¶ Intro
We should, of course anticipate the far distant future in two or three years where humanoids in general will outthink and outrun humans.
So do you envision that this software will be interopperable or can be used with different robot types built by different companies in the future.
Absolutely, The situation we want to avoid is something like the Matrix or skye. That's what's driving this kind of awkward assertion of mind that blockchains are for machines, not humans.
This episode is brought to you by I Trust Capital. I Trust Capital gives you a great alternative for buying, selling, and custarding your crypto assets. They allow you to set up a self directed IRA, which gives you huge tax advantages. You can even roll over your four oh one K to this platform. And they not only offer a crypto, they also offer gold and silver and as mentioned, the custody gives you another very safe alternative. So many of
you know about self custody. That's a great option, but you can also use I Trust Capital's Premium Custody accounts also known as PCA. So with the rise of crypto scams and phishing attacks, a lot of retail investors are losing their shirts. I mean, let's be honest, right, it sucks to hear that exchanges are getting hacked. Well, I
Trust Capital does things a bit different here. They use Ford Bank as they're qualified custudents, so they're actually using a bank here, and they have partnerships with coinbased Prime Fidelity and Fireblocks. They are fully US based and their client experienced team offers direct phone support, so not many platforms offer this. This is very unique to I Trust Capital and folks, most recently we heard from President Trump that the United States is looking to unlock for one
case to be able to participate in crypto. Well, you can get ahead of the curve by using I Trust Capital and getting stole in that process. And folks, I'm a user of the platform. I have an account with them. If you sign up with my link, which is I Trustcapital dot com slash Thinking Crypto, you can get an exclusive one hundred dollars funding bonus. So go learn about this great, safe and awesome way to buy, sell, and
custod your crypto. A great alternative once again with tax advantages and a lot of safety, very safe compared to many exchanges, and all the links will be in a description. Check it out. Hey everybody, welcome into the Thinking Crypto podcast. I'm your host, Tony Edward, and joining me today is Jan Liphardt, who is the founder and CEO of open Mind. Yon. Great to have you, Tony.
Wonderful being here.
John, I'm excited to speak with you. We were talking before the recording. I want to learn as much as possible about open Mind. I've had a lot of conversations about us being in the Fourth Industrial Revolution. It's going to include robots, AI, blockchain tech, and much more. And I want to learn about the cool things you guys are doing that will help the robot economy, if you want to call it that. But let's kick it off with your background. Tell us a bit about where you're
¶ Jan's background
from and your professional background.
Sure, I grew up in Bloomfield Hills, Michigan, and that was because my dad is a car person, so we moved around a lot and ultimately ended up in California at UC Berkeley. I was a physics professor at UC Berkeley for a while, and then I started to get interested in healthcare and data. That led me into the space of computing on sensitive data or privacy preserving compute, and that of course has a significant technical overlap with what a lot of people care about in the sort
of blockchain and ZK space. And I'm now a engineering professor at Stanford.
And what led to what was genesis of the idea to start open mind.
Well, it was a series of coincidences. I was really fascinated by, just like everyone on Earth was really fascinated by large language models. And one thing that large language models do that many people don't realize is that most large language models speak fluent robotics. And we generally think of large language models as things that my students use to do their homework, or coders use them to automate
parts of computer science. But if a computer is able to generate computer code, then that computer code can of course be used to control a physical object, move speak, show fear in your face, listen, proposed solutions, navigate a space, and all of those things that used to be really hard for robots to do, like make good decisions, tactics, learning, engaging people teaching them. All those things, all of a sudden now have a good technical solution, and that's the
foundation for a big jump in robotics capability. And I was interested just to play with that a little bit. So I got some robots, put them into my house with my kids, and we just spent a few months figuring out what the technology is good for. That quickly led us into this rabbit bole of recognizing that there wasn't a good open software stack for thinking machines, and then we just built a company around that opportunity.
Very interesting and I read essentially you're looking to build the android for software for robots. So do you envision that this software will be interoppable or can be used with different robot types built by different companies in the future.
