Peloton for Gamers: Virtuix Goes Public ft. Jan Goetgeluk, CEO of Virtuix

The AI XR Podcast.

The episode begins with a deep dive into Snap's new $2,200 XR headset and its potential impact on enterprise applications. The conversation the

Key takeaways

• Snap's new headset is viewed as an enterprise-first device due to its price point and AI capabilities via Lens Studio.
• Jan Goetgeluk discusses Virtuix's expansion into the defense sector, providing mission planning and rehearsal through 'virtual terrain walks'.

Main topics

• Snap XR Hardware
• AI Safety and Regulation

Notable quotes

The pioneers are the and the ones with arrows in their back.

Conclusion

The episode illustrates how specialized hardware like VR treadmills ca

Transcript preview

Speaker 3 (0:00) Our This Week in XR podcast is sponsored by our friends at Zapper, the world's leading augmented reality platform and creative studio. With over 11 years of experience working with the world's biggest brands through Zapper Creative Studio. Zapper also has an award-winning web AR platform, ZapWorks, that lets you create your own mobile AR magic. Finally, check out their Zap Box, the most affordable mixed reality headset on the planet. Start creating AR over at zap.works or talk to them about your next AR project at zapper.com. Good morning, everybody. I'm Charlie Fink with Roni Abovitz for the AIXR podcast. It's September 18th, 2026. Good morning. Good morning, Roni. Good morning, Charlie. Speaker 2 (0:45) And good Speaker 3 (0:46) morning, Speaker 2 (0:47) Ted, if you Speaker 3 (0:47) hear us. Ted is trying to log on with mobile without success. So hopefully Ted will get in another way. But let's start with the news, the big launch of the Snap specs yesterday. I guess people who pre-ordered the $2,200 headset will be getting them in the next couple of weeks. And certainly the people at the Snap event got a good 40 minutes each with the device, at least those of us with the press. And I have to say I came away. pretty impressed. I had low expectations. Evan gave a speech at AWE, a keynote, in which he described a device that was not like the developer device that they actually were demoing there. And I was like scratching my head, but now I see what he was talking about. And this certainly is a very capable first cut at a device which believe it or not, may have more applications and more appeal to enterprises than consumers, particularly at this price. But it has AI on board. It will allow you to program experiences on the device using Lens Studio on the fly, which was a big, big pain point with previous iterations of Enterprise XR hardware. It required someone with knowledge of game engines. It was difficult to upgrade. It was expensive. So it added another reason not to do it. But I think this kind of knocks that away and could enable a lot of Enterprise applications we only speculated about three or four years ago. I'm optimistic, but they're going to have to hang in a long time to have a shot at consumer because this is not for consumers. This is for first adopters, developers, lens creators, and enterprises. Speaker 2 (2:40) Charlie, I understand you were there. How would you describe the experience? Did you get to try one there? Yeah, they Speaker 3 (2:47) gave us about 40 minutes each, 30 minutes each with the device. I hate passing judgment on a device after 30 minutes because... If I have it next to me all day, I'll tell you how I use it, right? Sitting in a demo, they're just showing you the greatest hits. So the greatest hits are great. And the people working on the lenses are absolutely among the best and most experienced. Lots of people there from Apple, friends who are there from North, remember them? They were acquired years ago by, I believe, Google. And so... You know, obviously time has gone by and the founders and others have moved on, in this case, to Snap. So good on them. I mean, they've made a lot of acquisitions around this technology. They've certainly iterated a lot and waited to release the see-through display. They're the only one in the field since the HoloLens and the Magic Leap are no longer available except to hackers. So, you know, I think they've got the field to themselves if see-through is... got a chance, this is the chance. You know, obviously there'll be a succession of other devices, but not anytime soon, as far as I can tell. Speaker 2 (4:01) I mean, if nothing else, they keep a bit of a continuity going. And they probably are a reference design that I bet you Apple and Meta, everyone else has buying and is tearing apart. So if you're at Samsung, if you're at Meta, you're at Apple, other companies, you've tore it apart. And I bet you teams are making predictions about all the internal components and what they could be replaced with over the next five years to be smaller and more consumer grade. Speaker 2 (4:30) Inadvertently, I mean, this happened to us at Magic Leap, you became a gold standard for what someone can do next. When I look at that device, there's no doubt they tore down the ML1 and ML2. Because you could see echoes of what you do. And I think they're going to be surprised when they see echoes of their work, which are echoes of other works, echoing in an Apple product or a meta product in a theater coming to you soon. So the really interesting thing is, are the big players simply waiting for someone to get close enough, tear it down, lay out the recipe for scale, make it a little bit smaller? Just like Apple did with the folding phone, they let Samsung iterate on a lot of stuff. They probably tore it down, figured out the one thing they needed to solve, which was the crease, spent a bunch of years on it, and then boom, out they come out. That seems to be the big guy approach. Like if you think about AIXR. It must be that Google is watching Anthropic and OpenAI tear themselves apart. What do you do? What do you not do? And then at some point, they're just going to drop something that is better and does not kill everyone. You can imagine like let someone else get shot up up front. And if you're big enough, you just can lay back and lay back and lay back and tear apart and copy, tear apart and copy, and then jump forward. Like that lay back and wait, this is the sad part because you need Evan and Snap to go forward. But I think the big guys are now learning, like, let's just eat the bodies of those pioneers and step on them. I don't know. That's a little bit of a diatribe, but it seems that's the modus operandi. Speaker 3 (6:02) They say the pioneers are the ones with arrows in their back. I'd say it's even worse. I can't tell which one Snap is. But let's get to the AI news because that was pretty compelling, right? Saturday. Dario Amodei of Anthropic dropped an essay, We Must Pace the Frontier, arguing that frontier labs should slow down their rollout of new models and accept independent safety valuations. And he was immediately seconded by Sam Altman and Elon Musk. So once again, we have sort of a Mexican standoff where everyone is saying, I'll stop when he stops. as excuses for