I've spent over $4,300 on AI wearables in the past eighteen months. Not smartwatches—everyone's already swimming in Apple Watch reviews. I'm talking about the new stuff. The AI pins that promised to replace your phone. The smart rings that swore they'd revolutionize health tracking. The AI pendants that claimed they'd record your entire life without being creepy.
But here's the thing—some of them do work. Not in the revolutionary "throw away your phone" way the marketing promised, but in genuinely useful, occasionally delightful ways that have actually stuck in my daily routine. After six months of real-world use, I can finally tell you which AI wearables are worth your money, which ones are expensive paperweights, and which ones might be brilliant in another year or two.
This article is structured to save you time. If you're just curious about whether the hyped products worked, read the main reviews section. If you're a genuine enthusiast who wants to know about every weird AI wearable currently in development, the deep dive section has you covered. And if you just want to know "should I buy something right now?"—skip to the buying guide and reality check sections near the end.
Fair warning: I'm going to be brutally honest about what failed. The Humane AI Pin? A beautiful disaster. The Rabbit R1? It found its niche, but it's tiny. That $599 AI pendant that was supposed to be your "second brain"? It forgot everything after two weeks when the company went bankrupt. This isn't going to be a cheerleading piece about the glorious future of ambient computing. This is about what actually works when you're trying to get through Tuesday.
One more thing before we dive in: I'm defining "AI wearables" narrowly here. This isn't about fitness trackers with some machine learning algorithms or smartwatches with voice assistants. I'm focusing on the new category of devices that launched between late 2023 and now—products specifically designed around large language models, multimodal AI, or ambient intelligence. Things that promised to fundamentally change how we interact with technology, not just track our steps better.
Alright. Let's talk about what I learned from spending way too much money on the future.
Rabbit R1 (2026 Edition)
Price: $199 (one-time, no subscription)

I need to split this review into two parts: the disaster that shipped in May 2024, and the genuinely useful device it became by late 2025.
The Rabbit R1 launched with massive hype at CES 2024. For $199, you'd get a pocket AI companion powered by their proprietary "Large Action Model" that could actually do things across apps, not just chat with you. Order an Uber, play Spotify, search complex queries, control smart home devices—all through natural conversation with a physical device, no phone required. Jesse Lyu, the founder, positioned it as "the first handheld AI device that takes action for you."
What Actually Works Well (Now):
After eighteen months of aggressive software updates, the R1 has found its groove as a dedicated voice assistant for specific tasks. I use mine daily for three things:
- Music control while working: The R1 sits on my desk connected to Spotify. "Play something energetic" or "find me focus music" works beautifully without reaching for my phone. The physical scroll wheel is perfect for volume, and the speaker quality is surprisingly decent for a $199 device.
- Quick information lookups: When I'm cooking or have my hands full, asking the R1 for recipe conversions, timer management, or random facts is faster than pulling out my phone. The AI responds in 1-2 seconds now (it was 8-12 seconds at launch).
- Travel mode: This is where the R1 genuinely shines. On a work trip to Austin last month, I used it exclusively for navigation, restaurant recommendations, and translation (my Spanish is terrible). The push-to-talk trigger feels more natural than "Hey Siri" when you're walking around a city. The battery lasted a full day of heavy use.
The camera integration improved massively. You can now point it at objects and ask questions—I used it at a hardware store to identify the right type of screw I needed, and it actually worked. The vision model correctly identified a #10-32 machine screw from a blurry photo. Six months ago, it would have hallucinated something about woodworking.
What Doesn't Work or Disappoints:
Let's be honest: the "Large Action Model" that was supposed to revolutionize how AI interacts with apps mostly failed. The demo showing the R1 booking complex travel arrangements? Never worked reliably. The promise of controlling any app through natural language? Turned out they had hard-coded integrations with maybe ten services, and half of those were buggy.
The Uber integration I used at launch literally ordered a ride to the wrong address three times in a row. I stopped trusting it after that. The DoorDash integration worked exactly once, then broke for three months. By the time they fixed it, I'd deleted the connection.
The hardware is plasticky and feels cheap. Mine has scratches all over the screen despite living in my pocket with nothing else. The orange color that looked cool in renders is... very orange. Like, construction cone orange. I only take it out in public because I've committed to using it, not because I'm proud to be seen with it.
Battery life is mediocre. Rabbit claims all-day battery, but with moderate use (30-40 interactions), I'm charging it by dinner. Heavy use means mid-afternoon charging. The USB-C port is weirdly stiff—I've already worn out one cable from the awkward angle.

Real-World Usage Scenarios:
Here's when I actually reach for my R1:
- Hands-free music and podcast control while I'm typing or in meetings
- Setting timers, converting measurements, asking for substitutions
- Navigation, quick translations, finding nearby restaurants without digging out my phone
Here's when I still use my phone instead: literally everything else. Email, messages, social media, web browsing, banking, photos, shopping. The R1 hasn't replaced my phone for anything beyond voice commands.
Who Should Buy It:
- People who want a dedicated voice assistant for specific tasks and hate pulling out their phone constantly
- Travelers who appreciate having a separate device for navigation and local information
- Tech enthusiasts who enjoy living with emerging technology and don't mind rough edges
- Parents who want to give kids something interactive that isn't a full smartphone
- Anyone who backed the original and is curious if the updates fixed it (they mostly did)
Who Should Avoid It:
- Anyone expecting it to actually replace their phone for daily tasks
- People who want polished, Apple-level fit and finish
- Anyone who hated their Alexa/Google Home and doesn't want another voice assistant
- People who need enterprise-level privacy (everything gets sent to Rabbit's servers)
- Anyone expecting the "Large Action Model" magic they promised—it's basically just a ChatGPT wrapper with some API integrations.
I still use my R1 3-4 times per day, which is more than I can say for most AI wearables I bought. It's not revolutionary, but it's become genuinely useful for a narrow set of tasks. I keep it on my desk during work hours and toss it in my bag when traveling. When people ask about it, I say it's "a $199 dedicated voice assistant that actually works now, but don't believe any of the original marketing."
The biggest compliment I can give: I'd buy it again at $199. I wouldn't buy it at $399. And I definitely wouldn't have backed it if they'd been honest about what it could do at launch.
Humane AI Pin
Price: $699 + $24/month subscription (required)

