Solos AirGo V2: The Modular AI Smart Glasses — Transcript
Full transcript
- 0:00So, today we are looking at the Solos
- 0:02Airgo V2 today.
- 0:03>> Mhm.
- 0:04>> And this is a device that acts as an AI
- 0:06companion that just well, it simply
- 0:09looks at the real world with you.
- 0:11>> The Airgo V2, it launched at CES for
- 0:13around $299.
- 0:15>> That price point is entirely dictated by
- 0:17their physical design choices. They just
- 0:19look like standard lightweight
- 0:21eyeglasses.
- 0:21>> They do. You can wear them to a cafe,
- 0:24the office, the grocery store, and you
- 0:27just blend in entirely. No one's going
- 0:29to look twice at you.
- 0:30>> But the underlying architecture making
- 0:32that possible is what they're calling
- 0:33the smart hinge system. I found the
- 0:35mechanics of this so fascinating. The
- 0:37frames and the internal electronics,
- 0:39they aren't just glued together into one
- 0:41monolithic piece of plastic. You can
- 0:44actually pop the front frames off the
- 0:45hinges and swap them out to match your
- 0:47outfit, which is neat.
- 0:48>> And more importantly, you can swap
- 0:50between a frame that has a built-in
- 0:52camera and a frame that's completely
- 0:54blank.
- 0:55>> Right, so you have options. Even the
- 0:56temple arms, like the pieces that slide
- 0:59over your ears, they clip right in and
- 1:00out. So, I look at this device, I see
- 1:02the modular frames, the slick design.
- 1:05But if there is absolutely no augmented
- 1:08reality display projecting onto the
- 1:10lenses, can we legitimately call these
- 1:13smart glasses?
- 1:14>> That's the big question.
- 1:15>> Right, like
- 1:16>> I mean, that is the exact debate
- 1:18happening in the tech community right
- 1:19now.
- 1:20>> Oh, really?
- 1:20>> Yeah. And it's a totally valid critique
- 1:22if you're strictly applying the old
- 1:24definition of spatial computing.
- 1:26>> Okay.
- 1:26>> If smart requires a visual display, then
- 1:29no, these aren't smart. But the
- 1:30definition of ambient computing is
- 1:32evolving.
- 1:33>> Because rendering those graphics takes a
- 1:35lot of juice.
- 1:35>> Exactly. Rendering high-fidelity 3D
- 1:38graphics in real time requires a massive
- 1:41graphical processing unit.
- 1:42>> And a massive battery.
- 1:43>> Right, which generates a ton of heat
- 1:45right on your face. So, Solos basically
- 1:47sidesteps that entire thermal and power
- 1:49nightmare.
- 1:50>> Ah, I see. So, by stripping out the
- 1:52screen, they free up all that processing
- 1:54power and battery life to focus entirely
- 1:57on sensory input and AI.
- 2:00>> Precisely. The smart designation here
- 2:02comes from real-world contextual
- 2:05computing.
- 2:06>> Meaning what, exactly?
- 2:07>> Meaning it's not about projecting a
- 2:09little digital weather widget into your
- 2:12field of view, it's about the glasses
- 2:14knowing you are staring at dark clouds
- 2:16and answering you when you ask, "Hey, do
- 2:18I need an umbrella today?"
- 2:19>> Okay, that makes sense. Which means if
- 2:21this AI is going to process the physical
- 2:23world, it desperately needs to see.
- 2:25>> In other eyes.
- 2:26>> Right. And [snorts] the front frames
- 2:28house this upgraded 16-megapixel
- 2:31ultra-slim camera.
- 2:32>> Which is no small thing.
- 2:33>> Yeah, from a pure physics standpoint,
- 2:35fitting a 16-megapixel sensor into a
- 2:37standard eyeglass frame without, you
- 2:39know, burning the user's face is a
- 2:40serious engineering feat.
- 2:42>> When you compress optical sensors and
- 2:44image processors into a space that tiny,
- 2:46they generate immense localized heat.
- 2:48>> Right, they get super hot. Ow.
- 2:50>> Yeah, wicking that thermal energy away
- 2:52from the camera module and distributing
- 2:55it along the frame so it just dissipates
- 2:56harmlessly into the air.
- 2:58>> That's wild. And it's not just taking
- 3:00grainy still photos, either. We're
- 3:01talking full HD video recording,
- 3:03high-res photos, and even live video
- 3:06streaming.
