My name is Matthew Dowd. I'm the founder of a virtual reality company called Wild (@thatswilddotcom) and have been (almost exclusively) in the VR/AR space for the past 5 years. While I myself have not built AR glasses, I've followed friends' progress on such projects and am knowledgeable on the fundamentals of AR/VR technologies.
I believe Pickle (@pickle) is fraudulently representing their product, the Pickle 1 AR glasses.
This is a big claim. First, I'll go over every technical claim across their launch video and website. Then, I'll reveal what I discovered about Pickle's team and the company's history. Finally, I'll bet Pickle's CEO - Daniel Park - all of their pre-order revenues if they miss their Q2 2026 delivery estimate for US customers. If he delivers in Q2, I'll pay him all the money they've accepted from pre-orders; if he doesn't, he pays me (should he accept).
By the end of this exposé, you'll understand how wildly impossible their claims of current hardware and software capabilities are. You might even agree with me that these dubious propositions could only have been made knowingly by Pickle's team.
THE LAUNCH VIDEO
Their launch video makes numerous claims about Pickle 1's current hardware and software capabilities. These are both said by CEO Daniel Park himself and implied by the included videos and "live demo".
"Binocular full color"
For those who aren't aware, this refers to having a full-color display in both lenses of AR glasses, which grants a 3D effect when looking at visuals. Meta's recent Display Glasses, for example, are only monocular, with one display in the right lens. Meta's viral prototype they showed off last year (code named Orion) are binocular, but have a manufacturing cost of approximately $10k per unit, sub-2-hour battery life, and doesn't do most of its compute onboard - it streams it from phone-sized puck that must be within 12 feet of the glasses.
Pickle's description of the Pickle 1 would fall somewhere in between Meta's Display Glasses and the Orion prototype. But even the monocular Display Glasses are significantly bulkier and more expensive than Pickle 1. Due to low yield rates when manufacturing its waveguide displays (the same display technique Pickle says is in Pickle 1), I estimate the bill of materials (BOM) for Meta Display Glasses at its current production scale could be over $1,000.
Once again, Meta's Display Glasses only have one display, not two like Pickle 1. The jump from one display to two isn't linear, as thermal constraints compound. It also means paying for two advanced waveguide displays!
"One of the brightest, widest fields of view designed for everyday life."
Field of view and brightness have historically been a challenge for AR glasses. Meta's Display Glasses have a stated field of view of ~20 degrees diagonally. On Pickle's website, they take their statement in the launch video to another level: "Widest field of view in the world".
In Pickle's Discord community, CEO Daniel Park added details, stating that Pickle 1's FOV is "~30 degree[s]", though it is unclear how this is measured. In the small text on Pickle's website, they state that 'Widest Field-of-view' claim is based on a comparison with commercially available binocular waveguide standalone AR glasses as of December 2025".
Ok, let's give them the benefit of the doubt that they've squeezed out 10 degrees more FOV than Meta. On to brightness.
Brightness is a challenge in glasses because it heats up the chassis and consumes battery. On their website, they state Pickle has a "12-hour dual battery". They qualify this statement in small text on their website:
"Claims regarding "All-day comfort" and battery life are based on testing that assumes a 12-hour wear period with mixed usage: 4 hours of active AR display rendering, 2 hours of active AR spatial rendering, and standby mode. Actual battery life varies by network environment, features used, screen brightness, ambient temperature, and other factors."
Be this as it may, AR display glasses with standard display brightness still struggle with anything close to a 12-hour battery with similar daily usage levels. Meta's Display Glasses clock around 6 hours of mixed use - and those have a larger frame and only have to power one waveguide display.
Meta's AI glasses made with Ray-Ban have a slightly more comparable frame size to Pickle 1. Those glasses have no display, just camera and audio features for wake word activation - and only survive for 8 hours of typical use before needing to charge.
These bold claims paired with the team's experience in hardware (or lack thereof), their other claims, and their company's suspicious history make me doubt the presence of these displays in the shown glasses at all.
Meta's Orion prototype with it's two displays doesn't rely on an external battery, but offloads the compute to a separate puck, but still have less battery life than Pickle. Products have tradeoffs.
Could some sort of AR head mounted display with those specs conceivably exist? Sure, displays with those specs are probably out there somewhere. But could a startup that's raised under $10m, was founded 19 months ago, has no relevant experience, and is also claiming to have all-day standalone battery life with advanced computing features do it? I have my doubts.