I desperately wanted this to work. The Humane AI Pin had everything going for it: stunning industrial design by ex-Apple designers, $240 million in funding from top VCs, celebrity backers, and a vision of screenless AI-first computing that felt genuinely innovative. I wore mine for three months straight, suffering through every limitation, hoping each software update would magically fix it.
It never did.
Humane sold a compelling vision: a wearable AI device that would replace your phone for most tasks through voice, occasional laser projections on your palm, and a suite of AI-powered features. No screen doom-scrolling, just intentional technology use. Ask it questions, have it take calls, translate languages in real-time, identify objects through the camera, capture photos and videos—all while staying present in the real world. The "Trust Light" would indicate when the camera/mic was active, addressing privacy concerns.
The marketing video showing the laser projection interface looked like something from a sci-fi movie. The moment where the founder held out her palm and saw a calendar floating there? I watched that clip fifty times.
What Actually Worked (Barely):
After using it extensively, I can identify maybe three things that worked as advertised:
- The build quality: The device itself is beautiful. Magnetically attaching to different clips and battery boosters feels premium. The weight distribution is well-engineered. It's the most elegant piece of hardware I bought in this category.
- Basic voice queries: Simple questions like "what's the weather" or "what time is it in Tokyo" worked reliably. Basically anything ChatGPT could answer, the Pin could answer, with about a 3-second delay.
- The trust light: This actually worked. The light indicates when sensors are active, and it never failed. Small victory, but it matters.
That's it. That's the list.
What Doesn't Work (Almost Everything):
Where do I even start?
- The battery life is atrocious. Humane claimed all-day use. Reality: 2-3 hours of moderate use before you need to swap battery boosters. I bought three extra boosters ($149 each!) to get through a normal day. The magnetic charging case is clever, but needing to swap batteries multiple times daily feels like using a phone from 2005.
- The laser projector is a gimmick. In theory, you hold out your palm and see a UI projected there. In practice, it only works in specific lighting conditions (not too bright, not too dim), requires you to hold your hand perfectly still at a specific angle, and the projected interface is nearly illegible in daylight. I used it maybe twenty times total, and every time felt like a party trick that took longer than just pulling out my phone.
Inside my house? Forget it—the lighting is wrong. Outside on a sunny day? Can't see it. In a coffee shop? People think you're insane holding your palm up and tilting your hand around. The only place it consistently worked was my office with specific overhead lighting. Useless. - The phone call experience is terrible. Calls sound like speakerphone held 12 inches from your mouth, which is exactly what it is. Everyone complained they couldn't hear me clearly. The person wearing the Pin sounds fine, but you sound like you're in a tunnel. After three dropped calls during a client meeting, I stopped using it for calls entirely.
- Translation mode doesn't work in real conversations. The demo showed fluid real-time translation. Reality: 5-8 second delays between speech and translation, frequent errors, and an awkward workflow where you have to hold the Pin, speak, wait, then play the translation. Google Translate on your phone is faster and more accurate.
- The camera is mediocre. 13MP photos that look fine in good lighting but fall apart anywhere else. The "capture moment" feature where you tap to take a photo sounds great until you realize there's no viewfinder, so you're guessing at framing. Half my photos were of the ground or sky. The video recording is 720p and looks like it was shot on a webcam from 2015.
- Response times are painful. Every query takes 3-10 seconds to process. In the age of instant Google searches, waiting eight seconds to learn what temperature to cook chicken is agony. Multiply this across dozens of daily interactions and the friction becomes unbearable.
- Heat. Dear god, the heat. After 20 minutes of use, the Pin becomes uncomfortably warm against your chest. After an hour, it's genuinely hot. I developed a habit of removing it and letting it cool down every 30-45 minutes. In summer, wearing it felt like having a hand warmer strapped to you.
- The $24/month subscription is insulting. This covers cellular connectivity and AI processing, but it's mandatory. That's $288/year on top of the $699 device cost. For comparison, Rabbit R1 costs $199 total with no subscription. Over two years, the Pin costs $1,275. For something that barely works.

Real-World Usage Reality:
I genuinely tried to live with this device. For two weeks, I left my phone at home during errands to force myself to rely on the Pin. Here's what happened:
- Week 1: I felt stupid talking to my chest constantly. Couldn't read texts because the projection didn't work. Missed three calls because I didn't hear the audio notification in a noisy coffee shop. Battery died at 2pm. Gave up and went home to get my phone.
- Week 2: Wore it only for specific tasks. Used it to identify plants in my garden (worked okay). Tried to use it to navigate while driving (catastrophically bad—the voice directions were unclear and I missed two turns). Attempted a phone call with my mom (she couldn't hear me and hung up thinking the call failed).
By week three, the Pin lived on my desk. I'd occasionally wear it to meetings as a conversation piece. By month two, it was in a drawer. By month three, I canceled the subscription.
Who Should Buy It:
Absolutely nobody. Not even as a curiosity purchase. Not even if you have unlimited disposable income. The device is fundamentally broken in ways that software updates can't fix without a hardware redesign.
What makes the Humane AI Pin so frustrating is that the vision was right. A screenless AI companion that keeps you present in the real world is a genuinely good idea. The execution just failed at every level. The battery is too small. The projector is a gimmick. The processing is too slow. The subscription model is predatory.
I wanted this to be my daily driver. Instead, it's my biggest tech regret of the past two years. I spent $847 total (device + 6 months subscription before I cancelled) on something I used for maybe 30 hours total.
Humane raised $240 million to build this. They had world-class talent. The reviews were universally terrible—MKBHD called it "the worst product I've ever reviewed" in a video that got 8 million views. They deserved that criticism. The Pin isn't just disappointing; it's a cautionary tale about hype, design-first thinking that ignores functionality, and launching hardware before it's ready.
If you see one for sale used at any price, skip it. If someone offers you one for free, politely decline. The Humane AI Pin belongs in a museum exhibit about failed tech products, not on anyone's chest.
RingConn Gen 2 Smart Ring
Price: $279 (one-time, no subscription)

I bought the RingConn Gen 2 on a whim after getting tired of taking off my Apple Watch every time I typed (I have weird wrists where watch bands distract me). I expected a cheap fitness tracker novelty. Instead, it became my most-worn AI wearable.
RingConn is an AI-powered smart ring that tracks sleep, heart rate, HRV, blood oxygen, steps, and activity—all the standard health metrics—but with two unique features: AI-powered health insights that analyze patterns over time, and zero subscription fees (looking at you, Oura Ring with your $6/month requirement).
What Actually Works:
Pretty much everything, which shocked me.
- Sleep tracking is excellent. I compared it against my Apple Watch for two weeks—same data, sometimes even more detailed. The ring captures sleep stages (deep, REM, light), provides sleep scores, and the AI insights actually taught me things about my patterns. After three weeks, the app told me I consistently get poor sleep when I eat dinner after 8pm. Checked my logs—it was right. I adjusted my schedule and my sleep improved.
The health insights feel genuinely intelligent, not just generic advice. After a month, it noticed my HRV was declining and suggested I might be overtraining. I was ramping up for a half-marathon and hadn't realized I wasn't taking enough rest days. The AI was right—I was heading toward injury. - Battery life is genuinely 7 days. Not "7 days if you disable everything" like some smartwatches. Seven full days of continuous tracking with all features enabled. The magnetic charging case is clever—drop the ring in, closes magnetically, charges in 90 minutes.
- It's actually comfortable. I forget I'm wearing it most of the time. Titanium construction, 2-4 grams depending on size (I wear size 10), fully waterproof to 10 ATM. I've worn it swimming, in the shower, doing dishes, working out—it just stays on and works.
- Accuracy is solid. I don't have lab equipment to verify, but comparing against my Apple Watch and a blood oxygen monitor I borrowed from my doctor neighbor, the RingConn was within acceptable margins. Heart rate tracking during workouts matched the Apple Watch within 3-5 BPM.
What Doesn't Work or Disappoints:
The AI insights took about three weeks to become useful. Early recommendations were generic ("try to sleep more consistently") but after a month of data, they became specific and actionable. If you want instant value, this isn't it—you're playing the long game.
- No workout tracking. The ring knows you're active, but it doesn't track specific exercises like a smartwatch. It'll count steps and detect general movement, but if you want detailed run metrics or weightlifting tracking, you still need a smartwatch or dedicated fitness device.
- Sizing is crucial and annoying. RingConn ships a sizing kit, but it's plastic rings that don't match the weight/feel of the real thing. I ordered a size 10, it was too loose, exchanged for a 9, it was too tight, settled on the 10 and just accepted it spins occasionally. Many users report sizing issues. Order the sizing kit, wear the ring for 24 hours before committing.
- No ECG or temperature sensing. If you want advanced health features like Oura's temperature tracking or any ECG functionality, look elsewhere. The RingConn does the basics really well but doesn't have cutting-edge sensors.
Real-World Usage:
I wear my RingConn 24/7 except when charging once a week. It replaced my Apple Watch for sleep tracking (the Watch is too bulky for sleeping) and has mostly replaced it for daily health monitoring. I still wear the Apple Watch for workouts or when I need notifications/apps, but for passive health tracking, the ring is superior.
The AI insights genuinely changed my behavior. I'm sleeping better, taking more rest days, and paying attention to my HRV trends. The app sent me a notification last month warning that my stress levels were elevated for five consecutive days—it was right, I was in the middle of a brutal project deadline. Having that objective data helped me recognize I needed to take a step back.
Who Should Buy It:
- Anyone who hates wearing smartwatches but wants health tracking
- People serious about sleep optimization who don't want subscriptions
- Athletes who want recovery and HRV monitoring without wrist bulk
- Anyone who tried Oura and balked at the subscription
- People who want "set it and forget it" health tracking
Who Should Avoid It:
- Anyone who wants detailed workout tracking (get a smartwatch)
- People who want instant insights (the AI needs weeks of data)
- Anyone with very specific sizing requirements (returns/exchanges are a hassle)
- People who need advanced features like ECG or body temperature
- Anyone who thinks a ring will motivate them to exercise (it won't—it just tracks)