- 3:06>> Yep.
- 3:07>> It's essentially like a dashcam for your
- 3:09life, but one that is directly wired
- 3:11into a global encyclopedia.
- 3:14>> A dashcam is a really brilliant way to
- 3:15visualize it, but uh think about the
- 3:18reality of
- 3:19a camera mounted to human head for a
- 3:21second.
- 3:22>> Okay.
- 3:22>> We do not move like cars. We constantly
- 3:25micro-adjust our posture. We nod. We
- 3:27dart our heads around to look at sudden
- 3:29noises.
- 3:29>> Oh, true. Raw video from a head mount is
- 3:31usually pretty nauseating to watch.
- 3:33>> Exactly. It looks like a chaotic found
- 3:35footage horror movie.
- 3:36>> Right, right. So, how do they stabilize
- 3:38that for a live stream if you're walking
- 3:39around with these on?
- 3:40>> They utilize electronic image
- 3:42stabilization or EIS.
- 3:44>> EIS? Okay.
- 3:46>> Right. So, instead of trying to
- 3:47physically move the lens to compensate
- 3:50for your head movement, which would
- 3:52require bulky mechanical gimbals, they
- 3:55do it digitally.
- 3:56>> How does that work?
- 3:57>> The camera captures a slightly wider
- 3:59field of view than it actually needs.
- 4:02Then, uses gyroscope data from the
- 4:04glasses to understand exactly how your
- 4:06head is moving.
- 4:07>> Oh, so it tracks the tilt and panning?
- 4:09>> Exactly. And it dynamically crops and
- 4:12shifts the video frame in real time to
- 4:15counteract your jitter.
- 4:16>> Wow, okay.
- 4:17>> The result is a surprisingly smooth,
- 4:20really watchable video feed.
- 4:22>> That makes the live streaming feature
- 4:23genuinely usable then. But, I mean, that
- 4:26brings us to the elephant in the room
- 4:27whenever we talk about cameras mounted
- 4:29to people's faces.
- 4:30>> Privacy.
- 4:31>> We have to talk about privacy.
- 4:32>> People feel surveilled, the technology
- 4:34will be rejected no matter how good the
- 4:36AI is.
- 4:37>> Which is why the inclusion of a highly
- 4:38visible privacy indicator light on these
- 4:41isn't just like a neat design choice.
- 4:43>> No, it's vital.
- 4:44>> It's a non-negotiable survival mechanism
- 4:46for the product. Whenever the camera is
- 4:49recording or streaming,
- 4:50a distinct light on the frame turns on.
- 4:53>> Yep, it illuminates bright enough so
- 4:54everyone knows.
- 4:55>> So, you cannot covertly film a
- 4:57conversation. It immediately establishes
- 4:59this baseline of trust in public spaces.
- 5:02>> It forces accountability. So, the camera
- 5:05acts as the sensory input, right? It
- 5:06provides the raw data. But, raw data is
- 5:08completely useless without
- 5:10interpretation.
- 5:11To understand how these glasses actually
- 5:13function as an assistant, we have to
- 5:15look at the brains behind the lenses.
- 5:17>> Yes. The operating system they are using
- 5:19is called Solos Chat 3.0.
- 5:21>> Mhm.
- 5:22>> And the defining feature of the system
- 5:23is that it's a multimodal AI.
- 5:25>> Multimodal, yeah.
- 5:26>> Meaning, it isn't just waiting for you
- 5:28to type out a text prompt. It is
- 5:30simultaneously ingesting the live video
- 5:33feed of the street you're on,
- 5:35listening to the ambient audio of the
- 5:37traffic, and processing your voice
- 5:39commands all at once.
- 5:40>> Which is incredible. Multimodal
- 5:42capabilities are what bridge the gap
- 5:44between like a simple search engine and
- 5:47a genuine digital assistant.
- 5:49>> Right.
- 5:49>> But, what makes the Solos' approach
- 5:52unique is that they aren't relying on a
- 5:54single proprietary AI model.
- 5:57>> Oh, right. They opened it up.
- 5:58>> They did. They integrated four distinct
- 6:00major platforms: ChatGPT, Claude,
- 6:03Gemini, and DeepSeek.
- 6:05>> Wait, wait. So, if I have all four of
- 6:06those massive language models loaded in,
- 6:09how does it know which one to use?