0:51 - Video demonstrates 6DoF tracking, anchoring, CGI
At timestamp 0:51 in Pickle's launch video, they (perhaps unknowingly) suggested a few advanced software capabilities for their product.
It's clear that these visuals were rendered manually rather than captured from the glasses themselves; as this X user pointed out, the renders bear significant resemblance to After Effects filters. I suspect this is why these capabilities are demonstrated: they look good on video, but are hard if you want to actually put them on the glasses!
The first capability they show is 6 degrees of freedom (6DoF) tracking and object anchoring as the glasses move around space. For 6DoF VR headsets, you can both rotate your head on all axes and move forward/backward, left/right, up/down.
In Pickle's video at this timestamp, they show a semitranslucent window appear on the countertop, staying in place as the glasses move towards it which requires 6DoF. The digital window stays anchored in its position.
6DoF means the device needs an IMU - which isn't itself impossible, but adds suspicion to the device's shown chassis size given everything else they're claiming is inside. The software needed to support 6DoF for a device like this also isn't trivial. It would take time to affix an open source SLAM pipeline and even more for them to build their own.
In short: 6DoF with spatial anchoring on a device this small and compute constrained is hard for any company to build, let alone Pickle. I doubt they've done it given there history which we'll get into.
I suspect in the CGI room this wasn't flagged as a challenge to build into the device because the team lacks experience in VR/AR.
"Even when you're walking in direct sunlight, you can clearly see full color AR displays which no one else can see"
Once again, brightness is an industry-wide challenge for AR optics. Maybe Pickle uses transition lenses to artificially boost the readability of their displays? If so, that isn't mentioned anywhere that I could find and they certainly didn't show it off.
If they're suggesting their waveguide displays are bright enough and vibrant enough to see clearly in direct sunlight, that would be yet another innovation they've discovered that every other player in the market (including the Chinese manufacturers where the waveguide displays come from) has missed.
Meta has access to every display manufacturer and their Display Glasses' peak brightness is ~5,000 nits, itself very power heavy. These glasses came out just last fall and there is no sign of significantly brighter waveguides shipping for commercial products since then. It's also worth noting that Meta Display Glasses users have noted poor readability in direct sunlight, suggesting a greater leap would be necessary to meet Pickle's claims.
Put simply, Pickle having much brighter waveguides that also have a 30-degree FOV is extremely unlikely. It would mean that some waveguide manufacturer out there (there aren't many) recently achieved a breakthrough and has chosen Pickle to be the first device it's found in. Oh, and it's going to ship next quarter.
The second point here - "which no one else can see" - is another flag of potential ignorance. The only company that has gotten close to AR displays that don't emit light away from the wearer is Meta with their Display Glasses and, to a lesser extent, their Orion prototype. And they let everyone know, too - it was a big point at both products' announcements because of how big an achievement it was in optics technology.
Meta manufactures the waveguides in the Display Glasses with SCHOTT, Lumus, and OmniVision among other partners. It is possible that the rag-tag team of software engineers at Pickle struck a deal and began production with one or more of these same partners, but I think there's another simpler explanation.
The statement "which no one else can see" sounds innocent to someone who has a limited understanding of waveguides and the history of AR optics. It sounds like a cute product bonus. To someone unfamiliar, it sounds more like an obvious trait of AR glasses - not one that was only solved last Fall.
Bright, high FOV, and private binocular AR displays is not impossible, just unlikely today and almost certainly impossible with the other features and form factor they are claiming!
1:34 - Powered by Snapdragon
Pickle did YC earlier last year, so it's entirely possible they made an arrangement with Snapdragon! Snapdragon is the go-to chipset for AR glasses of all kinds, though Pickle didn't specify which Snapdragon chip they'd be using.
Given the language on their website, though, I doubt they've actually tested any such architecture. They claim "absolutely no latency", though this is such an egregiously vague claim that it could mean anything. AR glasses aren't primarily limited by chips, they're limited by thermals and power draw, as Wu Xu says here.
Fortunately, since I started this exposé, another X user received confirmation that Qualcomm is supplying chips to Pickle. This doesn't validate anything else about Pickle's product, though; as the Qualcomm representative says, the rest of the device is up to Pickle to design and achieve.
"At only 68 grams, it's the lightest personal computer in history"
I'm not going to get into the name choices with this company; plenty of people have expressed their thoughts on the name 'Pickle'. So the choice to call this theoretical device a personal computer is debatable.