exposing us to the dangers of increasingly powerful AI, perhaps AGI. Now, as the debate wore on into the week, of course, the administration is opposed, the government is opposed to any kind of regulation. And there were a number of relatively sane voices out there saying, why do they need the government to stop the release of dangerous systems? Isn't that in their interest? And then came another cynical view. I got this from Ted, actually, who sent me a link. If AI eventually helps create a bioweapon or empties bank accounts or penetrates a financial network, do they have product liability? Is that what this is about? They want the government to stand in between them and product liability? What's your take on what's going on here? Speaker 2 (7:37) So, Charlie, there's like eight million essays you just threw out, but let me start with the product liability one. Having taken a regulated high-tech and competing intelligence company public at MAKO and hopefully also with Synthby, you go through this quiet period with your S1, which is what Anthropic and OpenAI are supposed to be doing, especially because rumors are Anthropic is going to go public soon. You don't talk to the press. You just shut up. Everybody shuts up. So I don't understand what's happening. The other part is they look at all your risks and you've got millions of lawyers. The last thing you would ever do is support a rogue employee who quits who says huge product risk liability things about you. So the most damning thing that happened is not that the employee left and said there's a greater than 10 % chance what is happening at Anthropic kills us all. It's one of their safety alignment guys goes, yes, I agree. And I agree it's greater than 10 % wholeheartedly and enthusiastically and cheerily. You cannot take that back. So you now have to imagine that the underwriters of this IPO and all the funds are going, what is the product liability and lawsuit risk? Because every attorney general and every law firm is going to look for anyone who ever gets hurt, now or in the future, including as a private company, by using Anthropic to do anything. And that's a real problem. Because it's like if a big medtech company said, you know, a very large percentage of you will die when you take our Tylenol or whatever. You know, like I don't want to name any particular. But medtech companies go out of their way to prove that there's such a tiny fraction of a chance. They do this because they don't want to hurt people. They don't want to be sued. And there's government regs. So you have to ask the question. You don't need the government to regulate you to self-regulate first. You just don't build things that are that dangerous and you don't need the government to tell you that. Why are they just not doing it now? Why are they just not making it safe? So that's the bullshit meter. Why can't you just say, well, stop making this dangerous thing? It doesn't mean you're stopping development. Just don't development the parts that are hazardous. I'll give you an example. If I'm one of the major commercial airliners. I can't imagine a CEO saying, I can't build an airplane without it exploding every day. Then you suck and you should get out of the airplane business because the people that make airplanes make good airplanes. They have a very, very low – accidents once in a while, but it's like once in like 100 million kind of thing. Sure, they're safe for Speaker 3 (10:17) the Speaker 2 (10:17) cars. They're extremely safe. And if the CEO said 10 percent of our planes will blow up all the time, you're bad engineers and you're the wrong leadership. It doesn't mean good airplanes can't be made that don't have those risks. It means whatever you are doing, not the field. I think just so our listeners are unconfused, you absolutely can build powerful, safe competing intelligence that does not have to have these risks. It is total nonsense that they're saying it's not possible. That's like head of airlines saying, I can't build a plane that doesn't blow up one out of 10 times. That is also BS because they have. And a carmaker saying every time you get in a road, Charlie, would you buy a car from any of the U.S. or Japanese automakers said like every time you drive it, one out of 10 of you just blows up. You'd absolutely not buy it because some Japanese company would be like, no, our cars are built with reliability and Six Sigma. So these people are just bad engineering if they cannot solve this problem. It doesn't mean others can't. It means they are not, either deliberately or they don't know how. And if you can't build a safe car or a safe airplane, you don't build it anymore. Speaker 3 (11:27) Right. And Speaker 2 (11:27) the government's job is to actually stop you from building it. And again, here's the but China argument. People will build these things safely and powerfully. And powerful and unsafe is stupid. It hurts us more than it like, you know, helps us defeat, you know, our overseas foes or create equilibrium. So there's so much nonsense going on in the news about like, we can't help ourselves but build an unsafe thing. That is not true in other industries. I'm sorry. Other industries have figured out ways to build safe things that are powerful. Sorry for that diatribe. Speaker 3 (12:00) No, and save things that aren't that expensive. I think the argument now is that AI is getting so expensive and the compute is so expensive. You're better off just having a, you know, 4080, 5080 NVIDIA card. I know they're expensive. but it allows you to run AI locally. So you keep everything inside your walled garden. You don't have to worry about security issues. They're open source. You just download them. There's no cost to them. And given what people are racking up in API charges for the leading models, I think this is going to get more and more attractive. And again, that was my thesis behind why NVIDIA acquired Hugging Face, because that's where you would get those models. So Speaker 2 (12:45) there's an interesting, weird, crazy, mythical dimension to some of the West Coast AGI companies are trying to build like some godlike entity. But there's a whole slew of articles about what's happening in China and other countries, which is you people are crazy. We're just building practical things. Yeah. And that is actually a winning strategy. I don't think they're actually focused on building an AGI. They're focused on. practical computing things that make things faster and more efficient and do not create like untoward hazards. And I think the folks who've gone all in on a godlike AGI are realizing we made the wrong bet because so many other bets are in play that are practical and useful that don't bring any of those risks, including what's potentially happening in China where what you're hearing And the news is that they don't want to build AGI. They don't think it's a good goal. They think it's a stupid goal. They're trying to just build powerful, focused, practical systems. Speaker 3 (13:45) Yeah, I wish we had that focus as well. I think it's great for certain dumb things