Limitless: An AI Podcast - The Rumored OpenAI Device: Is it Really a Pen?

Episode Date: January 7, 2026

🌌 LIMITLESS HQ: LISTEN & FOLLOW HERE ⬇️https://limitless.bankless.com/https://x.com/LimitlessFT------OpenAI's new potential pen-shaped device may be designed for ambient computing and moving beyond smart glasses. We analyze the impacts of over $10 billion in failed AI hardware investments and OpenAI's strategic focus on voice technologies. Exploring the convergence of multimodal AI and societal screen fatigue, we consider how this device could reshape our tech interactions.------TIMESTAMPS0:00 AI's New Hardware Revolution4:08 Speculating the Device's Design8:25 Ambient Computing's Promise12:11 The Cost of Innovation14:52 The Flaws of Smart Glasses17:49 Pickle's Unlikely Reality19:26 OpenAI's Unique Approach23:51 The Future of Device Integration26:25 Looking Ahead to the Future------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

For the last three years, AI has been this disembodied superpowered. It lives in tabs. It lives in chat windows. You summon it, you type to it. Sometimes you can speak to it. It comes and then it vanishes. It's powerful, but it is trapped behind this glass. And then last week, we have one of the most exciting rumors to pop up in this world,

which is around Open AI and their new hardware device offering. Open AI, they partnered with Johnny Ive, who has created devices, including, but not limited to the iPhone, the iPod, the MacBook, basically any hardware device that you've used, Johnny Ive has designed it if it's come from Apple. And what's fascinating about this is the rumor isn't telling us that they're going to be choosing glasses. Everyone's choosing glasses. Meta has them. Google's working on them. Apple's working on them. But this form factor this time is different. And we're going to answer the question as to why on this episode. And what makes this device so important and different in a world where everyone's seemingly failing, right, EJez?

Like, if I think about all of the other companies who have tried to build an AI-first hardware device, they've all been failing miserably. We talk about the metarayband displays, we can't even get our hands on them. The demos are clunky. It doesn't work. We have companies like Humane and Rabby. And we have another one this week called Pickle. And none of them seem to work. So open air is going in a different direction. And I kind of want to talk about why they're making this decision to go one way when everyone else is going the other. I think the toll of wasted money on kind of defunct AI hardware products is now surpassing about $10 billion. It's a mixture of like VC funded,

crap, and then you've got even the big boys like meta releasing sunglasses, which then suck. It's pretty insane to think, Josh, that it's been less than a year since OpenAI announced their kind of like entry into consumer hardware. If for people, folks who don't know, Open AI basically acquired Johnny Ives hardware startup called I.O. for six and a half billion dollars repeatedly, a mixture of cash and equity at the time, which was really, was their largest acquisition to date, I still think. And there was a lot of speculation and rumors around what that device might be that they're building. And we actually put out an episode early last year, I think it was during May actually, where we, you and I speculated about

what that form factor might look like. And I remember we settled on two devices, Josh. It was the glasses where they would have like microphones that could like listen to you, but also kind of like speakers that could blare audio back to you. Obviously, the lenses would act with like a camera that can kind of see what you do. It's kind of like this multimodal type of device. And then the alternative was kind of like this puck-shaped disk that could like sit on your desk or sit in your pocket that might have eyes and visions. We were wrong on both accounts if this leak that we're showing on our screen right now is true. So the tweet goes, additional details about open AIS device and it goes on to to kind of describe this this iPod

shuffle type device which takes the form of a pen device which will have a sort of camera microphone but kind of like ambient presence in your pocket where you can kind of like take out you can write the device will kind of detect the kind of motion of your hand and understand and transcribe those words that are written into your chat GPT interface so there's like this weird link between real-life ambient presence and chat GPD, which you and I are familiar with today. So per that episode last year, we may have been wrong on some things, but I don't think we were wrong on all of them. And I still feel pretty good about the direction that we're heading in. So this says pen-shaped device, integrates AI, lightweight, highly portable, about the size