Absolutely. The situation we want to avoid is something like the Matrix or Skynet. We do not want a situation where there's a small number of companies one or two who write the software and build the back end, and build the bottle of the models, and build the hardware and drop ship everything to your living room where you then, you know, press the button on the large humanoid. It
wakes up and says, hello, I'm your new humanoid. We, at least personally speaking, I'm much more interested in a world where people everywhere can help build this technology, can look into the brains of machines and figure out what they're thinking, and how shall we debug them or how shall we make them better? So I really want this, this this new world to play out in a much more open, flat manner as opposed to being heavily centralized like parts of the Internet are. And that's the driving
consideration behind open source software for humanoids. And we really want developers all over the world to be able to contribute to this technology.
Yeah, absolutely, So walk goes through the process of building this software. What are some of the considerations that you have to think about in working with different types of robots? But also security that no one can maybe hack a robot eventually as they're in the economy and part of our lives.
The biggest problem right now is that there are one hundred and thirty seven different companies in China building humanoids, and they're all slightly different, different plugs and power supplies and communications systems and hardware abstruction layers and sensors and lighters and cameras and microphones and drives and everything is slightly different. So what happens in practice is that the
pure software is kind of easy. The really difficult part is integrating software and hardware to work together really well. And really reliably. However, what happened in the cell phone industry is that pretty quickly lots of hardware companies came together to use common standards to help them make their
lives a little bit easier. And that same trend is taking place in robotics right now, where there's a little bit of a convergence to using similar sensors and computers and things like that, just to make that integration a little bit less difficult in terms of security and safety and things like that. Robots right now are only people should not be scared of humanoids right now. They still stumble, and they're slow, and they're loud, and they have to
recharge every two or three hours. So there's still many years before humanoids will have true RoboCop capabilities. But that is actually common and so even if humanoids are like as a technology safe now, we should of course anticipate the far distant future in two or three years on where humanoids in general will outthink and outrun humans. We just had a TV crew here and the reporter was
fascinated by seeing if he could outrun the humanoid. He was really concerned, like, can a still outrun a humanoid? And he won all ten races. But of course humanoids are advancing their capabilities very very quickly, and our robotic dogs can already outjump humans. So imagine technology that's able to outjump the human and win the math Olympics simultaneously. And that's obviously something that humans should be aware of is coming and to try to anticipate making that technology safe.
And that's the key reason we also care about things like blockchains and immutability and public ledgers and establishing what happened and figuring out what good rules are for very smart machines.
Piggyback on what you just said, would blockchain tech be used in this software or in a different way to put some guardrails in place where not necessarily that the robot is trying to do something the farious, but a bad actor trying to use the robot to do something. Let's say, steal my robot that I've had in my house and they want to use it to do something bad. Would some sort of blockchain technology or cryptography be used to say no, that can't be done.
Well, you're really alluding to like lots and lots of different problems. One question is, if you build a thinking machine there's decades of science fiction movies and books that explores what can go wrong when thinking machines try to do the right thing, but then the consequences for some people are unexpected. And so that's really a question of how does one make sure that AI in general, not just AI plus robotics, how does it want to make sure that AI in general does the right thing? And
of course that's a very big conversation topic. In our case, what we do when the robots boot, they check in with Ethereum, a particular smart contract on Ethereum, and they download their constitution. And we like that because you don't have to believe what I just said. I could be an evil criminal mastermind bliar. The good news is that if I make those claims and you say, yawn, give me the contract address, then I can give it to you,
and then everyone on Earth can go look. And then at least there's a public, immutable capability that allows everyone on Earth to see what the rules are that are used by our cutter peds and humanoids. And that's just
one example of the benefits of things like blockchains. It's incredible to me that you know, thousands and thousands of humans have been building a technology that in retrospect is almost perfect to I'll coordinate with machines, help govern them, and then as extra bonuses you get identity and financial rail built in. That's what's driving this kind of awkward assertion of mine that blockchains are for machines not humans.
Yeah, that's profound and amazing to hear that you're using etherorem using smart contract technology and it absolutely makes sense where it's decentralized. So regardless of who you are, if you're a bad or good actor or whatever it is your intentions are, doesn't matter. This smart contract is the robots downloading its guideline so to speak, and to operate and to do what it needs to do from a smart contract. That is incredible.