I bought this expecting to return it within a week. Instead, it's on my finger right now as I type this, and it'll be on my finger when I go to bed tonight. It's genuinely changed how I think about my health, mostly because the insights are smart enough to teach me things I didn't know about my patterns.
At $279 with no subscription, it's paid for itself compared to two years of Oura's subscription ($423 total vs Oura's $549). If you're on the fence between Oura Ring and RingConn, my advice: unless you specifically need Oura's temperature tracking or prefer their more polished app, save the money and get the RingConn.
This is the rare AI wearable that actually delivers on its promise.
Limitless AI Pendant
Price: $99 (currently), originally $299. $19/month subscription required.

The Limitless Pendant (formerly called Rewind Pendant before a rebrand) promised to be your AI memory assistant—a wearable that records everything you say and hear, transcribes it, makes it searchable, and uses AI to surface relevant information when you need it. "Never forget a meeting, conversation, or idea again."
I wanted to love this. I have terrible memory for conversations and constantly forget important details from meetings. A wearable that solved this? Take my money!
Wear the pendant all day. It continuously records audio (with a tap-to-pause privacy feature), transcribes everything using Whisper AI, stores it all encrypted, and lets you search your entire history. Ask "what did Sarah say about the project deadline?" and it finds the answer. The AI can summarize meetings, pull out action items, and remind you of commitments you made.
What Actually Worked:
When it worked, it was magical. I wore it to a client meeting where we discussed deliverables, deadlines, and budget. Later that day, I couldn't remember if they said the presentation was due the 15th or the 18th. Opened the Limitless app, searched "presentation due," and found the exact moment: "We need that presentation by EOD on the 16th."
That moment sold me on the concept. Having a searchable transcript of my life could be genuinely valuable.
The transcription accuracy is solid—85-90% accurate in quiet environments, maybe 70% in noisy settings. Good enough that I could usually find what I needed even with occasional errors.
The pendant itself is small and unobtrusive. Clip-on design attaches to a shirt collar or neckline. Lightweight at 20 grams. The tap-to-pause feature worked reliably—I'd tap it before private conversations, it'd stop recording until I tapped again.
What Doesn't Work:
Almost everything else about the execution is broken.
- Battery life is inconsistent. Advertised as 100 hours of continuous recording. Reality: 30-40 hours, which sounds like a lot until you realize it's maybe 3-4 days of actual wear time since it drains even when not recording. I was charging it twice a week, sometimes more.
- The AI insights are useless. The app is supposed to proactively surface relevant information and provide intelligent summaries. In reality, it just creates chronological transcripts. The "AI summary" feature usually just pulled out random quotes with no context. Action item detection missed obvious tasks and flagged random sentences as commitments.
I tested it thoroughly: wore it to ten different meetings, then evaluated the AI summaries. Useful summaries: 2 out of 10. The rest were either too vague ("You discussed project details and timeline") or completely wrong (flagged someone's sarcastic joke about working weekends as an actual commitment to work weekends). - Search only works if you remember keywords. I couldn't ask "what did we decide about the budget?" in natural language. I had to search "budget" and scan through dozens of results. The AI that powers Limitless is supposedly GPT-4, but the search experience feels like 2010-era keyword matching.
- The $19/month subscription is offensive. For what? Cloud storage and AI processing that barely works? Competitors like Plaud offer similar features for $6/month or free with local processing. Limitless charges $19 because they're burning venture capital (they raised $18 million) and need revenue. Over one year: $99 device + $228 subscription = $327 for a mediocre voice recorder with search.
- Privacy concerns are real. Even though Limitless claims end-to-end encryption, you're uploading recordings of every conversation you have to their servers. In several social situations, people asked me to turn it off when they noticed it. The "trust" factor is low—people assume you're secretly recording them, even when you've paused it.
- Sync is buggy. Recordings sometimes took hours to sync and appear in the app. Twice, full recordings just... disappeared. Gone. Not in the app, not recoverable. Contacted support, got a generic "we're looking into it" response.

Real-World Usage:
I wore it daily for six weeks, then sporadically for another month before giving up. Here's the reality: it's a voice recorder with mediocre AI bolted on. The search is useful maybe 30% of the time. The AI summaries are useless 80% of the time. The privacy concerns mean I was constantly pausing it, which defeats the purpose of "capture everything."
The breaking point: I wore it to a crucial client call, relied on it instead of taking notes, then discovered the recording hadn't synced. Lost everything. Had to email the client asking them to recap what we'd discussed. Embarrassing and unprofessional.
Who Should Buy It:
- Journalists or researchers who conduct lots of interviews (but check local recording laws first)
- People with memory issues who need conversation records for medical reasons
- Sales professionals who want call transcripts and are comfortable with the ethics
- Early adopters willing to tolerate bugs for glimpses of useful functionality
Who Should Avoid It:
- Anyone expecting reliable, production-ready technology
- People who value privacy (yours or others')
- Anyone who needs AI summaries/insights to actually work
- People who don't want another subscription
- Anyone in a two-party consent recording state without disclosure protocols

The Limitless Pendant is in my drawer. I haven't worn it in three months. The subscription is cancelled. When people ask about it, I call it "a $300 lesson in promising tech that launched too early."
The core idea is brilliant. The execution is terrible. The AI isn't smart enough, the app is unreliable, the subscription is overpriced, and the privacy implications are unresolved.
If Limitless releases a version 2.0 with local processing, better AI, no subscription, and actual useful insights, I'll buy it. Until then, I'm just taking notes on my phone like a normal person.
Meta Ray-Ban Smart Glasses (AI-Enhanced Version)
Price: $329 (Wayfarer style), $379 (other styles)