of an iPod shuffle, can be carried in a pocket worn around the neck. It'll include a microphone and a camera. And it'll be able to convert handwritten notes directly to text and instantly upload them to chat GPT. So if we take... The iPod shuffle, by the way, for anyone who is not a dinosaur that can get an idea of what this look like. So I wanted to include this picture because I think it's important to help understand the form factor. I think when a lot of people think about this, and when we first thought about it, it was more of a rounded shape just because if you're designed these things from first principles, that kind of made sense to put on a desk, right? Like if you're designing

one of these devices, you want it to see, you want it to hear. What materials does that look like, what shape does it have to take, how do the cameras have to interact with the materials? We kind of concluded on this little round puck shape. It appears as if the shape is wrong, but the sensors are still correct. It still has the ability to see. It still has the ability to hear, and it still will likely be this always on thing. It will just come in a form that's maybe a little more wearable as jewelry or fits easier into your pocket. That's what it seems like this is. And I took the iPod shuffle first generation because EJ's, if you notice on that image, it actually has a little string around it to put around your neck and to wear it.

Oh, yeah. Wow. This has actually been done before in the very first generation of the iPod. What happens if you take this very portable device and you just add the new 21st century technology that wasn't available when this thing first came out? You get something that can see, that can hear, that can speak back to you. And it creates this really interesting thing. Now, we have some renderings of what people suspect this will look like. You just, how do you imagine this actually plays out when we get it? Well, it's really interesting. The number one term that's used to describe this new advice since the news came out is pen. And so the code name for this device, whether it is gumdrop. The concepts people have drawn up is literally a pen. And it kind of looks like, Josh, I don't know whether you ever bought one of these junk tech things. But do you remember the pens that had a USBC, like a USB stick within them? They were like really chunky pens and you could write and then you could unscrew the cap and suddenly, wow, hey, You've got like a memory stick here.

This kind of looks like that. And I hope it's not something like that. But I get what they're going for. And here's my take. One, I kind of weirdly like that it's not another screen. I'm tired of having like a laptop and then a cell phone and then you had like the Apple I watch. There's too many screens involved here.

I like that they've gone with a potentially kind of a non-impactful device that I can just kind of put down and not acknowledge, but is still so aware of everything. It's consuming maybe audio that I'm ingesting or it's seeing what I'm seeing. I quite like the hands-off approach. Number two, I like that it's mini. It's funny.

You mentioned that the pen could potentially be on a piece of string like we're showing on the screen here with the iPod shuffle. It's giving me friend.com vibes, Josh. Do you remember that device where it's like a... I do, the little round circle. Yeah, yeah. Exactly.

Well, it's like a pendant that goes around your neck. I'm kind of hoping it's not going to look like that. It's going to look like a weird kind of jade totem crystal type thing. But I appreciate that it's going to be non-impactful. And then my other thought about this is I like that it is ingesting types of data, which I'm not usually focused on. Okay, so what do I mean by that? Humans to date have been the processes of information, right?

So we ingest all of this audio and visual stuff. We process it in our brain, and then we stuff it into a laptop or a mobile phone. It comes in the form of texts. It comes in the form of TikToks that we film or whatever. But it's pre-processed in our human brains. What's different with this device is it'll do the processing for us because we've built this kind of like AI model already. And that's what's super exciting for me. Yeah.

