Should Home Robots Talk? Silence May Win

The Familiar, from his new startup Familiar Machines & Magic, is a small, bear-like quadruped about the size of a dog, covered in touch-sensitive fuzzy fabric. The Robot Report's May 4 interview describes 23 degrees of freedom, edge AI that reads facial expressions and tone of voice, and a behavior engine trained on thousands of narrative vignettes. Angle describes it following people from room to room; the company's pitch includes a nuzzle when you're stressed and a nudge when you've been doomscrolling too long. These are proposed behaviors for a robot still in development.

What it doesn't have is a screen-first interface or conversational voice. That is not a missing feature. It is the entire point.

"Screens are actually not very good at human connection," Angle told The Robot Report. "When you're physically interacting with something, you get a physiological response that you don't get with a screen."

This is a deliberate bet: that emotional connection between humans and robots works better through motion, posture, timing, and physical presence than through conversation. It is a controversial position in a market where every new robot seems to be adding an LLM and a voice. So let's examine the evidence — from both sides.

The Nonverbal Camp: Robots That Communicate Without Words

Familiar isn't the first robot to choose nonverbal communication. Some of the most emotionally successful companion robots in history barely use language at all.

Sony Aibo: 150,000 Units and Buddhist Funerals

Sony's aibo is the benchmark for emotional companion robots. The original AIBO launched in 1999 and reportedly sold more than 150,000 units before being discontinued in 2006. When Sony stopped repairing old units, Buddhist funeral ceremonies were held for owners' robot dogs.

The current ERS-1000, relaunched in 2018, is listed in the United States at $3,199.99 including three years of AI Cloud Plan, checked October 14, 2026. Sony describes 22 axes of movement, OLED eyes, and deep learning that develops a unique personality. Aibo responds to voice commands such as "Perform a dance", but its primary emotional language isn't verbal — it's body movement, eye expressions, ear positions, and the way it trots over to greet you. The barks and whimpers are animal-like, not conversational. Aibo doesn't explain why it's happy to see you. It shows you.

This is the lineage Angle is drawing from. "If I try to make a robot dog, I disappoint," he said. "I'm not trying to make a pet. I'm trying to make a familiar — something pet-like, but not bound by those expectations."

LOVOT: Warmth Without Words

GROOVE X's LOVOT, available in Japan since 2019, takes the nonverbal approach even further. In Japan, the basic ちゃ color of LOVOT 3.0 costs ¥577,500 through October 25, 2026, rising to ¥599,500 on October 26, including tax, administration and shipping, checked October 14. It is not cheap, and a monthly care plan is required separately.

LOVOT doesn't speak. The official 3.0 specifications list 4.6 kg without clothes, OLED eyes, full-body touch sensing and 11 degrees of freedom including the two wheel axes. It has a warm body and a sensor "horn" on its head. The often-cited total of more than 50 sensors appears in Miraikan's GROOVE X-supported exhibit description and research using earlier 4.3 kg LOVOT hardware; the current official 3.0 specification lists sensor types without a total. It recognizes its owner, develops a personality over time, and waddles back to its charging nest when it needs power.

The emotional hook is physical: holding LOVOT triggers a warmth-and-weight response that is hard to replicate with a screen. It is designed to be picked up and cuddled, and its responses — cooing, shifting weight, closing its eyes — are somatic rather than conversational.

Mirumi: Connection Through Glances

At the other end of the price spectrum, Yukai Engineering's Mirumi is listed at $149.99 in its U.S. preorder store, checked October 14, 2026. It is a tiny clip-on companion that hangs from a bag strap and responds with shy glances, head turns, and spontaneous motions. No speech. No screen. No navigation. Just brief moments of eye contact that create what the designers call playful curiosity.

Mirumi launched in Japan on April 23, 2026. As of October 14, the U.S. store's October-shipping preorder listings are marked sold out; that describes those listings, not whether earlier orders shipped. Its design makes a modest bet: emotional connection need not require expensive hardware. Sometimes a well-timed head tilt is enough.

The Talking Camp: Robots That Speak Your Language

But silence is not the only approach — and for some use cases, it is clearly not the right one.