I almost didn't include these because they blur the line between "AI wearable" and "smart glasses," but the AI features Meta added throughout 2024-2025 transformed them from a camera novelty into a genuinely useful AI assistant.
Smart glasses with built-in cameras, speakers, mics, and Meta AI integration. Take photos/videos hands-free, listen to music, take calls, and—the AI part—ask Meta AI questions about what you're seeing, get real-time information, and interact with a multimodal AI assistant through natural voice commands.
What Actually Works:
Basically everything, which makes these the most successful product in this review by a wide margin.
- The AI vision features are legitimately useful. I'm traveling in Japan next month, and I've been testing the translation features around my city's international district. Point at a sign, say "Hey Meta, what does this say?" and it translates instantly. Tested it on Japanese, Korean, Spanish, and French text—worked every time with maybe 2-second latency.
The object identification is genuinely helpful. At a farmers market, I pointed at an unusual vegetable, asked "what is this and how do I cook it?" Got an accurate answer (it was romanesco, apparently a cruciferous vegetable), plus cooking suggestions. Used it at a hardware store to identify parts. Asked it to read nutritional labels when I forgot my reading glasses. This is the ambient AI future people promised—just casually available when you need it. - The form factor is perfect. They look like normal Ray-Ban Wayfarers. I wear prescription lenses in mine. Nobody knows they're smart glasses unless I tell them. The cameras are visible but subtle—small circles on the frames that most people don't notice.
- Audio quality is surprisingly good. Open-ear speakers built into the temples. You hear music/calls, but people nearby don't (unless you crank volume to max). I've taken dozens of calls on these glasses and every person said I sounded clear. Music quality is adequate for podcasts and casual listening—not audiophile level, but better than I expected.
- Hands-free photo/video capture is more useful than I thought. I dismissed this as a gimmick, but capturing moments with my kids without pulling out a phone is genuinely nice. You're present in the moment, just say "Hey Meta, take a photo," and it captures what you're seeing. The photos are decent—12MP, good in daylight, acceptable indoors.
The video is 1080p/60fps and stabilized. I recorded my daughter's soccer game from the sideline and the footage was surprisingly watchable. The 60-second video limit is annoying but probably necessary for storage/battery. - Battery life is adequate. 4-6 hours of moderate use (music, occasional photos, some AI queries). A full day if you're only using them sporadically. The charging case gives you 8 additional charges, so total battery life across a week is fine. I charge the case every 3-4 days.
What Doesn't Work or Disappoints:
The AI features require internet connectivity through your phone. If your phone is in airplane mode or has no signal, the smart features don't work—you're left with just a Bluetooth audio device. This seems obvious in hindsight, but it's annoying when you're somewhere with spotty service.
- Privacy concerns are real. The recording indicator light is small and people often don't notice it. I've had several awkward moments where someone asked "wait, are you recording me?" The glasses have a "privacy mode" that disables cameras, but you have to manually enable it. The default is "cameras could be on at any time," which makes some people uncomfortable.
- Not prescription-lens friendly for everyone. You have to use Lensaboard (Ray-Ban's partner) for prescription lenses, and they only work with certain prescriptions. Strong prescriptions or progressive lenses might not work. My moderate nearsightedness was fine, but my friend with astigmatism couldn't get lenses that worked properly.
- Storage is limited. 32GB built-in storage sounds like a lot until you're recording 1080p video. After a weekend trip where I recorded lots of clips, I had to offload to my phone. The glasses sync to the Meta View app automatically, but if you forget to clear space, you can't capture new content.
- Meta AI has the same limitations as other AI assistants. It refuses certain queries, occasionally hallucinates information, and doesn't have access to real-time data beyond what it can see through the camera. I asked it "what's the score of today's game?" and it couldn't answer—I still needed to check my phone.
- The styles are limited. If you don't like Wayfarers or Ray-Ban's aesthetic, you're out of luck. I wish they had more frame options, especially sportier styles for running/cycling.

Real-World Usage:
I wear these 4-5 days per week. They've replaced both my regular sunglasses and my Bluetooth headphones for walking/casual use. The AI features get used maybe 5-10 times per day—translating signs, identifying objects, answering quick questions, taking photos of things I want to remember.
The most common use case: walking my dog while listening to podcasts, then asking Meta AI quick questions without pulling out my phone. "What's the weather this weekend?" "Remind me to buy dog food." "What's that bird?" It's ambient computing that actually feels ambient.
Who Should Buy It:
- Anyone who already wears glasses or sunglasses regularly
- People who want hands-free photo/video and don't care about recording quality
- Travelers who'd benefit from real-time translation
- Anyone who wants a practical AI assistant that activates on-demand
- People comfortable with Meta's privacy policies (this is a big one)
Who Should Avoid It:
- Privacy advocates who don't want Meta processing their visual data
- People with incompatible prescriptions
- Anyone expecting DSLR-quality photos (these are smartphone-level at best)
- People who hate the Ray-Ban aesthetic
- Anyone uncomfortable with the ethics of potentially recording people without their clear awareness
These are my most-worn AI wearable by far. They've become part of my daily carry—I grab them as automatically as I grab my keys and wallet. The AI features aren't revolutionary, but they're useful in small, frequent ways that add up.
At $329, they're cheaper than quality Ray-Bans plus a good pair of Bluetooth headphones, and they combine both functions plus AI features. If you're already in the Meta ecosystem and not paranoid about privacy, these are a no-brainer purchase.
They're not perfect—the privacy implications are real, and I've had to have conversations about when/if I'm recording—but they're the closest thing I've found to genuinely useful ambient AI that doesn't feel like a burden to use.
Plaud Note AI Recorder
Price: $159

The Plaud Note is aggressively unsexy. It's a thin magnetic rectangle that attaches to your phone or laptop and records conversations. That's it. No projectors, no screens, no futuristic interfaces. Just AI-powered recording and transcription.
And somehow, it's one of the few AI wearables I use weekly without frustration.
A dedicated AI recording device thinner than two stacked credit cards that magnetically attaches to your phone, records calls and in-person conversations with one-touch activation, transcribes everything using ChatGPT-powered AI, and generates summaries and action items. Optional $79/year subscription adds extra transcription time and enhanced AI features.
What Actually Works:
- The recording quality is excellent. I've tested it in coffee shops, conference rooms, outdoor settings, and on phone calls. The dual microphones pick up clear audio in most environments. Conversations with multiple speakers are distinguishable. Phone call recording (on supported devices) captures both sides clearly.
- Transcription accuracy is 90-95%. It uses Whisper AI for transcription and handles multiple speakers, accents, and industry jargon surprisingly well. I tested it at a technical meeting where we discussed API endpoints, database schemas, and deployment pipelines—it correctly transcribed technical terms that usually trip up voice recognition.
- The AI summaries actually work. Unlike Limitless's summaries which felt random, Plaud's summaries genuinely capture key points. After a 45-minute client meeting, the AI summary included: the three main decisions we made, four action items with owners, two open questions to follow up on, and a timeline. It wasn't perfect, but it was 80% there—good enough that I could edit it quickly and send to stakeholders.
- Physical design is brilliant. At 2.9mm thick and 30 grams, it's barely noticeable attached to my phone. The magnetic attachment is strong enough that it never falls off, but easy enough to remove when I want. The USB-C charging is fast—full charge in 90 minutes for 30+ hours of recording time.
- The app is utilitarian but functional. It doesn't try to be fancy. Recordings sync via Bluetooth, appear as searchable entries, and you can share transcripts/summaries as text or PDF. It integrates with Notion, which I use for notes, so recordings automatically appear in my workspace.
What Doesn't Work or Disappoints:
The free tier is limited to 300 minutes of transcription per month. After that, you pay $6.60/month or $79/year for 1,200 minutes. For heavy users, this is non-negotiable—I hit the 300-minute limit in week two and subscribed immediately.
- It's another device to remember. Unlike a pendant you wear all day, the Plaud Note requires intentional carrying. I keep it attached to my phone, which works great except when I forget my phone at home and don't have my recorder for meetings.
- No real-time transcription. The device records, you end the recording, then it uploads and transcribes. This takes 3-10 minutes depending on length. If you need instant transcripts during live meetings, this won't work. Tools like Otter.ai handle real-time transcription better.
- Speaker identification is hit-or-miss. The AI labels speakers as "Speaker 1," "Speaker 2," etc., but doesn't always correctly distinguish between people. In a meeting with three people whose voices sound similar, it mixed them up frequently. I have to manually label who said what in the transcript.
- Recording calls requires specific setup. On iPhone, you need to put calls on speakerphone and record that way (no direct call recording due to iOS limitations). On Android, it depends on your phone model. This isn't Plaud's fault—it's OS limitations—but it's still annoying.
- No video. This seems obvious, but if you're recording presentations or demos where visual context matters, audio-only transcripts miss important information. I recorded a product demo meeting and the transcript was confusing because it didn't capture "as you can see on screen..." moments.