- 6:10>> Well, you can choose.
- 6:11>> Really? Can I essentially hot swap the
- 6:13AI's brain depending on what specific
- 6:16task I'm trying to accomplish?
- 6:18>> can, and that is a massive strategic
- 6:19advantage here.
- 6:20>> Wow.
- 6:21>> Because, you know, no single AI model is
- 6:23the best at everything. They all have
- 6:24their own distinct architectural
- 6:26strengths.
- 6:27>> Could you give me an example of that?
- 6:28>> Sure. So, let's say you're looking at a
- 6:29complex engineering schematic.
- 6:32You might route that through Claude
- 6:34because it has a stellar reputation for
- 6:36structured logic and technical analysis.
- 6:38>> Okay, that makes sense.
- 6:39>> But, if you're traveling and need to
- 6:40quickly identify a specific landmark in
- 6:43real time, you might switch over to
- 6:45Gemini since it's deeply integrated with
- 6:48Google's live search index.
- 6:50>> Oh, that is so smart. You're constantly
- 6:52choosing the most specialized tool for
- 6:53the job.
- 6:54>> Exactly.
- 6:55>> And you interact with these brains
- 6:57entirely hands-free via a wake word.
- 7:00But, there's a specific phrase in the
- 7:02slide deck used to describe these
- 7:03interactions that I think is critical.
- 7:06>> Multi-turn natural conversations.
- 7:09>> Yes, multi-turn.
- 7:10>> That is the feature that prevents you
- 7:11from wanting to throw the glasses into a
- 7:13river after 10 minutes.
- 7:15>> Right. Because with older voice
- 7:16assistants, interaction was extremely
- 7:19transactional.
- 7:19>> It was exhausting.
- 7:20>> to say the wake word, ask a question,
- 7:22get your answer, and then say the wake
- 7:24word all over again just to ask a
- 7:26follow-up.
- 7:26>> Yeah. Hey, assistant, what is this
- 7:28building? It is the Louvre. Hey,
- 7:29assistant, when was it built?
- 7:31>> Exactly. It feels like you are
- 7:32interrogating a robot rather than having
- 7:35a dialogue.
- 7:36>> Multi-turn capability changes that
- 7:39entirely by utilizing what's called a
- 7:41continuous context window.
- 7:43>> Okay. How does that work?
- 7:44>> So, the AI actively holds the memory of
- 7:47your ongoing conversation in its
- 7:48short-term token memory.
- 7:50>> Right.
- 7:50>> You ask, "What is this building?"
- 7:52>> Mhm.
- 7:52>> After it answers, you naturally just
- 7:54follow up with, "Who designed it?" And
- 7:56the AI inherently understands that it
- 7:59refers to the Louvre.
- 8:00>> That's huge. The interaction flows
- 8:02exactly like a normal human
- 8:03conversation.
- 8:04>> Exactly.
- 8:05>> Think about the practical applications
- 8:06of that for a second. You're staring at
- 8:08a menu in Tokyo, and it's translating
- 8:10the text in real time.
- 8:11>> Mhm.
- 8:11>> Or, you know, you're at the hardware
- 8:13store staring at a massive wall of
- 8:14screws, totally lost.
- 8:16>> You've all been there.
- 8:17>> Right.
- 8:17>> Yeah.
- 8:17>> You just look at the wall and ask the
- 8:18glasses, "Which one of these do I need
- 8:20for half-inch drywall?" And it guides
- 8:22you.
- 8:22>> But, facilitating that kind of seamless
- 8:25magic takes a massive toll on the
- 8:27hardware.
- 8:28>> I can imagine.
- 8:29>> Having an ongoing, continuous dialogue
- 8:32with a cloud-based AI while
- 8:33simultaneously streaming video requires
- 8:36two critical things,
- 8:38>> which are?
- 8:38>> a private way for the AI to speak back
- 8:41to you, and an enormous amount of power.
- 8:43>> Uh, yeah. That's the ultimate wearable
- 8:45paradox, isn't it? You want infinite
- 8:47battery life, but you don't want to wear
- 8:49a heavy brick on your face.
- 8:51>> Precisely.
- 8:52>> Let's tackle the audio first, because if
- 8:54I'm having a detailed conversation with
- 8:56Claude about a private work email,
- 8:58I really don't want my AI blasting its
- 9:00responses for the entire subway car to
- 9:02hear.