The claim of it being 68 grams with all of the features listed so far is ridiculous. I'm not contesting that the glasses they showed in their video are 68 grams, I just don't think the glasses shown have most or any of the features they listed and 68 grams would be impossible for a device that does.
Meta's Display Glasses are 69 grams standard with 1 camera and 1 waveguide display (plus speakers, battery, etc.). This is below the supposed Pickle 1 spec. When I tried them, this still felt heavy and not "all-day".
Snapchat's AR Spectacles - an internal device right now - has dual low-FOV waveguide displays, onboard compute, multiple cameras, and more like the Pickle 1 and come out at 226 grams(!).
Meta's Orion prototype is slightly higher specced than Pickle 1 and weighs 98 grams, but has way lower battery life and the majority of its compute is done on its external puck unlike Pickle 1.
The sad part here is that even if Pickle made the device they claim feature-wise at a higher weight (which would still be a multi-year journey given the state of mobile chip performance, optics technology, etc.) it would still be a great product! It is disappointing they are making such clearly outlandish claims.
A 68 gram AR glasses product is not impossible, just not with the features Pickle claims to have!
1:46 - Real-time binocular occlusion from camera feed, hand tracking (?)
I spotted another slipup in their representations of Pickle 1's software. At 1:46, you can see the software overlaying and anchoring elements on the guitar using 6DoF and spatial anchoring (as discussed above), but then another capability is shown: real-time binocular occlusion!
The user's hand, while playing the guitar, goes over one of the overlaid dots - and the dot disappears. This would indicate either the software is using real-time depth tracking or estimation to occlude virtual objects behind physical objects, or they are tracking the user's hands to occlude virtual overlays behind a user's hand.
So now we've added one or two other potential features to Pickle's lengthy list of capabilities: depth occlusion and potentially hand tracking.
Right now the only headset that natively and by default supports hand occlusion is the Apple Vision Pro, which packs a dedicated chip for camera feed processing and composition for occlusion. Meta Quest and Samsung's Galaxy XR have beta options for similar features. Performance of real-time hand occlusion on all of these headsets is far from perfect - even with more compute than Pickle 1 could possibly have. Even on Vision Pro, a $3,499 device with an M5 and R1 chip, lags behind quick hand movements when occluding virtual objects.
As with previous demonstrations in Pickle's launch video, this is a detail an uninformed person probably couldn't point out as being challenging (or impossible) and was probably included in the video because it looks pretty. There is no way their device runs this, and I doubt they could make a device with the other features and form factor they're selling that could in even 3 years.
2:03 - 2:08 - Shows different overlays that span way beyond stated FOV
From 2:03 to 2:08 in their launch video, they simulate PickleOS on the lens of the glasses. At one point, the rendered image appears right in the center, the next appears in the bottom right of the lens.
For a display with "~30 degrees" FOV, range of display real estate isn't possible. While other AR companies like Snapchat and Meta also have a history of misrepresenting the FOV of their displays when the real image gets cut off at the edge of the display, this doesn't help Pickle's case.
2:24 - “Live Demo”
At 2:24 in Pickle's launch video, CEO Daniel Park does a "live demo" (right…).
During this "live demo", the image quality coming from the glasses is amazingly crisp! Certainly 720p, if not higher. It's also a landscape sensor orientation, which would break the mold from Meta Ray-Bans' portrait sensor.
Then I did some digging and things didn't add up. In Pickle's Discord server, CEO Daniel Park, when asked about Pickle's camera, said this:
"Our core use case is ambient capture, not creator-grade recording. Unlike Meta, we're not targeting high-resolution capture…"
That's what I thought, too: it seemed like Pickle 1 would attempt some sort of always-on camera system for passive contribution to its 'memory'. This was reinforced by part of their website:
"Ideas, contents, music, products and more. It quietly observes your context, then proactively surfaces the perfect suggestion."
I suspect the so-called "live demo" wasn't so live, and wasn't recorded on Pickle 1's cameras that Daniel says are not made for creator-grade recording. It tosses the validity of the entire "live demo" into question - to say the least.
3:35 - Real-time animation of companion avatars
Finally, at 3:35 in Pickle's launch video, they show off real-time animated avatars for Pickle 1's built-in companion. Especially at 3:35, the lip syncing is spot-on.
For such a small device that would already be powering the displays, rendering the companion avatar, waiting to receive an LLM response, analyzing camera inputs, etc. I would be surprised if this is representative of the real-time avatar they can provide at the speed that was shown, assuming any of this product exists at all.
OTHER OBSERVATIONS
There were a few other questionable claims from Pickle's team and observations from their website.