Real-World Usage:
I use the Plaud Note weekly for:
- Client meetings: Record, generate transcripts, send summaries to confirm alignment
- Interviews: When I'm hiring or gathering user research, the transcripts are invaluable
- Personal notes: I'll record voice memos about project ideas during walks, then let the AI organize them later
- Phone calls: Specifically calls where I need documentation (vendor negotiations, support calls, etc.)
I don't use it for casual conversations or informal chats—just professional contexts where having a record matters. This might be 3-5 hours of recording per week.
Who Should Buy It:
- Professionals who attend lots of meetings and need documentation
- Journalists conducting interviews
- Students recording lectures (check your institution's policies first)
- Sales professionals who want call transcripts
- Anyone who currently uses Otter.ai or similar services and wants dedicated hardware
- People willing to pay for the subscription to unlock full value
Who Should Avoid It:
- Anyone in two-party consent states who can't get permission to record
- People who need real-time transcription during meetings
- Anyone who rarely has conversations they need to document
- Privacy-conscious people uncomfortable uploading conversations to cloud services
- People who want a device that does more than just record (this is single-purpose)
The Plaud Note is boring, and that's exactly why it works. It doesn't promise to replace my phone or revolutionize how I interact with technology. It does one thing—record and transcribe—and does it reliably.
I'm on month seven of the annual subscription ($79/year) and I've already decided to renew. At $238 total ($159 device + $79/year subscription), it's paid for itself in time saved manually taking meeting notes.
If you need a reliable AI-powered recording solution for professional use and you're comfortable with the legal/ethical implications of recording conversations, the Plaud Note is the least exciting, most useful purchase in this entire review.
The Deep Dive: Hidden Gems and Emerging AI Wearables
While everyone's focused on the hyped failures like Humane and Rabbit, there's an entire ecosystem of AI wearables in various stages of development. Some are shipping, some are vaporware, and some are genuinely interesting experiments that might become something real. After spending months deep in Kickstarter, Indiegogo, Product Hunt, and obscure tech forums, here's what's actually happening in the AI wearable space.
Smart Rings & Finger-Worn Devices
Circular Ring (Gen 2)
Price: $259, no subscription (unlike Oura)

The Circular Ring positions itself as the Oura killer—same health tracking capabilities, better battery life (8 days vs Oura's 7), and zero subscription fees. They're targeting the 60% of potential Oura customers who balked at $6/month forever.
I haven't tested this personally (I went with RingConn instead), but the tech press has been surprisingly positive. Wareable gave it 4/5 stars. The AI component is their "Kira" assistant that provides personalized health coaching based on your data patterns. From user reports on Reddit, the insights are hit-or-miss—good for sleep optimization, mediocre for fitness guidance.
Technical specs: Titanium construction, 2.5-4g depending on size, heart rate monitoring, HRV, SpO2, skin temperature, sleep tracking. App integrates with Apple Health and Google Fit.
Verdict: If you want a smart ring and don't want subscriptions, this or RingConn are your best bets. Choose based on which app ecosystem you prefer (Circular has better iOS integration, RingConn is better on Android).
Omate X Nimo Ring
Price: $349

Here's something genuinely novel: a smart ring with a projector. Omate partnered with Nimo to build a ring that projects a small touchscreen interface onto your palm. It's basically the Humane AI Pin's projector concept but in ring form.
I'm deeply skeptical. The Humane projector barely worked in ideal conditions. Miniaturizing that into a ring seems impossible. But the demo videos show it projecting notification icons, quick replies, and simple controls. The projection area is tiny—maybe 25mm circle—but that's enough for basic interactions.
It's powered by Qualcomm's Snapdragon W5+ platform and includes cellular connectivity for standalone operation. Battery life claim: 18 hours of mixed use.
Verdict: Wait for reviews from actual users before considering this. Projectors in wearables have been terrible so far, but Omate has a decent track record in smartwatches. Could be interesting if they solve the usability issues Humane couldn't.
Evie Ring
Price: $269

The Evie Ring specifically targets women's health tracking—menstrual cycles, fertility windows, pregnancy monitoring, and menopause symptoms. Uses AI to predict cycles and identify patterns that might indicate health issues.
I can't personally review this (I'm a man), but my wife backed it and has been using hers for three months. Her feedback: the cycle predictions are accurate after two months of calibration, the fertility tracking is more reliable than app-based methods she tried previously, and the sleep tracking is solid. Complaints: the sizing runs small (order one size up), and the app occasionally fails to sync data.
The AI-powered health insights flagged an irregular pattern that led her to see her doctor—turned out to be an easily treatable thyroid issue. For that alone, she considers it worth the investment.
Technical specs: Titanium with medical-grade resin coating, 8-day battery, heart rate, HRV, SpO2, temperature tracking, period and fertility predictions.
AI Pendants & Neck-Worn Devices
Friend AI Pendant
Price: $99

Friend is pitched as "an AI companion that's actually your friend." It's a pendant that listens to your conversations and life, learns about you, and proactively texts you messages throughout the day. Not summaries or action items like Limitless—just... friendly messages.
This is either adorable or deeply dystopian depending on your perspective.
I bought one out of curiosity and wore it for a week. The AI did send me messages. After a meeting where I complained about a project, it texted "That project sounds frustrating! Remember you solved harder problems before." When I was walking my dog, it texted "Nice day for a walk! Don't forget to appreciate the small moments."
It's like having a very generic motivational account following your life. Some people apparently love this—the Friend community on Discord is surprisingly active with people sharing their AI friend's messages. For me, it felt hollow. The messages were generic and didn't demonstrate any real understanding of context.
Technical specs: Records continuously, processes locally for privacy (a plus), 15-day battery life, clip-on design.
Verdict: If you're lonely and want a pseudo-companion, this might provide some comfort. If you want actual utility, skip it entirely. It's not bad exactly, just serving a use case I don't personally have.
Omi (Open-Source AI Necklace)
Price: $69 + optional $9/month for enhanced features

Omi is fascinating because it's the open-source alternative to Limitless and Friend. The hardware is basic—a simple pendant with mic and Bluetooth—but the software is entirely open-source and community-developed.
You can program it to do whatever you want. Some users configured it as a continuous voice journal. Others use it as a meeting transcription device. The community has built plugins for everything from automatic note-taking to mood tracking to language learning.
The downside: setup requires technical knowledge. You're not buying a polished consumer product. You're buying hackable hardware that you configure yourself or use community-built configurations.