- 9:03>> Yeah, that would be a deal-breaker.
- 9:05But, Solo solved this using directional
- 9:07open-ear speakers built directly into
- 9:09the temple arms.
- 9:10>> So, there's no physical earbud jamming
- 9:12into your ear canal.
- 9:13>> Right, nothing in your ear.
- 9:15>> Which is fantastic for situational
- 9:16awareness. You can still hear traffic or
- 9:18someone calling your name.
- 9:19But, how does an open speaker near your
- 9:22ear remain private?
- 9:24>> It's pretty clever.
- 9:25>> Like, why doesn't the person sitting
- 9:26next to me hear it?
- 9:27>> It relies on a principle called phase
- 9:29cancellation.
- 9:30>> Okay, I've heard of that with
- 9:31noise-canceling headphones.
- 9:32>> Similar concept, yeah. Instead of just
- 9:35pushing sound waves out in all
- 9:36directions like a traditional speaker,
- 9:38these arms use an array of micro
- 9:40speakers firing at precisely timed
- 9:43intervals.
- 9:44>> Okay.
- 9:45>> Imagine dropping two pebbles into a pond
- 9:48where the ripples meet, they either
- 9:49amplify each other or they completely
- 9:51cancel each other out.
- 9:52>> Right, destructive interference.
- 9:54>> Exactly. These speakers essentially
- 9:56construct a targeted acoustic bubble
- 9:59right over your ear canal.
- 10:01>> That is so cool.
- 10:02>> For you, the sound waves arrive
- 10:03perfectly in sync, creating crisp, clear
- 10:06audio. But for the person sitting two
- 10:08feet away, those same sound waves arrive
- 10:11out of phase and literally erase each
- 10:13other in the air before reaching their
- 10:15eardrum.
- 10:15>> Wow, that is brilliant engineering. But
- 10:18okay, processing phase array audio,
- 10:21capturing stabilized video, pinging four
- 10:23different AI servers. The battery drain
- 10:26has to be astronomical.
- 10:28>> Oh, it is.
- 10:28>> Like we're talking about a recipe for a
- 10:29dead device in 45 minutes.
- 10:31>> laws of physics dictate that battery
- 10:33capacity requires physical volume. You
- 10:35cannot cheat the chemistry. But they
- 10:37engineered an incredibly clever
- 10:39mechanical workaround using that smart
- 10:42hinge system we talked about earlier.
- 10:44>> The battery temples.
- 10:45>> Yes.
- 10:45>> This is honestly my favorite hardware
- 10:47design choice on the whole device.
- 10:49>> It's so smart.
- 10:50>> arms out.
- 10:52The pieces resting on your ears actually
- 10:53are the batteries.
- 10:54>> Yes.
- 10:55>> And because they're hot swappable, you
- 10:56don't have to take the glasses off,
- 10:58power them down, and tether yourself to
- 11:00a wall charger when they die.
- 11:01>> No cords required.
- 11:02>> You literally just pop off the dead arm
- 11:05and click on a fully charged one. It's
- 11:07like reloading a high-tech gadget in a
- 11:09spy movie.
- 11:09>> That's a great comparison.
- 11:11>> Click out the old clip, snap in a fresh
- 11:12one, and you're immediately back in the
- 11:14action.
- 11:15>> It bypasses the physical limitations of
- 11:17battery size entirely, and the base
- 11:20battery life on a single set of arms is
- 11:22highly respectable.
- 11:23>> What are they quoting?
- 11:24>> They're quoting about 10 hours of
- 11:26continuous music playback or 7 hours of
- 11:29active call time.
- 11:30>> Okay, wow. Which covers a full work day
- 11:32for most people.
- 11:33>> Easily. And for heavy power users, they
- 11:36designed an optional charging case.
- 11:38>> AirPods case.
- 11:39>> Sort of.
- 11:40But it looks identical to a standard
- 11:43premium eyeglasses case.
- 11:45Inside though, it houses an 1,100
- 11:47milliamp hour battery.
- 11:49>> Oh, nice.
- 11:50>> So you can be actively wearing the
- 11:52glasses and talking to the AI while your
- 11:54spare set of battery arms is sitting in
- 11:56the case inside your bag recharging.
- 11:59>> That's amazing.
- 12:00>> Yeah.