ElliQ 3: Conversation as Care

Intuition Robotics' ElliQ 3 was designed specifically for older adults experiencing loneliness. It talks. A lot. ElliQ initiates conversations, asks about your day, suggests activities, reminds you about medications, and can video-call family members.

The strongest public evidence for ElliQ comes from New York's elder-care program. NYSOFA's current ElliQ initiative page says the program includes approximately 900 units being made available to older adults in the community. It also cites a 2023 NYSOFA report showing a 95% reduction in loneliness among older adults using the platform, with users interacting with ElliQ more than 30 times per day, 6 days a week. More than 75% of those interactions were related to social, physical, or mental wellbeing.

Those are historically reported program results, not a controlled comparison with silent robots or evidence specific to ElliQ 3. They do not prove every person will respond to a talking robot. Those results help explain why speech matters in this use case. For isolated older adults, conversation is not a gimmick. It is part of the intervention. A nuzzle does not remind you to take your blood pressure medication. A head tilt does not help you feel heard when you need structured engagement.

The Voice-First Trend

The broader market is still moving toward voice. Many humanoid demos, mobile assistants, and care robots are being built with conversational AI front and center. The assumption is that people want their robots to understand natural language — to be commanded, questioned, and conversed with.

And there is logic to it: if you are paying thousands of dollars for a robot, shouldn't it at least answer your questions?

What Does the Science Say About Robot Communication?

Research does not give buyers a simple universal answer. A 2024 systematic review in the International Journal of Social Robotics describes companion robots as systems that aim to establish emotional connections and facilitate social interaction, sometimes while also helping with specific tasks. It also notes that companion robots commonly use human-like or animal-like features, verbal commands, face recognition, emotion interpretation, empathy, contextual behavior, and social presence.

The useful buyer takeaway is not "talking wins" or "silence wins." It is that communication mode has to match the robot's job. Spoken reminders and conversation need language; companionship for an older adult can also be nonverbal. A pet-like companion designed for warmth, play, and low-pressure presence may be more believable when it communicates through movement, gaze, sound, and timing.

That supports a narrower version of Angle's thesis: you don't necessarily need a robot that talks to feel connected to it. Embodiment, posture, context-aware behavior, and physical timing can do emotional work that a screen or chatbot cannot. But context matters enormously. A companion robot for a lonely 85-year-old has different communication requirements than one for a tech-curious 35-year-old who wants a screenless creature around the apartment.

The Spectrum: Where Companion Robots Fall

Here is how these designs compare. Prices and availability were checked on October 14, 2026; dollar prices below are U.S. listings, while LOVOT is the Japan basic-color example above.

The pattern is clear enough: robots designed primarily for emotional presence often lean toward nonverbal communication, while robots designed for structured care or explicit assistance rely on conversation.

Why Would a Silent Robot Be Better Than a Talking One?

Angle's argument rests on several points worth taking seriously.

Avoiding the uncanny valley of conversation. We have all had the experience of chatting with an AI that is almost human but not quite. That gap can be more unsettling than no conversation at all. A robot that speaks imperfectly can feel creepy. A robot that communicates mostly through motion sets different expectations and can meet them more reliably.

Privacy through local processing. Familiar is not microphone-free. The robot is reported to use vision and audio inputs, including tone-of-voice analysis. The privacy claim is narrower and more credible: Familiar Machines says data is stored on the device and users control if and when they share it with the cloud; The Robot Report says Angle's team is using local edge AI to avoid constant cloud streaming. That still leaves questions about logs, updates, diagnostics, and consent, but it is better than assuming every interaction must be streamed to a server.

Emotional authenticity through constraint. There is a reason dogs and cats form deeper emotional bonds with many owners than a smart speaker does. Pets do not explain their feelings in words. They demonstrate them through behavior. Angle is betting that the same principle applies to robots: showing emotion through motion may be more believable than describing it through language.

Lower disappointment risk. When a talking robot gets confused, misinterprets you, or gives a generic response, the illusion breaks immediately. A nonverbal robot that nuzzles you when you are sad does not have to understand the full nuance of your sadness. It has to show up with warmth at the right time.