First, by this point you probably realize how outlandish the $1,300 list price is for Pickle 1 given how many unique innovations it would take for the product to actually come into existence.
Pickle's website also confirms that Pickle 1 uses waveguide displays. As I went through above, the specs of these specced displays are almost impossible to be in existence - let alone available to Pickle or possible for their form factor - by Q2 2026, so I doubt they exist given the tradeoffs with waveguide.
The other questionable part of Pickle 1 using waveguides is how curved the glasses' lenses are. They look different from any waveguide glasses today - curved, almost like runners' glasses.
Waveguides bounce light in between an internal substrate to reproduce an image on the lenses. Curving the lens means a custom waveguide architecture to account for the curve; that means that Pickle is claiming not only highest FOV, brightest optics, but also a custom architecture for their curved lenses.
I only found one curved waveguide system from Tooz, a joint venture between Zeiss and Deutsche Telekom. They claim it is the only curved waveguide on the market and have significantly thicker lenses to account for the less-than-Pickle curvature.
The Tooz curved waveguide glasses only have an 11-degree field of view and have other tradeoffs, like less flexibility for various IPDs and more image distortion. Accommodating a smaller range of IPDs (the distance between one's eyes) runs up against Pickle's starting point of making glasses that look good on everyone.
Waveguides also require contact with the glasses’ lenses in order to project the display onto the glass, roughly proportional to the field of view of the display. If the field of view is bigger, the waveguide projector needs more space in contact with the edge of the lens to project the bigger image.
Looking closely, there doesn’t appear to be much contact between the outer edges of the lens and where the waveguides presumably are (at the edges of the front of the frame). Given the cameras are also there, there doesn’t seem to be much space.
Other waveguide glasses completely encircle the lenses for sufficient space for the projector. I’m not an expert in AR glasses construction, but that seems tight to project a 30 degree field of view with such little space and would be yet another engineering feat for Team Pickle.
A few more things seem to be missing.
They mention that Pickle 1 will charge on either a stand or a case in their FAQs:
"Your order comes with a dedicated Charging Case and Charging Stand. It offers an all-day battery life of up to 12 hours with mixed use, including AR computing features."
Meta Ray-Ban glasses have a contact point under the nose bridge that connects to the case's battery. The Even Realities G1 have similar exposed charging points at their hinges. But in the photos of Pickle 1, there is no sign of any spot to charge the device.
Likewise, any trace of speaker grills, microphones, or an IR sensor for detecting when a user puts the glasses on are nowhere to be found. Speakers and microphones especially are essential for creating music with Pickle 1's AI companion and giving voice commands to the device as they demonstrate.
The speakers and microphone are clearly visible on Meta Ray-Bans. The IR or capacitive sensor for detecting use is under plastic to ensure reliability. There is no plastic cutout for any such sensor on the images of Pickle 1.
When I was exploring their Discord, a Pickle co-founder said Pickle 1 "is not going to CES this time." and Daniel Park said "we need participants' consent first…" before releasing their presumably more transparent in-house demo video. It reminds me of Nucleus Genomics using fake, shapeshifting photos of clients on their website which brought them much criticism.
At this point, you might already be convinced that Pickle 1 is just smoke and mirrors, but the story only gets better.
This is the Pickle we know today, but there are more.
So. Many. Pickles.
1 PICKLE, MANY FLAVORS
11 Months Ago:
Pickle is revealed as a YC startup for AI body doubles on Zoom calls.
On their YC Launches page, the Pickle co-founders describe Pickle as "Pickle replaces your Zoom camera with a virtual body double that always looks great—even if you're calling in from bed."
It's easy to verify this. In my research, I found CEO Daniel Park's previous posts about Pickle's AI Zoom avatars and a clip of their YC demo day presentation for the same thing.
9 Months Ago:
CEO Daniel Park celebrates raising $200k from an investor in Pickle, still at this point an AI Zoom avatar company.
This means that at most, Pickle had pivoted from this to their revolutionary AR glasses just 9 months ago. Right?
7 Months Ago:
Pickle Co-Founder SJ Lee joins the company.
5 Months Ago:
Pickle pivoted! But not to the revolutionary Pickle 1 they're pre-selling to people today.
This time, Pickle's rag-tag team is at it again with a Cluely knockoff called Glass. As CEO Daniel Park says on LinkedIn:
"Meet Glass by Pickle. Fast, lightweight, and fully open-source, Glass runs invisibly on your desktop, structures context in real time, and proactively surfaces meeting actions and summaries—all 100% free."