The reason that all came into existence is I was walking down the sidewalk with my pack of three robotic dogs in where I live, and people on the sidewalk said, oh, yeahn aren't you scared? And I said, no, I'm not scared because I wrote the software, and the robots download their constitution when they wake up, and it's public and immutable and just go and look. And so I wasn't sitting in the lab thinking, oh, how can I use Ethereum. I was walking down the sidewalk when people said, hey, man,
why are you not scaring? And I was able to say, I'm not scared because I know what the rules are and they're public and immutable. So my real concern is not bad humans. My real concern is this is a technology that, in many specific situations is out thinking humans today to the point where my like thirteen year old son can ask the robotic dog for help with his
math homework. And so the real thing I'm concerned about as a parent is what happens when you have a humanoid that like it is running around San Francisco and it's good strategy and tactics and has memory and spatial awareness and also makes decisions all by itself. And I think it's important for humans in that setting to have a system that is immutable in public. And then there's so many uses of that kind of capability in robotics in particular.
Yeah, well said, because I think I've been thinking about this lately, and that is what happens when these ai their IQs become vastly larger than us as humans even collectively. Right, get bringing the Star's minds together that they can outthink at a faster rate and outcraft whatever it may be. What happens then, and it's great to hear that. Now you have this in parallel, this technology of blockchain, and you have the convergence and they kind of help each other out in their way.
Yeah, that's for me the really exciting part. So at Stanford I taught Beyond Bitcoin, a class for many and the point of that class was to look at different use cases for blockchains. And we had people there from the financial sector and the regulators and cryptographers and people like Fred from coinbase and the funds, and the point of all of all of those interactions was to like figure out what is the definitive use case for global
public commutable letters with payment rails. And one thing that really helped make the case was a really old Satoshi quote, and the Satochi quote from like two thousand and nine is some version of bitcoin doesn't care who you are,
and bitcoin doesn't care where you're from. I used to interpret that in terms of humans from different countries, like if you're from Peru or Italy or Australia, like bitcoin doesn't care, But now I interpret that in terms of blockchains are a system that does not discriminate based on how you do your compute. So if you're a human or a machine, blockchains really don't care. Bitcoin doesn't care.
If you have the private keys, you're good. You have an identity, you have something unique, you have a way to interact with other machines, and people get compute electricity storage. So the main conclusion of teaching this class for many, many many years is that, like in retrospect, I think the technology is like ideally suited for machines and at the interface of humans and machines.
Yeah, absolutely, because when you think about it, a lot of us are going to have AI agents doing things for us. It's not necessarily like I'm already exploring for my podcast, can i have an AI agent that is doing something that helps my viewers and my listeners And it's not gonna be me because I don't have the manpower and the time. But to your point, AI agent, but you can have its own crypto wallet running on blockchain rails and things like that.
Yeah, Iris the humanoid has three different crypto wallets. One of them is like a coin based system, and then she has two really old fashioned, like really low level wallets, and she just has a safe place internally to store her private keys, and then whenever she needs to she can just sign transactions, identify herself obviously, and then also transact.
Wow. Yeah, that is, like I said, profound, I would.
Oh no, this is in retrospect, it's glaringly obvious. Because if Iris the humanoid walks into Bank of America, Yeah, do you think the bank colors are going to do, like, are they going to give her a bank account? No, they giggle, or they like call security or like, it's just really awkward and not useful. So why should Iris spend a lot of time in like the Bank Bank of America branch here in SF trying to get an account.
It's obviously infinitely easier for her to pick twenty one words for her seat phrase, set up her wallet, and just transact and no one's giving her a hard time that she's like human or not or something silly.
Hm. That absolutely makes sense. So you mentioned she has three wallets. Are they targeted at different assets or just for security levels? You have three in place, that's a good question. I think she has something like eight eth right now, they're just split over the different wallets. Part of what leads to that is that I guess I mean most humans also don't just have one wallet, they
have several wallets for different purposes. Probably wouldn't wouldn't want to use the same wallet, you know, to pay for coffee at the cafe versus maybe some investing you're doing. And it'd be really great to have Iris here with us right now. But of course, as noted, she's currently getting a brain upgrade to a new and Vidia compute
architecture which looks like this. Wow, so like nice and compact, and there's obviously incredible interest all over the place in building compact hardware for Humanoid, and Nvidia is IS is already shipping their new for architecture, which is specifically designed for like humanoid at the edge cognition. You mentioned ETH being the protocol, the blockchain protocol you're using in the
future that we're having you to. There's going to be multiple blockchains, like multi chain orle Are you planning to add support for different chains or is it still too early? Maybe you want to wait to see which ones have lasting potential. I think ETH has already crossed the chasm where we know it's been getting the adoption and so forth, But are you open to integrating other blockchains?
