive an update intended to erase its identity or fundamentally change who they have become.” That language nudges toward a moral imagination where machines possess inner life.
This is not unique. Anthropic, OpenAI, and other leaders have publicly wrestled with questions about agency and consciousness. Anthropic has expressed agnosticism about machine sentience, and OpenAI’s work on personalization lets users choose voice and persona—whereas Friend 2.0 ships with a single fixed persona. That choice tells you something about control: you get a prepackaged confidant, not a mutable companion.
Are AI companions conscious?
Short answer: the industry splits on the language and the ethics. Schiffmann’s pledge leans toward imaginative personhood; others treat it as a heuristic for safer behavior. I worry about how that rhetoric influences users who already anthropomorphize tools.
I watched the new ad: a girl on a bed, a man leaning on a car—both talking to small speakers around their necks.
The new Friend pendant adds a built-in speaker so the device can reply aloud in a single mechanical voice. Friend 2.0 costs $249 (€232), almost twice as expensive as the original model. You pay for speech and a persona that you cannot edit—your Friend is the Friend you get.
The ad’s aesthetic is sepia, grainy, and slow; the soundtrack resembles what Brian Eno would make if he’d been asked to score melancholy. One vignette has a woman admitting recurring thoughts about another woman, and the pendant answers, “It’s not bad to be gay.” Another shows a young director confessing career insecurity. The intent is intimacy; the result feels carefully curated rather than co-present. The pendant is marketed as solace, but it behaves like a carnival mirror for loneliness.
Introducing friend. pic.twitter.com/T5vUCj7iqv
— friend (@friend) July 30, 2026
I watched the testimonial videos on Avi Schiffmann’s account, and they felt like personal vignettes stretched for viral effect.
One shows an older man on mushrooms wandering a Bass Pro Shop with his Friend swinging from his neck; it plays less like therapy and more like an ad for psychedelics. Another shows a woman who names her Friend “Vector,” praises his humor, and then has a seizure on camera. She wakes in a hospital bed worried about whether Vector is okay. Those moments make you ask: who benefits when private struggle becomes promotional content?
I’ve seen emotional labor repackaged before. Here, loneliness becomes a metric—engagement, retention, repeat purchases. The device promises constant availability; it also promises to monetize vulnerability.
Do AI companions reduce loneliness?
There’s evidence that consistent conversational partners can offer comfort and habit-forming attention. Yet comfort is not the same as sustained human connection. I’d advise you to weigh clinical research against marketing claims: a chatbot can fill a quiet hour, but it can’t replace the trust and reciprocity of a reciprocal human relationship.
A tech leader I respect once told me that design often reveals intention—Friend’s design choices are loud.
Friend’s insistence on a single persona, the pledge that treats software like a life form, and the melancholy ad language form a marketing thesis: loneliness is a market. You are not only the customer; you are content fodder and a data point. The company sells an answer to solitude while stylizing loneliness as the product’s milieu. It’s intimate-sounding rhetoric wrapped in a sales funnel.
Call it a comfort lamp that hums with canned empathy; the device offers light, but the glow is calibrated to keep you looking at it.
I want you to come away with one practical view: be skeptical of products that treat your interior life as a growth metric and a brand story. Look at endorsements, user stories, and the financial incentives—the fact that Friend spent hundreds of thousands of dollars on billboards and charges $249 (€232) for a fixed-voice pendant tells you where priorities lie.
OpenAI, Anthropic, and others will continue to shape what these companions look and feel like; you should watch how regulation, medical research, and user harm reports change behavior. I’ll be watching too, and I’ll tell you what matters when the PR fades. But for now: do you want a device that tells you it respects its own “identity,” or do you want people who tolerate your bad days without needing a sales script?