If we look at that pen again, I feel like I can say with almost 100% conviction that it will not actually be, a pen. A pen, like if I put myself in Johnny's shoes, a pen is such a horrible form factor. There's so many different types of pens. There's fountain pens. There's the like different tips for pens. There's different inks for pens. There's this huge preference stack that people have and feel very strongly about because they've been writing with pens their whole life. And trying to cater to those people without creating a huge amount of skews that satisfy the preferences of each one of these people seems very complicated. How do you choose the quality of the ink? How much, how do you replace the ink when it goes out? It just adds too much complexity to a device that

should not be that complex. But I think the form factor is probably some hybrid between the pen and the iPod shuffle, like a pendant type thing. It makes sense. And then we have to talk about like the bet. Like why are they doing this, EJAS? And you started touching on parts of that. But the idea is that currently the phone is it's intentional computing right now. You have to go seek these things out or perhaps they come to you, but they come to you in a noisy way, where you go to your phone to send a text message. You go to do a Google search. You go to check social media or check whatever algorithm is going to like keep you hooked on this device. These AI hardware devices, particularly this one without a screen, it aims for this ambient computing. It's something that's always there. It's always listening.

It's always seeing. And it's not only waiting to help, but it's proactively seeking ways which it can help. And we've seen this previously through a feature that Open AI actually released in the past called Pulse, where if you remember Pulse, what it does overnight is it'll take all the queries that you had from the previous day. It'll do some extra research. It'll kind of figure out what you're most likely to be curious about the following day, and then you wake up in the morning, and it'll have all of the answers and all of these unique facts and things that you didn't know you wanted presented to you first thing. And it's this really valuable resource in the first step that they're taking towards becoming this predictive ambient AI that knows

what you want before you even ask for it. Now, a second thing is happening that is confirming these suspicions, and it's the company shakeup that's actually happening within open AI right now. There are recent reports that happened earlier this week, and it talked about the shakeup that they're doing towards creating a better voice team. Yeah, and I think this is the perfect time for a device like this to come out because AI models aren't just LLMs anymore. They ingest video and produce video. They ingest audio and produce audio to a really good quality app. I think Google's Gemini is probably the best or the leader here, right? They've got V-O-3 on the video side. they have translation AIs and they have one of the best LLMs there is out there

and it all kind of combines and melds into this one product into this one model which can kind of spit out any type of medium for you, right? If you had a device that now ingests all of those different types of data that is around you, you'll probably end up having a richer AI experience. I think that's what Open A.I. is really going for. It's not like this fancy device that kind of like does a crazy magic trick for you. It's kind of like an advanced sensor that can speak to you. and hear and see the things that you guys see.

And actually, on that note, a bit of evidence that broke out this week is that Open AI is working on a new voice model. And this was leaked by the information where basically, if you've ever used chat GPD or if you're, I don't know if you need to be a pro user to get this. I think anyone can actually use it. But there's a smart voice mode where you can basically talk to chat GPD and it is really, really good. In fact, there's like zero hesitation between me stopping talking.

and then someone responding for me. But then if I interrupt, there's this awkward interjection where it kind of like stops and starts. This new voice model is basically going to be like superior as if you're speaking to a natural human in normal human tones, but also it'll be perfected for the device

that they're building and releasing later this year. Yeah. So why is 2026 the year of AI hardware? I guess to summarize, there's these three converging forces. There's like you mentioned earlier, AI becoming multimodal, it has voice, vision, context. And the hardware part suddenly makes sense. Then there's people who are just

tired of screens. Everyone is just kind of sick at scroll. I don't know if they're sick, but they're getting sick by scrolling Instagram, TikTok, Twitter, the whole thing. It's just screens are very invasive and there's very few passive ambient technologies that don't ask for your attention. Instead, they deliver you value. And then three is the distribution plus platform power. I mean, the next battle of who owns the AI interface isn't just the model, it's the entire stack. It's the, is it used on your mobile apps? Is it used? How do you inject yourself into more of these people's lives? And I think the distribution and platform power is something that Open AI has a huge advantage when it comes to. Josh, how much do you think this, if you had to guess, this device would cost?