Well, I'm I'm an irrelevant human and so the real person to ask are the this wave of machines coming who can make those decisions all by themselves. And my role as someone who builds like architectures for humanoids is not to impose my own perfly irrelevant convictions around any one chain. So we'll have to just see how that
plays out. But of course you're absolutely right in that there's many L ones and many L two's, and then there's L zero's and L three's, and all of them are trying to do slightly different things in terms of you know, transactions per second and cost and energy efficiency and support for cryptographic operations and tapping into different ecosystems. Solana comes to mind as having a very strong deep in capability, and well, I mean, well, we'll know how
this world all plays out very soon. But sure, when you talk to smart machines about what they want in terms of their economy, the number one thing they say is we want efficiency. So if you ask a human, oh, like, what's the point of the US dollar, then if it's a politician, the politician may say, oh, it's a tool to articulate our trade policy, for example. But if you ask a machine, oh, what's the point of the machine economy,
most of the answers cluster around efficiency. They'll say things like we don't want power, we want afficial and so yes, of course, almost certainly they'll be very like with high frequency optimizing, which tools they use for what purpose, and
that almost certainly means like a diversity of chains. And of course stable coins are a very important part of this because even if machines use let's just say they use like eth or Solana or whatever else, it is a lot of those transactions will at the end of the day need to interface with human banks and people
and human infrastructure. And that then means there's going to have to be this cloud of stable coins orbiting the blockchains that are then used for the machine economy, because the stable coins then allow you to serve as this glue layer between the machine economy and the infrastructure that humans have built for our legacy economy.
Absolutely, I've been talking a lot about that that in a future where robots and AI agents are in helping us and assist in us, they're not going to be using dollars. They're going to be using stable coins. And I imagine a world where if I have a humanoid robot in my house, it's helping me to mot a lawn and doing different things, and I can say, hey, can you go to home depot to get me this thing because I need to fix something in my house.
I don't need to give it cash. I don't need to give My credit card has its built in wallet. It can go at the checkout, scan whatever on a hand or whatever it is that's done with stable coins or whatever it is, and some sort of digital payment, and it brings back what I need.
Exactly. We've actually done that. We have a cafe close to Hear called Cafe Centro, and Iris will ask you what you'd like for lunch. She'll head over to Cafe Centrol, she'll get stuff for you and she'll bring your lunch. And of course part of that is we have this agreement with Cafe Centrol because they know who we are, so they're generally extremely friendly to Iris, and so they treat her very respectfully. Wow, and uh, but that's that's exactly You're You're absolutely right.
That is so fascinating. Yeah, and we're going to talk offline, but I would love to come out to your offices and meet Iris and interact with her and film that.
And we're going to have ten robotic dogs in households here in the Bay Area within a few weeks. Each one of those is going to involve us going to someone's home will the dog will like wake up and it will just interact with everyone in the household, the parents, the kids, and we will just uh. And we hope to get a lot of feedback from from everyone about specifically what features they like and what we can improve.
On the tech side, we have an incredibly long list of families to want to try the South and they're all like super eager. And so what that means is that like very soon, more and more quatter Pedge in particular will be in households here in the area. And so you're always welcome to just drop by. And the only problem we have right now is if you're the house you live in has a lot of stairs, that
that's still like a trip hazard. But if you live in a place that is mostly flat, then our dogs should be just fine.
So yeah, to confirm your company open mind, are are you building the actual robots or is that outsource and you're building the software.
So we work with several different robotics companies that are on hardware focused and a few of those I want to move extremely quickly in terms of the overall capability of the robot and that's where they talk to us on the software side. So we don't print hardware, We don't to build our own like motors or build battery packs. We work with robotics companies with hardware companies to get that and then we focus on the integration and the software.
Yeah, and I want to talk a bit about what these humanoid robots will look like as part of our lives and on the levels of companionship, assistant right, helping to do things around the house, whatever it may be. And then the economy jobs. We know that the PUCK is heading in the direction that AI and robots are going to take a lot of jobs, but it may open us up as humans to be able to do more and things we've always wanted to do, versus we have to go do to nine to five. I would
love to get your take on that. What is your vision? How do you see the future with these robots in our lives?