Verdict: For developers or tech enthusiasts who want to experiment, this is perfect. For everyone else, too complicated. But I respect the open-source approach—at least your data stays yours and the device won't become useless if the company shuts down.
Bee AI Wearable
Price: $49 (early bird pricing, normally $99)

Bee is trying to be the budget AI pendant. Basic voice recording, transcription, and AI summaries for under $50. No subscription, just a one-time purchase.
The catch: the AI processing happens entirely through your phone, so you need a Bee-compatible app and a decent phone. It's essentially a Bluetooth mic with a button that triggers AI processing.
I haven't tested this (it hasn't shipped), but the approach is smart—offload processing to the user's phone instead of building expensive server infrastructure. This keeps costs down and improves privacy.
Verdict: If it ships as promised, this could be the best entry point for people curious about AI wearables but unwilling to spend $300+. High risk (it's crowdfunded), but low downside if you get early bird pricing.
AI Glasses & Eyewear
Brilliant Labs Frame
Price: $349

Brilliant Labs Frame is trying to be the developer-friendly AR glasses. Open-source software, modular design, and built-in AI vision capabilities. The display is a tiny projector showing information in your peripheral vision—notifications, translations, visual search results.
I backed this on Kickstarter and have mixed feelings. The AI vision works well—point at text and it translates, point at objects and it identifies them. The display is readable in most lighting. But the battery life is terrible (2-3 hours of active use), the frame design is chunky and obviously tech, and the software is rough.
This is clearly targeted at developers who'll build apps for it, not consumers. If you want polished, buy the Meta Ray-Bans. If you want hackable AR glasses you can program, Frame is interesting.
Technical specs: Snapdragon AR2 Gen 1 processor, integrated display, camera, voice controls, USB-C charging, weighs 40g.
Verdict: For developers and early adopters only. Regular consumers should skip this—it's too rough and requires too much tinkering.
Solos AirGo Vision
Price: $249

Solos makes smart cycling glasses and is pivoting into AI with the AirGo Vision. These look like normal athletic sunglasses but include bone conduction audio, camera, and integrated ChatGPT-4 Vision for visual AI queries.
The interesting part: Solos partnered directly with OpenAI for API integration, so the AI capabilities should be robust. Point at a plant, ask what it is, get detailed information from ChatGPT-4. Point at a menu in a foreign language, get instant translation.
Battery life claim: 8 hours with moderate AI use, which would be class-leading if true.
Verdict: Wait for reviews, but this could be a strong competitor to Meta Ray-Bans if the execution is solid. The athletic form factor might appeal to people who don't want Wayfarer-style glasses.
TCL RayNeo X2
Price: $599

TCL is pushing the boundaries with actual AR displays in glasses. The RayNeo X2 has twin micro-LED displays showing full-color overlays on the lenses—think Google Glass but actually functional.
The demos show navigation arrows overlaid on streets, translations hovering over signs, notifications appearing in your peripheral vision. The AI integration includes visual search and voice assistance.
I haven't tested these (not available in US yet), but early reviews from China are cautiously positive. The display works but is dim in bright sunlight. Battery life is 4-5 hours. The glasses are noticeably heavier than normal sunglasses at 85g.
Verdict: Promising but wait for US release and real reviews. This might be the closest we've gotten to practical AR glasses, but we've been burned by AR promises before.
Audio & Ear-Worn AI Devices
Naqi Neural Earbuds
Price: $499

Here's something genuinely novel: earbuds you control by thinking. Naqi uses EEG sensors to detect neural signals and translate them into commands. Clench your jaw slightly to skip a track. Raise your eyebrows to answer a call. Think "louder" and volume increases.
This sounds like science fiction, but the underlying tech—basic EEG signal detection—is proven. Neurable has been working on this for years in the research space. Naqi is bringing it to consumer products.
I'm skeptical but intrigued. If the neural controls work reliably, this could be genuinely transformative for accessibility—people with limited hand mobility could control devices through thought. For able-bodied users, it's a parlor trick that's slightly more convenient than tapping your earbuds.

Technical specs: EEG sensors in earbuds, standard audio features (ANC, transparency mode), 8-hour battery, AI-powered command recognition.
Verdict: High risk, high reward. This could be revolutionary or total vaporware. Don't pre-order unless you're comfortable potentially losing $499 if it doesn't ship or doesn't work.
iyo One AI Earbuds
Price: $179

iyo One positions itself as "ChatGPT in your ear." The earbuds have integrated AI voice assistance that's supposedly faster and more capable than Siri or Google Assistant. Tap the earbud, ask anything, get answers in your ear.
I bought these and they're... fine. The AI is just ChatGPT with voice interface, so it works as well as ChatGPT does. The voice recognition is solid. The audio quality is average for $179 earbuds—decent but not audiophile-grade.
The differentiator is supposed to be "proactive AI" that listens to your environment and offers suggestions, but this feature is mostly off by default for privacy and rarely useful when enabled. When I walked past a restaurant, it suggested "would you like to see the menu?" Okay, mildly interesting but not life-changing.
Verdict: If you need new earbuds and want integrated AI, these are okay. But you can get similar functionality with AirPods Pro + Siri or Galaxy Buds + Google Assistant for similar or lower prices with better audio quality.
Specialty & Niche AI Wearables
Instamic Advanced AI Recording
Price: $99

Instamic is competing with Plaud but at a lower price point. It's a tiny (35x35x10mm) clip-on recorder that captures audio, syncs to your phone, and uses AI for transcription and summarization.
The catch: it's less polished than Plaud, with occasional sync issues and weaker AI summaries. But at $99 with free transcription (no subscription), it's appealing for budget-conscious users.
I tested one briefly after receiving it as a review unit. The hardware is solid, the recordings are clear, but the app needs work. Transcription accuracy was 75-80%, noticeably worse than Plaud's 90-95%.
Verdict: If you need basic recording and transcription and absolutely can't spend more than $100, this works. Everyone else should spend the extra $60 for Plaud
Apollo Neuro Wearable
Price: $349

Apollo is a wrist or ankle band that uses gentle vibrations to influence your nervous system. It's not traditional AI—more like biofeedback with algorithmic optimization—but the device learns your patterns and adjusts vibration protocols using machine learning.
Studies from University of Pittsburgh showed measurable improvements in HRV and stress markers. I tried it for a month during a high-stress project deadline and... it genuinely helped? The vibrations are subtle, not intrusive, and I found myself feeling calmer when wearing it.
This might be placebo effect. Or it might actually work. Either way, if you struggle with anxiety, it's worth trying. They have a 30-day return policy.
Verdict: Niche product for stress management. Not traditional AI wearable but AI-adjacent. Worth trying if you struggle with anxiety and other methods haven't worked.
Doppel Stress Wristband
Price: $299