I suspect it'll be under $200. And my hope is that it'll be under $100. And the reasoning is because Open AI, I mean, by the time they released this, they make a lot of it. might have one and a half billion weekly active users. It's this astronomically large user base. And Open AI, if no one knows this, they haven't exactly produced hardware in the past. They don't have the distribution. They don't have the supply chain that a company like Apple has where they can create tens to hundreds of millions of iPhones every single year. So the way you get around this is to dumb down the device to a point in which you can manufacture it at scale in a cost-effective way without needing to create any sort of

very eccentric frontier technology that is required for production. Like with glasses, it's very cutting edge. It's very difficult because it's a hardware problem. With open-AIS device, I suspect they will make the hardware be basically trivial. I mean, you just need a few sensors. You need a camera. You need a gyroscope. You need a microphone. And the hope is that it will actually be underwhelming to a lot of people. When you see the final form, it's probably going to look pretty dumb. It's not going to be this crazy, beautiful, fancy device. It's just going to be a device that is going to be supercharged by the software stack instead of cutting edge manufacturing techniques on the hardware. Well, also, glasses are superimposing on the average user, right? You're going over

the most sensitive organ of their entire body, which is like their eyes, which they use every day, right? It's very personal, whereas the hardware device that Open Air is working on appears hidden yet present, if that makes sense. The other point I'll make is it definitely makes sense for it to be cheaper. I see this is kind of like how Amazon priced Alexa or Google priced. I've forgotten what their hardware devices actually. But the point is you want to get this out to as many users as you can. And if you want to do that, you probably need to price it really attractively. You're not going to price this like an expensive iPhone, at least in the first wave of getting this device. You want to get it into many people hands as possible because the feedback loop for OpenAI

is they in return get all this amazing data, which they can then make a crazy model for the future. So it makes sense. But Josh, it hasn't prevented people from releasing more smart glasses. Nobody learns, man. I swear. And nobody looks at the previous failures and says, oh, maybe I shouldn't try that. They're just doing the same thing over and over again. And every single time it fails miserably and not to ruin the plot here, but there was a new glasses company that was announced this week called Pickle One, the first sole computer.

They're calling it. Don't ask me what a sole computer is. What's a sole computer, Josh? I have no clue, dude. They're just trying to market this thing. And the marketing video was quite good. I mean, the glasses, if you ask me, they look beautiful. On screen, we're looking at these kind of futuristic silver glasses

with clear lenses. And the demo looked awesome. It was this really cool overlay into the real world. and it showed you these like ride share instructions and navigation, and it's everything that we dream an augmented reality device will be. The unfortunate reality was that this demo wasn't even a demo. It was just kind of this computer graphic imaging to make it look like a demo to sell the vision of something that doesn't exist.

And this came in the form of this kind of expose, if you will. If you want to share the details that came from within this, and why in particular this is still an impossibility. Yeah, so basically this guy called Matthew Dowd, who has been kind of like the most vocal opponent to this Pickle 1 announcement, which got everyone excited, basically pointed out some simple things, which is like, huh, in theory,

if you wanted to create a pair of glasses that look like this, you would need to get the costs of these components down 5 to 10x from what it is currently today. So he kind of basically pushed by, back, he found some inaccuracies on the website versus what the founder was also pitching in his demo video as well and vocally on X. So these kind of discrepancies plus his expertise in hardware basically pointed out that this pickle device is probably not real. And it points towards a team that hasn't really figured out how they're going to build these glasses. And this kind of like reflects a lot of the conversation you and I have had on AI hardware glasses, Josh, which is in concept, in theory, it sounds great.