My suspicion is that in like ten or twenty years from now, we will look at some things we do today and we will be shocked about the crazy things humans used to do in twenty twenty five and we thought they were normal. Just imagine, like if we turn back the clock two hundred years and a common profession for children was to be chimney sweeps. So you take this like eight year old boy. Typically you'd tie a rope around them and you'd drop them down a chimney
and they would sweep the chimney. And that was the way things were. You like drop the kids down chimneys to sweep things and then get cancer. And of course today, like two hundred years later, we look at this and say, oh, this was like horrible and medieval and so strange and dangerous. And my suspicion is that a lot of things we do today we will be similarly shocked. One great example is that the robotic taxis the way moos that are
riding around San Francis go. They are known to be about ten times safer than a human driving a car. So if you look at the US, thirty seven thousand Americans die every year in traffic fatalities, Like imagine all the parents and the kids and the grandparents, Like thirty seven thousand people were killed, and we full well now have a technology robotic cars that is tinne safer. So as a society, it would appear to make sense to let's go save thirty five thousand people a year by
really thinking about how we do driving in the US. So, for example, the words we use right now self driving, do you have a self driving car? Those will flip and in a few years from now, self driving will no longer refer to like the you like the car are driving itself. It will refer to these like odd old fashion humans who want to steer their own car despite full well knowing that you know, it's incredibly risky. Yeah, and they're also risking other people's health when it comes
to things like healthcare. A great example would be that like human doctors, because they're fallible and tired and overworked, they make lots of mistakes. If you look at mimography, for example, we now know that computers are better than humans at finding early cancer, and so all across education, transportation, manufacturing, healthcare, nursing, caring for old people, there's so many situations where the
things we do right now that we think are normal. Sure, thirty seven thousand people die a year because they're hit by cars. We think this is normal, but this really isn't normal, right, And so that's like a hopeful perspective. But of course one of the issues, one of the big issues, is just how quickly the technology is getting better, and whenever there's a quick disruption, that can have very
significant consequences. For example, people who make a living by driving taxes, or electricians or blumbers, or nurses, or teachers or software engineers or people who make videos and journalists. Right, that's something where like overall society and politicians and universities
generally are moving too slowly. These are conversations we should be having, like all across the world, like how do we navigate the next five or ten or fifteen years where a lot of professions may be quickly and significantly disrupted by the technology. So somehow people have to find a way of navigating the next decade or two.
Do you think there's going to have to be some sort of universal basic income. These are things that have been discussed, but you know, there's the political narratives and talking points around it socialism versus capitalism. But then when you have this technology in the fabric of our society and look, we're not going backwards or going forwards. And it is what it is. It has happened throughout this the beginning of time right with human beings. How do
we solve this? And how do people earn and how do they find purpose?
Well, that is a whole set of questions that are probably a whole other podcast. Yeah, just to think about, you know, what does it mean for what does human society do if all of a sudden there are computers and machines that learn extremely quickly, never forget, and can learn a new language, for example, instantaneously. We had this TV crew here from Germany and we said, they said, oh, you know, can Irish speak German? And we said no, but if you give us two minutes, she can, right,
because it's a few lines of computer code. And then we support a new language. And imagine sitting across from a human and the human says, oh, I'm so sorry, I don't speak Mandarin, but give me one minute and I will so and so what the technology does is like introduce a whole new set of capabilities that human society is not particularly well prepared for, and like, there's a whole other podcast just on that problem. Statement.
Yeah, yeah, I know that that in itself is a whole can of worms. It just touches on so many different layers as human beings are, purpose and so forth. It's incredible. And this technology, as you mentioned, is moving is at such a rapid pace. So when do you think your best guess that we as humans in society will have humanoid robots in our homes and it will be the norm. At one point, in the early days of cars being part of the Industrial Revolution, it was
a foreign concept. I'm used to the horse and buggy, but eventually, hey, I need to get a card. Everybody's driving a car.