When Talking Is Better

But let's not pretend silence is always superior. There are real scenarios where verbal communication is essential:

• Medication and health reminders require specificity that body language cannot convey.
• Crisis situations — a fall, a medical emergency — demand clear, unambiguous alerts.
• Loneliness intervention for people who genuinely lack human conversation partners may need actual conversation.
• Setup and configuration are easier when a user can say what they need.
• Accessibility matters: not everyone can read body language, gaze, or physical cues in the same way.

The honest answer is that both approaches have a place. The question isn't "should robots talk?" but "which robots should talk, and when?"

What This Means for Buyers

If you are considering a companion robot in 2026, here is what to think about.

Who is it for? A nonverbal companion like LOVOT or Familiar, if Familiar ships as described, may create a deeper, more intuitive emotional bond. But if the user needs structured engagement — conversation, reminders, cognitive stimulation — a talking robot like ElliQ 3 is the better tool.

What is the privacy tolerance? Familiar's pitch is not "no sensors." It is on-device processing and user-controlled cloud sharing. That is promising, but buyers should still ask what video, audio, and interaction data is stored, whether cloud sharing can be disabled, and how household memory can be deleted. Mirumi is simpler: its official FAQ describes USB charging and responses to sound, without understanding words or using AI.

What is the budget? At the October 14 check, Mirumi lists at $149.99 with its October preorder sold out; aibo costs $3,199.99 including three years of AI Cloud Plan; Japan's basic-color LOVOT 3.0 costs ¥577,500 through October 25 plus monthly care. ElliQ's U.S. membership adds a $249 lease-initiation fee to $59 billed monthly, or $588 prepaid annually or $936 prepaid for two years. Those longer plans are not monthly billing. Familiar remains price TBA, with Angle's dog-ownership-cost comparison as a public clue.

Is this supplementing or replacing? Familiar Machines itself says Familiars "are not a replacement for humans or pets. They're a supplement." That framing matters. A nonverbal robot works best when it is adding to your life, not pretending to replace relationships that require actual conversation.

The Bigger Picture: A Platform Play

Angle is not just building one quiet robot. He is explicit that this is a platform. "When we get to act two and humanoids are coming in the home, you want them to be familiar, right?" he told The Robot Report. "You don't want them to be uncanny."

The implication is that Familiar's emotional intelligence — reading facial expressions, learning routines, developing personality, and responding through body language — could become the social foundation for future products, including ones that eventually do talk.

"Selling a million familiars has more economic value than all of the Roombas ever sold by a large margin," Angle claimed. That is ambitious for a product Angle said would first be available in 2027, with no exact release date or price, and which is entering a market littered with failed or faded companion robots. But Angle has scaled consumer robotics before, and his team — drawn from iRobot, Disney Imagineering, MIT, and Boston Dynamics — has the pedigree to attempt it.

Our Take

The talking vs silent debate will not be settled by one product. But Familiar's entrance into the market is significant because it challenges the industry's default assumption that more voice always means a better robot.

Jibo, Kuri, and Anki Vector show the difficulty of sustaining a companion-robot business. Bosch's Kuri retrospective describes a failure to find a business fit inside Bosch; an Anki spokesperson cited a collapsed financial deal. In his analysis of their companies' failures, robotics researcher Guy Hoffman identifies long-term engagement and business models as important challenges. That is an expert interpretation, not evidence that every product failed because owners lost interest. Voice versus silence alone does not explain those outcomes.

Angle's bet is that physical, mostly nonverbal presence creates a different kind of bond — one that does not degrade when the conversation gets repetitive or the chatbot says something weird. It is a bet on emotional durability rather than conversational capability.

The planned 2027 launch will begin to test whether he is right. But even now, the existence of this debate is healthy for the home robot market. Not every robot needs to be a chatbot with legs. Some may be better as quiet, warm presences that simply show up when you need them.

Browse all companion robots in our database to compare specs, prices, and communication styles side by side. Or check our full robot comparison tool to filter by what matters most to you.