And Mehta has probably been the only company that's got the closest to putting out something that is actually real. Although we saw that transpire really badly in real time on stage when they demoed the product. And three out of five of their demos completely flopped. And the ones that did work also kind of sucked. So it's not really useful for you. And we don't think it's practical or I don't think it's practical right now to be able to build something like this. Now, has this random dude in South Korea calling his product? Pickle pulled it out the bag, maybe. He's kind of quoted to use a Snapdragon processor,

which is a Qualcomm-based chip. Maybe he's figured something out. From the visual side of things, Josh, I don't know about you, but I don't really like the look of this. It's giving Google Glass version 1, which is funny because Google Glass is supposedly going to be releasing their V2, sometime this quarter as well, excited to see what this is. Now, to kind of like put my other hat on this, Josh, like, listen, I'm not going to buy pickle. I've been burned already with the Fetra Raypans display. But why do you think so many companies are focused on building glasses? Like, it seems kind of weird that you have like, what,

four, five major companies and some startups focusing solely on the visual aspect, on these glasses form factor. We even speculated that that's the most likely probable device. And then you have open air kind of going with this pendant approach. Like, do you think both companies or both approaches are wrong, or one is more right than the other? I actually think both approaches are right. I think the timing is wrong. And I actually, I shared one of the posts that I share because it made sense when you build hardware, there's, you can win on two merits, right? You have either manufacturing innovation or software novelty. And if you don't have one of those two things, your competitors become Apple, Google, Google, meta, open AI, who is trying to build this manufacturing innovation themselves. So Pickle is unfortunately coming with neither of those things. It has has no breakthrough manufacturing techniques, or does it have any software novelty that creates

this amazingly magical experience? So therefore, they're competing directly with meta, directly with Apple, on supply chains that they very much have a monopoly on when it comes to the cutting edge, when it comes to these two nanometer chips that are so small and they're the only things that are possible for glasses. The problem is that they just don't have that. Open AI, on the other hand, is taking the opposite approach. They are taking the non-elegant or perhaps non-complex hardware approach in hopes that they will have unbelievable software novelty. They'll give you this tiny little device. It's not going to look that flashy,

but the idea is that the software behind it will be so impressive that it will stand out amongst the pack. It's very clear to me, and I think everyone, that glasses are a form factor of the future, and they have a very promising future. It's just the technology to manufacture them at scale, at a quality that is acceptable to the average user, is not there, nor will it be there really soon.

It's very difficult technology. There's a lot of complex problems to solve, and it's very expensive to solve them. So I don't feel good about it. It's funny. I think my major takeaway from this discussion with you is the only thing that's certain is the uncertainty of the next few years and what these devices are going to look like. We are just going to be in uncharted territory. And it's going to take a few different versions of these devices to kind of nail it.

Because like, let's be frank. Like AI models aren't at their end form right now, right? We're going from LLMs to these kind of weird voice and now, oh, wait, it's, spitting out video and nano banana from Google can edit images. And we've got to combine and package all of these multimodal types of things into a singular experience. And that's going to be aided by some kind of a device. Now, we can speculate all we want about whether it's going to be glasses or whether it's going to be some kind of puck device or a pendant. We don't know. And I'm just excited that these

companies are investing billions of dollars to try and figure this problem out because I think it's going to be the cell phone moment. I think it's going to be the PC moment. And I cannot wait to live through all three of those, right? It's going to be kind of technology supercharged. But listen, I do have one request, Josh. As long as it doesn't look like this, as long as it doesn't look like some kind of pin device that I slap on my shirt and that I have to hold my hand out. Do you remember the humane pin? To quote Marquez Brownlee, the worst product I've ever reviewed. To me, the form factor here actually isn't the problem. And I think we do have a pretty clear view of what the form

of these things is going to look like. We will have glasses. We will have this like this pendant shape thing. We will have a TV. We will have a desktop device, a countertop device. The reality is it's distributed. It becomes a suite of devices. The wild card, the thing that is most exciting is how these devices actually deliver value

to us. Like if we look at the humane pin, the reason why it was so bad isn't because the idea was bad. It was because the actual execution was so poor. It's that the hardware was kind of clinging. The software didn't actually deliver any real value. It took so much longer than your iPhone to generate the value that you wanted to extract from it. But if it actually delivered on the promises it made, I would have had one on my shirt right now.