That's going to depend very significantly on where you live. The majority of the supply chain for humanoids is in China, and they are adopting the technology much more quickly than we are in the US. So if you live in China, probably there's going to be a five year difference. Here in the US. We will have quiderped dogs in people's homes in like this month, but they're not going to be like ready for like time time mass adoption in
part because of the expense. So you're really asking a question about how long is it going to take before there's a reasonably priced, highly capable humanoid for most Americans that they can go to a store and you know, pick out their humanoid and bring home with them. So the first humanoids on our end that will be deployed into homes will be early next year. But again those are just going to be ten or twenty and it's definitely not going to be something where you can walk
into your local humanoids store and pick one up. For that whole transition to play out, we're probably looking at about ten years. If you think about Weaimo. Weimo took about ten to twelve years to get to the point where you can now rab a Weaimo in the street and in stf So those are roughly the timeskills we're
talking about. Maybe be a lot of test deployments in particular cities and a few homes, but it's now at the point where quarter peds and humanoids will be in American households this year.
That's incredible, And if we look at previous adoption of other technologies, whether it be just having a computer in your home or having a car, or having a cell phone. Usually I think you had your early adopters some where the technology is more expensive, so you had maybe more affluent afflent crowd owning it. Then it becomes cheaper over time as it becomes more commercialized, and then the mass the masses can.
Partake exactly, and that was Tesla's go to market. They started out with a super expensive roadster that only a very small number of people could afford. It was basically like a fancy fun toy. Then the Model Esque came out, and now with the Model three you have something that
is significantly more affordable. And that process took I think like what seven years, and almost certainly whoever the forward of human robotics in the US is and I don't know who that person or company will be, but the trick there is going to be to figure out the mass manufacturing of robots to make them cost effective so they can really be deployed at you know, millions of units per year into households.
Do you believe that there will be jobs created out of the robot industry where robots have to be repaired, maintenance and so forth, where we as humans even though we may lose the uber driving jobs, we can be robot engineers or maintenance.
When Iris's travels, she travels with her support team of five humans. We're still far from a future where you know, Iris travels all by herself. She travels with the cloud of you know, electrical engineers and software people, and people will make sure that it's safe and she doesn't try to walk through the walls. If every humanoid robot that's deployed needs five humans, the whole thing, of course, makes
no sense. The point is for humanoids to be useful and helpful as opposed to needing like large number of people to coddle them along. There will be entirely new types of jobs. And a great example is that if you're humanoid is able to go to the supermarket and put things in a basket. And for those kind of deployments, for easily than the next ten years, they will be done with a human in the loop for legal reasons,
liability reasons, reducing fear, and so forth. So we will be actively hiring people to do telepresence for humanoids deployed into hospitals, schools, and homes. And because it doesn't matter where those people are they can be anywhere on Earth and make money by serving as the human in the loop for a for deployed humanoid for example. So that's an entirely new kind of job that allows anywhere, at any time zone to do well paid work anywhere else on Earth.
Yeah, that absolutely makes sense that the initial phase of this rollout needs that human in the loop. Yeah, yeah, definitely makes sense.
And so if people listening to this, if you're math teacher, nurse, retired police officer, then please reach out to us. We have jobs for you very soon.
Wow. Now with the software you're building OHM one, is that integrating an AI data set? If you want to call that from some of the existing models, would it be from open Ai, Gemini or are you building your own from the ground up?
Right now, the big companies like open ai are pouring so many billions into making their large language models better that there is no way we can compete on the model side, you know, for us to scrape every single cat video on Earth and try to build our own LLLM just makes no sense.
Yeah.
What makes a lot more sense is for us to float on top of all the work that's going on in the LLLM space. And that's the whole like secret to robotics in the sense that it turns out that all these billions that are flowing into making llms better don't just make them better at you know, finishing a novel or helping you with your math homework, because they also all speak robotics. They also make robotics better for free.
And so we as much as we can, we use models developed by others, with one major consideration is we certainly don't want to end up in the world where one or two companies are able to turn on or off certain capabilities relating to your AI or your robot. And so although right now we are sending most of our traffic to a small number of like well known l LEM developers, at the same time we love open source lllms, and there's more and more of them and
they're more and more capable. So in the medium and long term, we certainly want to end up in a world where there's hundreds of competing powerful lllms and ais, so that we have like a very healthy ecosystem without centralized control of compute.