But there's more: Glass allegedly wasn't by Pickle at all. Soham Bharambe, a backend developer who built an open source version of Cluely called Cheating Daddy, shows how Glass from Pickle was virtually entirely stolen from Cheating Daddy, thereby breaking the license he put on it.
Soham posted a thread showing how Pickle stole the project and explains how it broke the license here, I recommend you check it out:
https://x.com/soham_btw/status/1940952786491027886?s=46
Ok, so surely now Pickle had a come-to-Jesus moment and gave themselves enough time to build the breakthrough, definitely real Pickle 1, right?
3 Months Ago:
Surprise! Another pivot! This time, Team Pickle seems to have drawn inspiration from Whispr Flow for 'Whisper by Pickle'.
This one seems to at least be on track with their PickleOS software offering with bubble memories, bubble clouds, etc. but still no trace of glasses.
Also - though I couldn't verify this - at some point at the end of last year, they may have attempted something called Pickle Cube. In the image someone sent me, it's described as "Portable device that captures audio and call AI whenever you need."
The image of Pickle Cube I received seems to match their branding, and uses a logo found in Pickle Co-Founder SJ Lee's LinkedIn header, though I couldn't verify it beyond this.
2 Months Ago:
This is the first time I could find Pickle engaging in glasses, when CEO Daniel Park says they held their first in-person demo.
I have not seen anyone other than Daniel confirm that the glasses worked at this demo.
At least one X user, Will Yeh, said the "units" at the demo did not have displays in them and did not do anything during the demo.
Today:
To recap, Pickle - the ex-AI Zoom avatar, ex-Cluely knockoff, ex-Whispr Flow knockoff, ex-Pickle Cube (?) company - is now not just marketing, but is selling AR glasses they say are better in every way than any other AR glasses product on the market and contains better components than any available components in the world.
They are claiming that this product has all the best (and better) components in one package, at a weight and battery life that is impossible by all standards.
They seem to have started on this product 5 months ago at best.
They are selling this to real people for $200 of their hard-earned dollars as a deposit, then more when/if they deliver.
They are claiming they will ship in Q2 2026.
DREAM TEAM
Daniel Park (CEO)
Ho Jin Yu (CTO & Co-founder)
SJ Lee (Co-founder)
Sanio Jung (Co-founder)
Lucas Yoo (Product & Design)
Tars Yang (Founding Engineer)
Wongyun Lee (Head of Pickle Studio)
Krithi Nalla (Design)
H. Jun Huh (Founding Engineer)
I won't overstay my welcome by going into each of the team members here, but know that one of the co-founders joined just 7 months ago and none of them have discernible hardware experience (let alone VR/AR experience) aside from Lucas Yoo being an undergrad industrial engineering research assistant.
I know people who are actually building AR glasses - real ones. They draw on legacy talent that has driven the evolution of AR optics since the beginning. Knowing their teams and their rates of progress, I know that Pickle's product proposition - one that they are selling to people right now - is impossible on the timeline they claim and probably is impossible on even a 2-4 year timeline given the industry acceleration that would be required.
We're talking tens of thousands of people from research-level optics and battery scientists, to supply chain managers, to factory production lines. Fake it till you make it, if that's what they're trying, doesn't work if it relies on tens of thousands of people to believe your bullshit.
My opinion is that if the founders of Pickle were being honest, they know they can’t deliver that.
There are clearly some areas where they just aren’t knowledgeable enough in AR to know how ludicrous their product goals sound, but given the history of this company I suspect they know that multiple major parts of their product and timeline are false even though they’re accepting orders.
That, in my opinion, constitutes fraud.
Thank you to everyone who offered up this information and helped me piece together the many pasts of Pickle, including @mil0theminer!
THE OFFER
I am hereby placing a bet:
Pickle says that Pickle 1 (as described and shown on their website and launch video) is a completed product. They are currently accepting pre-orders for it. They state it will ship to US customers in Q2 2026.
If Pickle ships their product as described and shown by the end of Q2 2026, I will personally pay Pickle CEO Daniel Park the full amount of revenue they made through pre-orders.
If Pickle does not ship their product as described and shown by the end of Q2 2026, Daniel pays me the full amount of revenue they made through pre-orders.
If Daniel truly knows that this product exists and truly believes the timeline will be met (which should be achievable given the completeness of the product), this should be an easy opportunity to make some money.
Sign when ready!
-Matthew Dowd