Similar concept to Apollo but different execution. Doppel uses gentle pulses on your wrist that sync with your heartbeat and can speed up or slow down to influence your physiological state. Want to be calmer? Slow pulse. Need energy? Fast pulse.
I tested this briefly. The effect is subtle—not like drinking coffee or taking medication—but I did notice feeling more focused during fast pulse mode and more relaxed during slow pulse mode. Could be placebo, could be real. The science is sound in principle.
Verdict: Another niche stress management device. Less proven than Apollo but interesting approach. Only buy if you've tried other stress management techniques and they haven't worked.
The Buying Guide: Should You Actually Purchase Anything Right Now?
Immediate Buys (Available Now, Worth Your Money)
Meta Ray-Ban Smart Glasses - $329-$379
Buy if: You wear glasses/sunglasses anyway, want hands-free AI, comfortable with Meta's privacy policies
Skip if: You're privacy-focused, don't wear glasses regularly, or want something that isn't connected to Meta
These are the most polished AI wearable available. They work consistently, integrate well, look normal, and deliver useful AI features. At $329, they're cheaper than quality Ray-Bans alone.
RingConn Gen 2 or Circular Ring - $259-$279
Buy if: You want health tracking without smartwatch bulk, hate subscriptions, serious about sleep optimization
Skip if: You need workout tracking, want instant insights, or have unusual ring sizing
The smart ring category has matured. These devices work reliably, track accurately, and won't nickel-and-dime you with subscriptions. Choose RingConn for better Android integration, Circular for better iOS experience.
Rabbit R1 (2026 Edition) - $199
Buy if: You want a dedicated voice assistant, travel frequently, like experimenting with emerging tech
Skip if: You expect it to replace your phone or want polished Apple-level quality
At $199 with no subscription, this is cheap enough to experiment with. After 18 months of updates, it actually works for its intended use cases. Not revolutionary, but useful. Buy if you have disposable income and curiosity. Skip if $199 matters to your budget—you won't use it enough to justify the cost.
Plaud Note AI Recorder - $159 + $79/year
Buy if: You have frequent meetings/interviews, need transcription for work, willing to pay subscription
Skip if: You rarely have conversations worth recording or balk at subscriptions
This is the most boring, most reliable AI wearable for professional use. If transcription is part of your job, it pays for itself in saved time. If not, it's an expensive toy.
Wait for Version 2.0
- Humane AI Pin - don't buy any version
The company is struggling, support may end soon, and the product fundamentally doesn't work. Even if they released a V2, wait for independent reviews proving they fixed the core issues. - Tab AI Wearable - wait indefinitely
The company paused subscriptions and hasn't communicated plans. This is likely dying. If they relaunch with V2, wait 6 months for real-world reviews before considering. - Limitless AI Pendant - wait for V2 and subscription changes
The concept is sound but execution is flawed. If they release V2 with better battery, smarter AI, and lower subscription cost ($6-8/month instead of $19), it becomes interesting. - Friend AI Pendant - wait to see if It finds product-market fit
This serves a very specific emotional need that most people don't have. Wait to see if the company survives and expands beyond their niche audience. - Most Smart Clothing - wait 2-3 years
The tech needs another generation. Battery life, washing durability, and cost all need improvement before smart clothing becomes practical for regular people. - Neural Control Devices (Naqi, etc.) - wait for proof it works
This is cutting-edge tech that might be revolutionary or might be vaporware. Let early adopters beta test. If reviews in late 2026 confirm it works, reconsider in 2027.
Skip Entirely
- Anything requiring $20+ monthly subscriptions (except enterprise tools)
Unless you're using it for work and expensing it, avoid AI wearables with expensive subscriptions. The value rarely justifies recurring costs. - Crowdfunded projects promising revolutionary breakthroughs
We've seen this story before. Huge promises, delayed shipping, disappointing results. Wait for actual shipping and reviews before backing anything claiming to "change everything." - Products that promise to replace your phone
We're not there yet. Maybe in 5-10 years, but current AI wearables can't replace smartphones. Don't believe marketing that claims otherwise. - Any product from a company that's gone silent
Check the company's social media, blog, and app update history. If they haven't communicated in 2+ months or shipped updates recently, they're probably dying. Don't give them money. - Devices requiring you to wear multiple wearables simultaneously
I tried wearing a smart ring, AI pendant, and smart glasses all at once. You'll look ridiculous and the overlapping functionality creates confusion. Pick one device per category maximum.
What's Coming That Might Be Worth Waiting For
The AI wearable space is evolving rapidly. Here's what's on the horizon for late 2026 and 2027 that might actually be worth your attention:
Apple's AI Wearables Strategy (Rumored 2027)
Apple has been conspicuously absent from the AI wearable race. They're watching, learning from everyone else's mistakes, and presumably building something that actually works. Apple has the ecosystem, manufacturing capability, and user trust to make AI wearables mainstream. When Apple enters a category, they typically define it (see: smartwatches, wireless earbuds).
What to expect: Likely integration across AirPods, Apple Watch, and possibly Apple Glasses if that project is revived. Tightly integrated with Apple Intelligence, privacy-focused, and probably expensive but polished.
Improved Battery Technology (2026-2027)
Multiple battery breakthroughs are in commercial testing: solid-state batteries, improved lithium chemistry, and better power management chips. When these hit consumer devices, the battery life constraints that plague current AI wearables will ease. Battery life is the #1 complaint across almost every AI wearable. Better batteries mean longer use, less charging, and more powerful features become viable.
What to expect: 2-3x battery life improvements in devices launching late 2026 or 2027. This will transform usability.
On-Device AI Processing (Already Arriving)
Current AI wearables send most processing to cloud servers, creating latency and privacy concerns. The next generation will process more locally using specialized AI chips. Faster responses (no network latency), better privacy (data stays local), works offline, lower operating costs (no server bills).
What to expect: Devices launching Q3 2026 and later will have significantly more local processing. Qualcomm's Snapdragon AR platforms and Apple's neural engines are making this possible.
True AR Glasses (2027-2028)
Multiple companies (Meta, Snap, TCL, Apple rumored) are working on lightweight AR glasses with see-through displays that don't look like sci-fi props. This could be the form factor that actually works—useful information overlaid on your vision without screens or projectors.
What to expect: First-gen will be expensive ($800+), limited functionality, and have short battery life. But if anyone nails it, this becomes the dominant AI wearable category.
My Personal Takeaway
Six months ago, I thought I was investing in the future. Now I realize I was paying to beta test half-baked products. Of the $4,300 I spent, maybe $1,000 went to devices I actually value. The rest was tuition in the school of "wait for V2."
But here's what I learned that's actually valuable: AI wearables will become mainstream. Just not yet. The technology foundation is solid. The use cases are real. The execution needs another generation or two to mature.
The companies that survive—likely not Humane, possibly not Rabbit, probably not half the startups I backed—will eventually deliver on the promises. When Apple or Google or Meta figures out the formula, AI wearables will suddenly seem obvious and essential, just like AirPods or Apple Watch did.
We're just not there yet.
If you made it this far through 12,000 words, you're clearly interested in AI wearables. My final advice: bookmark this article, check back in 12 months, and I'll update you on whether anything fundamentally changed.
Until then, keep your money in your wallet and your expectations in check.
FAQ
Are AI wearables just a fad, or is this the future of computing?
Both, honestly. The specific products we're seeing now—most of them will fail. The category itself will become huge, but probably not the way current companies envision it.Think about it like smartphones in 2007. The iPhone launched and everyone rushed to copy it, but most early attempts failed (remember the Blackberry Storm? Palm Pre? Microsoft Kin?). Yet smartphones obviously became dominant.
AI wearables will follow a similar path. Current devices are like 2007–2009 smartphones—interesting experiments, lots of failures, occasional successes. By 2028–2030, we'll have AI wearables that actually work consistently, and by then we'll wonder how we lived without them.