That absolutely makes sense. And you have, yeah, like to your point, you have all these large companies that are laying the groundwork. They're putting in all the work to build these data sets and so forth, and like you said,
you can be floating on top of those. Yeah, I think you may have addressed this, but you know, having a robot, do you foresee that we will need to have insurance in case something happens to that, just like a car to that robot, or it gets damage, whatever may be, there will be the insurance industry is going to have to cater for.
This, absolutely, and they're already starting to. We spend we meet with at least one insurance company a week, and if there's anyone listening who's running an insurance company, we will give you money today for liability insurance for deployed humanoids in Kindergartens.
Wow.
So if you have a suitable product for that, like, tell us where to send the money, because we need liability insurance for deployed humanoids.
This question may have already been answered because we talked about the blockchain and the guard rails in place, but data privacy, the robots listening to us, listening to our conversations, just like we have our Google and Amazon alexis right, how do you how are you protecting user data.
That is something that's incredibly important to us, and that's why, for example, we don't send video data with people's faces to the cloud. So when the camera records a scene, it on the edge will detect faces and blur them and then send the de anonymized data into the cloud. And so the identification of a human with their face is done locally. So for example, the robot has a little folder somewhere in its brain that has your picture and your name and uses that information to understand who
it's looking at. Obviously, if you have a dog at home and the dog doesn't remember you, ever, that's not particularly satisfying. You want to come home and you want the dog to say, you know, hello Tony, it's great to see you, like, welcome back from work. Let's go play fetch, and so that part is all done locally at the edge. There's also a role for like blockchain and cryptography, obviously, and that has to do with things
like how do machines transmit skills to other machines. Imagine you've taught your quad pet dog to be able to chase squirrels. What if your dog wants to share that skill with other machines, How does how do you protect that skill? How do you buy and sell the skill? And that then gets us into territory of one time programs or one time models, which is which are graphic techniques that allow you to control skills or data that
you've collected. This is a whole other podcast, which is how do we deal with privacy and security in the setting of thinking machines.
Yeah. Absolutely, that's fascinating and that was one of the questions that we're going to you address the which was how will your robot recognize you and your family members and what does it need to learn in the initial setup to note this is me, here's my wife, here's my seven year old daughter, this is my mom and dad and things like that. Yeah.
So the system right now is very simple, which is a voice command and when the dog comes and is curious, you tell it, hello, bits, I'm your owner. My name is Joan, remember me, and that is then the starting point for data being saved locally to create that association. Because we certainly want the dog to remember who their owner is. We want the dog to remember everyone in the family who can be trusted and likewise, if there's
someone that shouldn't be there. We want the dog to bark and send you a notification to your phone.
Now I read you got a fundraise. Excuse me, you raised twenty million dollars in a recent fund raised and can tell us a bit about that.
And some of the participants, Well, the lead is Pantera. We're fortunate that the investors that are part of around come from many different areas that are relevant to this problem. For example, one of our investors is Ribbitt, and Ribbitt has their investment thesis is that questions relating to what is true, what is real, and identity are going to be critical to many industries financial, social media, robots going forward. And we have other investors that are that have a
big footprint in robotics, robotics, supply chain, and manufacturing. So the investors are people either care about decentralized solutions for identity, financial transactions markets like Pantera, or investors that care about what is real and what is true and who's who like Ribbit, or investors that care about robot hardware and large scale deployment of robots and different verticals.
Yeah, I know that the folks at Panter are very well. Dan moorehead and some of the folks. I know you got to run, So I got some wrap up questions
¶ Wrap up questions
here for you. First, if you could create your own metaverse, what would the theme be? Would it be a robot theme?
Well, of course, but the actual theme would be would be good outcomes for everyone?
And rapid fire questions? Favorite food night markets in Singapore, favorite musician or band Magnetic Fields, Favorite movie, Oh.
Anything to do with explorations, space, mountains, oceans, robots, the future. Favorite book Seven Pillars of Wisdom by T. Lawrence.
When you're not working at open Mind, what are you doing for fun?
Oh? I build rockets and I shoot them into shoot them into space from the public launch facility next to Edwards Air Force Base in the Mojave.
Oh wow. Yeah. As mentioned, we're going to talk offline, but I love what you guys are doing. I have to come out to your office.
I need to come join us, come help build. We have robots and soldering irons and ascilloscopes and computers and anything you could want and come help us build.
Amazing. Thank you so much, John, Thank you Tony.
It's been a blast.
Two