- 12:01>> built in an auto power on and off
- 12:02feature, right? You take the glasses
- 12:04off, you set them on your desk, and the
- 12:06internal gyroscopes realize they aren't
- 12:08being worn, so they instantly put them
- 12:10to sleep to conserve juice.
- 12:11>> And you put them back on, they wake up.
- 12:13>> It's all designed to completely remove
- 12:15the friction of interacting with the
- 12:16technology.
- 12:17>> But you know, power is only half of the
- 12:19equation here.
- 12:20>> What's the other half?
- 12:20>> The other invisible tether we need to
- 12:22discuss is data.
- 12:23>> Ah.
- 12:24>> Processing high-res video and sending it
- 12:27to an AI in the cloud requires moving a
- 12:29massive amount of data extremely fast.
- 12:31>> Right, because to keep the glasses
- 12:33lightweight and affordable, the actual
- 12:35heavy computing isn't happening inside
- 12:36the frames.
- 12:37>> No, there's no room for that processor.
- 12:39>> Your smartphone is acting as the
- 12:40middleman, pulling the data from the
- 12:42glasses and routing it to the cloud.
- 12:43>> Exactly.
- 12:44>> And they use a dual connectivity system
- 12:46to manage this traffic jam, don't they?
- 12:48>> Yeah, it's a very elegant networking
- 12:50solution.
- 12:51For the low-intensity tasks like
- 12:53connecting to the companion app, taking
- 12:55phone calls, streaming a podcast, or
- 12:57just basic voice commands, the glasses
- 13:00use standard Bluetooth.
- 13:02>> Okay.
- 13:02>> Think of Bluetooth as a quiet, narrow
- 13:04country lane. It requires very little
- 13:07power, making it incredibly energy
- 13:09efficient, which is perfect for a slow
- 13:11trickle of audio data.
- 13:12>> But a live 1080p video feed is not a
- 13:15slow trickle.
- 13:15>> Not at all.
- 13:16>> It is a massive convoy of heavy freight
- 13:18trucks.
- 13:19>> Yeah.
- 13:19>> If you try to force that much data down
- 13:22a narrow dirt path,
- 13:23you get massive traffic jams. You get
- 13:25latency.
- 13:26>> And latency kills the experience.
- 13:28>> If the AI is trying to identify a
- 13:29fast-moving object, even a 1-second lag
- 13:33makes the entire system useless.
- 13:34>> Exactly. So, to circumvent that
- 13:36bottleneck, the glasses also include a
- 13:39low-power Wi-Fi connection.
- 13:41>> Oh, built right in?
- 13:42>> Yes. The millisecond you activate a
- 13:44high-bandwidth task, like live streaming
- 13:47or asking the AI to analyze a live
- 13:49visual feed, the glasses dynamically
- 13:52spin up a direct Wi-Fi connection to
- 13:54your phone.
- 13:55>> Wow.
- 13:56>> They essentially build a dedicated
- 13:58multi-lane high-speed highway to
- 14:00guarantee the data transfers instantly
- 14:03without lag.
- 14:04>> This seamless. And all of this routing,
- 14:06the model selection, the device
- 14:08management, it all happens through the
- 14:10Solos Airgo companion app.
- 14:12>> Right. It is entirely dependent on your
- 14:13phone.
- 14:14>> But this connectivity pipeline opens the
- 14:16door to something far more profound than
- 14:18just translating menus or finding
- 14:19drywall screws.
- 14:20>> definitely.
- 14:21>> Solos announced a major integration with
- 14:23Envision, a company that specializes in
- 14:25assistive technology.
- 14:26>> Which is huge.
- 14:27>> We've been talking about this as a
- 14:29productivity hack. But with Envision,
- 14:31they are planning tools like dedicated
- 14:33scene description, text recognition, and
- 14:36object identification designed
- 14:38specifically to help blind or low-vision
- 14:41users navigate the physical world.
- 14:43>> I'd argue it's even more fundamental
- 14:45than a productivity hack. It's a
- 14:47paradigm shift in independence.
- 14:49>> Truly. A visually impaired user puts
- 14:51these glasses on and the AI can read the
- 14:54fine print on a prescription medicine
- 14:56bottle to them.
- 14:57>> Right. Or it can verbally announce what
- 14:59number bus just pulled up to the transit
- 15:01stop.