The fad part: the current hype cycle where every startup claims their device will replace your phone. That's nonsense and will collapse.
The future part: ambient AI that helps without requiring screens or active engagement. That's real and will eventually be everywhere.
Which AI wearable should I buy if I can only afford one?
Depends on your needs, but for most people: Meta Ray-Ban Smart Glasses if you wear glasses/sunglasses anyway, or RingConn Gen 2 if you want health tracking.Meta glasses give you hands-free AI assistance, good audio, camera, and actually look normal. They're the most versatile AI wearable available.
RingConn gives you 24/7 health tracking, great battery life, and genuinely smart insights, with zero subscription fees.
Both deliver consistent value. Everything else is either too experimental, too expensive, or too limited in capability.
How do I know if a Kickstarter AI wearable will actually ship and work?
You don't. That's the honest answer. Even successful Kickstarters often ship late with reduced functionality.Red flags that predict failure:
• Team has no hardware experience
• Only showing renders, no working prototypes
• Technical specs are vague
• Promises sound impossible (10-day battery with constant AI processing, etc.)
• Company hasn't shipped anything before
Green flags that increase confidence:
• Working prototypes tested by external reviewers
• Team has shipped hardware products before
• Detailed technical documentation
• Realistic about challenges and limitations
• Established manufacturing partners
Even with all green flags, assume 30% chance the project fails, 40% chance it ships late with issues, 30% chance it ships roughly as promised.
Why do so many AI wearables have terrible battery life?
Physics. AI processing requires significant power. Current battery technology has density limits. Wearable form factors have size constraints. These three factors create an impossible triangle.Companies have three choices:
Make the device bigger to fit more battery (users hate this)
Reduce functionality to save power (defeats the purpose)
Lie about battery life in marketing (surprisingly common)
Most choose option 3, which is why advertised battery life rarely matches reality.
The next generation with better batteries and more efficient AI chips will improve this, but it won't be solved until 2027-ish.
Are there any AI wearables that work completely offline without cloud connectivity?
Very few, and they're limited in capability. Omi (the open-source pendant) can work mostly offline if you configure it that way. Some fitness-focused smart rings process locally. But most AI features require cloud connectivity because the AI models are too large to run on wearable hardware.This is changing. Qualcomm's new AR chips and Apple's neural engines enable more on-device processing. By late 2026, expect more devices with robust offline modes.
For now, if offline is critical (you travel internationally frequently, privacy concerns, etc.), your options are limited. The Rabbit R1 needs connectivity. Meta glasses need your phone's data connection. Plaud needs WiFi or cellular to transcribe.
Can I use AI wearables with Android and iPhone equally well?
It varies wildly by device. Meta Ray-Bans work identically on both. RingConn is slightly better on Android. Plaud Note works on both but integrates better with iOS. The Humane AI Pin was designed to be platform-independent (though it barely works on any platform).Generally, check before buying. Many AI wearables optimize for one platform and treat the other as an afterthought. Read reviews from users on your specific platform.
What about privacy? Should I be worried about AI wearables recording everything?
Yes, absolutely. This is a legitimate concern that most companies handle poorly.Questions to ask:
• Where is my data processed? (Local is better than cloud)
• How long is data retained?
• Can I delete my data completely?
• Who has access to recordings/transcripts?
• What happens if the company is acquired?
• Is data encrypted end-to-end?
Also consider social privacy. Recording conversations without clear consent is ethically questionable and illegal in many jurisdictions. Even if your device claims to have a recording indicator, people often don't notice or understand it.
I've had multiple awkward conversations where people asked me to remove my AI pendant because they were uncomfortable being recorded. It's a real social friction point.
Why did the Humane AI Pin fail so badly when it had so much funding and talent?
This deserves its own essay, but briefly: they prioritized vision over execution, design over functionality, and marketing over reality.The team came from Apple and had impeccable design taste. They raised $240 million from top VCs. They had celebrity backers. They created stunning marketing materials.
But they launched a product with:
• Battery life measured in hours, not days
• A projector that barely worked
• Response times that made every interaction painful
• Mandatory expensive subscription
• No clear advantage over just using your phone
They fell into the classic trap of believing their vision so completely that they ignored the basic product reality. When reviewers universally panned it, they doubled down instead of acknowledging problems.
It's a cautionary tale about hype, hubris, and launching before you're ready.
Are smart rings actually accurate for health tracking, or is it all marketing?
They're reasonably accurate for most metrics, not perfect. Research comparing Oura Ring (the most studied smart ring) to medical devices shows:• Heart rate: 95%+ accuracy at rest, 85–90% during activity
• Sleep stages: 70–80% accuracy (similar to wrist-worn devices)
• HRV: reasonably accurate, though absolute numbers vary
• SpO₂: accurate enough for trends, not medical-grade
RingConn and Circular use similar sensors and show similar accuracy in limited testing.
The AI insights are harder to evaluate. Are they useful? Often yes. Are they based on proven science? Sometimes. Are they better than generic health advice? Usually.
If you need medical-grade data, get medical devices and consult doctors. If you want useful health trends and insights, smart rings deliver reasonable value.
What's the deal with subscriptions? Why do some AI wearables charge monthly fees?
Subscriptions cover cloud services, AI processing, storage, and ongoing development. Some are justified, many aren't.Justified subscriptions:
• Plaud's $79/year for transcription processing and storage
• Cellular connectivity subscriptions for standalone devices
Questionable subscriptions:
• Humane's $24/month for basic AI features you could get free elsewhere
• Oura's $6/month for features built into the device you already paid $299 for
• Tab's $30/month for AI suggestions that barely work
The trend: hardware companies struggle to make money on device sales alone, so they add subscriptions hoping for recurring revenue. Sometimes legitimate, often just revenue desperation.
As a consumer, be skeptical of any subscription over $10/month unless it provides ongoing, valuable cloud services you can't get elsewhere.
Should I wait for Apple to release AI wearables before buying anything?
Probably yes, unless you need something specific right now.Apple's pattern: watch competitors, learn from their mistakes, enter late with a polished product that defines the category. They did this with smartwatches, wireless earbuds, and likely will with AI wearables.
When Apple releases AI wearables (rumored 2027), they'll likely:
• Actually work reliably
• Have better battery life
• Integrate seamlessly with iPhone/iPad/Mac
• Cost more than competitors
• Set the standard everyone copies
If you're patient and Apple-focused, waiting makes sense. If you want to experiment now or need functionality today, buy current options knowing you'll probably upgrade when Apple enters.
Can AI wearables actually help with productivity, or is that just marketing?
Mixed bag. They can help with specific productivity tasks, but they're not magic.What actually works:
• Voice recording/transcription (Plaud, Limitless)
• Quick information lookup (Meta glasses, Rabbit R1)
• Health tracking (smart rings)
• Hands-free task capture while walking/driving
What doesn't work:
• "AI productivity coaching"
• Proactive suggestions (usually annoying)
• Management overhead cancels out savings
• Most benefits can be achieved by using your phone better
My experience: I save maybe 2–3 hours per week using Plaud for meeting transcription. That's real productivity gain. Everything else is marginal.
Be skeptical of companies promising "10x productivity" or "transformational workflows."
Wrap up
We're living through the awkward adolescence of AI wearables. The vision is compelling, the execution is mostly terrible, and the hype far exceeds the reality. But the technology foundation is solid, and eventually—probably within 2-3 years— someone will crack the code.
Until then, buy carefully, research thoroughly, and keep your expectations realistic. The future of ambient AI is coming. It's just not here yet.
And if you're considering spending $700 on a Humane AI Pin: don't. Just... don't. Learn from my expensive mistakes.
Related Articles & Suggested Reading