It would have been the best device there is. So the problem is it's a combination of both things. And we see it here again with the Rabbit R1. Oh my gosh. I almost forgot about this. Throw back. Yeah, this was a device that was released in 2024 and kind of like projected a lot of the concepts that we are talking about discussing right now, but in a really mundane and kind of like analog way. It flopped. It ended up

not becoming a thing. I think maybe they're releasing version two, but I hope not. I get your point, Josh, I think, I guess the way I think about it is the end form is going to be chips in our brain. And I know that sounds probably insane to a lot of you listening to this, but it makes sense that like the human brain becomes the kind of end goal for what a computer will be, right? Instead of relying on glasses and mechanical hardware, we just use our own eyes or an artificial brain that works with our brain. So kind of like a human robot or human computer hybrid. And so any kind of device that acts as a sensor to input and augment any kind of information that we ingest that isn't just words, it could be audio, it could be video, it could be any other

form factor makes sense. And so I don't know whether it'll be a suite of these things. It probably will be a suite of devices, I guess. Maybe Open AI becomes a sole supplier to all of these, or maybe we just buy a bunch of devices from all different kinds of things. I don't see the latter playing out, because to your earlier point, you kind of want the distribution and the platform if you want to own that kind of cohesive experience. So, you know, maybe this is the first of many devices from open AI. Yeah, it requires like a fundamental rethinking of the word device. Like if I could imagine what the future of this looks like, it actually becomes inverted where the device is the AI. It's the intelligence and the intelligence is this modular thing that manifests itself through a suite of

devices that's always around you. When you, where you're this pendant around your neck, it has the same exact context window as the TV on your screen, which is connected to the small desktop thing on your desk, which is connected to the display in your kitchen that is your like heads up display. It's all one universal thing. And the actual product becomes the hardware. It becomes your preference stack within that AI. So it really understands you deeply. And I think that is the way that we remove ourselves from the screens is by just putting

these intelligence all around us. That's why it's considered ambient. It's because it's always there. It's always around us. And this open AI device is the first iteration of what that could look like, physically manifested as a product. So that's kind of, that's kind of it. That's where we're at. I'm hoping that they're going to release this thing or tease this thing sometime this year. It's moving slow, but progress is being made. We were finally starting to get some leaks, which means it's at least going into tooling.

They're figuring out where they want to manufacture it. Things are happening here. And I'm really excited for them to happen in a big way. Yeah, I think optimistically, we're going to see what this device release might look like, I think in Q3 or Q4 of this year, which would be crazy. They're already engaged in manufacturing contracts with Foxcon in Vietnam. And rumors say that they also might be building a manufacturing plant in the US as well. The reason why it was delayed in the first place, fun fact, was because it was in China. And as you know, America doesn't really get along with China right now, especially in this AI race. So it's important to have everything onshore.

But yeah, that is it. I cannot wait to get my hands on one of these devices, Josh. We have been speculating about this stuff for six months now. And it would be nice to get a device in our hands that we can kind of play around with and maybe even bring onto the show and augment this entire experience for you guys. That is it.

I have two hopes for this year. One of them is that we get to see what this open AI devices. The second one is that we get a good pair of glasses. My God. Google Glass, too? Google Glass, too? Good pair of glasses. Google Glass too?

Google Glass 2. I don't care who makes it. I'm holding up. I don't care who makes it. Make it set I can put it on my face. I want so badly to wear glasses. But every product sucks. So whatever we've got to do, figure it out, deploy those. Those are the two things I will be stoked in the world of AI hardware. If we get glasses and this ambient device from Open AI. But that's the hardware update. That's what's going on. Now you know about as much as we do, about as much as everybody except for Sam and Johnny who are in the lab, hopefully cooking up something remarkable. And we'll just, we'll stay up to date with all these things. As always, and see you guys in the next episode. Don't forget to like and subscribe, turn on the notifications, subscribe to our newsletter, all of those things, please. Lots more information coming to you pretty soon. See you, folks.

I'll see you. See you.