- 15:01>> a consumer tech gadget into a literal
- 15:04seeing eye assistant.
- 15:05>> And historically, specialized medical
- 15:07devices that perform those exact
- 15:09functions cost thousands of dollars.
- 15:11>> And they often look highly clinical,
- 15:12which carries its own social stigma.
- 15:14>> Exactly. Packing that life-changing
- 15:17utility into a $299 device that just
- 15:20looks like standard eyewear is
- 15:22revolutionary.
- 15:23>> And they're actively encouraging the
- 15:25tech community to push this even
- 15:26further.
- 15:27>> Yeah, that's DK.
- 15:28>> Right.
- 15:29Solos released an Ergo SDK, a software
- 15:32development kit, meaning they are
- 15:34opening up the underlying architecture
- 15:36so third-party developers can build
- 15:37their own custom applications.
- 15:39>> Because hardware is ultimately only as
- 15:41valuable as the software ecosystem that
- 15:43supports it.
- 15:44>> Right.
- 15:44>> By releasing an SDK, Solos is signaling
- 15:47that they don't just want to sell a
- 15:48standalone product. They want to
- 15:50establish a platform.
- 15:52>> So anyone can build for it.
- 15:53>> Yeah. Developers can now build highly
- 15:56specialized enterprise workflows.
- 15:58Imagine an application built for
- 15:59warehouse logistics where workers can
- 16:01identify inventory and cross-reference
- 16:03databases entirely hands-free.
- 16:05>> Oh, that would be incredibly efficient.
- 16:07>> Or specialized navigation apps for
- 16:09university campuses.
- 16:10>> Looking at the reality of what the Ergo
- 16:13V2 offers,
- 16:14does that $299 price tag actually hold
- 16:18up? Let's weigh the pros and the cons.
- 16:20>> I mean, the value proposition here is
- 16:22remarkably strong on the pro side.
- 16:24>> Agreed.
- 16:25>> You're gaining immediate hands-free
- 16:27access to the most powerful multimodal
- 16:30AI models on the planet, OpenAI,
- 16:32Anthropic, Google.
- 16:34>> Plus the hardware itself.
- 16:35>> Right. You're getting a highly modular,
- 16:37repairable hardware design with that
- 16:39brilliant hot-swappable battery system.
- 16:42The camera provides stabilized,
- 16:44genuinely usable footage.
- 16:46>> And achieving all of that at a $299
- 16:48price point makes contextual computing
- 16:50accessible to a massive demographic.
- 16:52>> It really does. But the limitations are
- 16:54equally real, and they're inherently
- 16:56baked into the philosophy of the device.
- 16:58>> glaring con for power users is going to
- 17:02be the absolute lack of an augmented
- 17:04reality display.
- 17:05>> Yeah, that's a hard stop for some
- 17:06people.
- 17:07>> If your primary goal is having digital
- 17:09maps overlay directly onto the pavement
- 17:11or seeing your text messages floated
- 17:13midair, this device fundamentally cannot
- 17:15do that.
- 17:16>> Second is the severe dependency on your
- 17:19smartphone.
- 17:20>> Right, it's not a standalone unit.
- 17:21>> No, these glasses are brilliant sensory
- 17:23inputs, but they are not standalone
- 17:26computers. Your phone is doing all the
- 17:28heavy lifting, routing the data, pinging
- 17:30the cloud.
- 17:31>> Which means this system is going to
- 17:33aggressively drain your phone's battery
- 17:36alongside its own.
- 17:37>> Oh, absolutely.
- 17:38>> And finally, the social friction.
- 17:40Despite the inclusion of that very
- 17:42necessary privacy indicator light, the
- 17:44cultural unease of wearing a camera on
- 17:46your face in public spaces is going to
- 17:48linger.
- 17:49>> It definitely will. It'll take years of
- 17:51normalized use for that discomfort to
- 17:53fully fade. It is a deliberate,
- 17:55calculated trade-off. Solos looked at
- 17:57the landscape and decided that a
- 17:58lightweight, affordable, AI-first
- 18:01wearable was a much faster path to mass
- 18:03adoption than a heavy, incredibly
- 18:06expensive, visually focused headset.
About this transcript
This page contains the full transcript of Solos AirGo V2: The Modular AI Smart Glasses by ProReview, generated from the public captions YouTube serves with the video. The transcript has 3,375 words across 616 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.