¶ Intro
Right now, there was Bitcoin, which is slow as balls, then if the Ethereum, which was still sort of slow but offered like the ability to become programmable, and then sort of moved on for there we got we got l twos to go solve the etherium like transaction speed problem. And then we have all kinds of novel novel systems which are less and less decentralized, but more and more practical.
So this episode is brought to you by Treasure. Treasure makes beautifully crafted hardware wallets that make it easy for you to safely store your cryptocurrencies. I'm a big believer in self custody. I've been using hardware wallets for years and Treasure was the pioneer of hardware wallets and they
started this journey in twenty thirteen. Their hardware wallets are open source and they support a variety of tokens from Bitcoin to the top all coins and even new coins such as black Rocks, Biddle and I'm personally a user of the Treasure say. It's a beautiful device and easy to use, and Treasure offers a handholding service where they can help you on board and set up your device so you can speak directly to a Treasure representative who
can hold your hand through the process. And if you use the code thinking Crypto all one word, you can get five percent off your purchase of a Treasure device. So to learn more about Treasure and all their great devices and services, visit the link in the description. Hey, folks, welcome into the Thinking Crypto podcast. I'm your host, Tony Edward, and joining me today is Bailly Wang, who's the co founder and CEO of Prisma X. Billy, great to have you.
It's pleasure to pleasure to be here.
Yeah, Bailly, I'm really excited to speak with you and learn about what you and the folks at Prisma X are doing with robotics and AI, and of course there's the convergence with blockchain and crypto. But let's kick it
¶ Bayley's Background
off with your background. Tell us about yourself, your time at MIT, and your professional background as well.
Yeah, yeah, so, i I've been a founder on my life. I was a robotics research at MIT for a long time. I worked on a robot called the MIT Mini Cheetah, and the Mini Cheetah was very influential in the industry because it got the cost of the hardware down about one hundred x and really enabled people to do things like focus on applications and use cases and software. But when push came to shove, there wasn't really enough automation to make the robots do anything useful at the time,
so that project was retired. The researchers left, went and worked, went and worked at other companies. So ultimately, that's what we want to solve. We want to at prism X. We want to give these machines the intelligence they need to be able to interact with the real world to be able to complete real tasks, and we think that crypto and blockchain are a great way to coordinate the
tasks that need to happen to make this takeoff. I've also been involved in blockchain for quite some time now, since I mined my dollar bitcoins back in the early twenty tens. I was one of those guys who went to spent ten bitcoins on a pizza and felt really happy about it because it was it was free money, right. We used that Mitax electricity the mind bitcoins and used the bitcoins to buy lunch. Felt great when we were poor undergrads. Now that those pizzas are worth a million dollars.
It's maybe a little a little less great, but that's that's that's that's why I've enjoyed crypto. Like through it all the narratives have shifted, the industry has changed a lot. We've seen everything from selling plots of land in the metaverse to decentralized lending systems to and everything in between. And the one thing that's remained constant is that crypto has the ability to captivate and engage a million people and something completely new, very very very quickly, like literally
literally overnight, over over over a week. You could have a million people be excited about what you're doing. And I think that's really important for emerging frontier technology like robotics, where we do need new ways of working with things. We do need new types of jobs, new types of interactions, and crypto is a way to get these people engaged and to coordinate what they're doing.
Billy, that's amazing that you were mining bitcoins so early at a dollar, that's incredible. Were you part of the
¶ MIT Crypto & Gary Gensler
MIT Crypto Club, I know they had formed that over the years.
Yeah, yeah, I was part. I was part of the original community. I saw the first rise of bitcoin to thirteen hundred dollars a coin hundred thousand dollars. Bitcoins were a pipe tream back then. We laughed about it, but no one thought it could happen. We were part of this thing called the Experiment where we gave a third
of a bitcoin to every MIT undergrad. When this price is slumped, we wanted to inject some life back into the ecosystem and see whether we could get any applications coming out of it.
Wow, this is going to sound like a weird question, and there's a lot of people who hate this guy. But Gary Ganser, former SEC chairman, he was teaching at MIT, I believe around twenty eighteen. Were you there? Did you attend any of his courses by any chance?
I was not there. I was in eighteen. I was still in Cambridge, but I was more on the robotic side of that point. Point.
Got it okay? I had to ask because that's a big controversial time, you know, against her in the last under Biden. You know, they had attacked crypto and at one point he was teaching about it at MIT. So it's just a weird change that took place. But I guess it was all politically driven.
Politics is a weird thing.
¶ 4th industrial evolution
So Billy, you know, I really want to get your perspective as an entrepreneur and a builder on kind of
the future economy. Right, you have all these technologies converging, AI, robotics, E V space exploration is booming, you have crypto and blockchain, and it seems like this is an industrial revolution that we're part of the fourth and these technologies are converting and they kind of have a symbiotic relationship where you know, as you mentioned, like I don't foresee robots transacting in cash, but maybe with stable coins in there, and they're running
on blockchain networks to power the data that they need.
Yeah, yeah, I think you put it right. It's a new industrial revolution. And with these revolutions, they're not bad. They don't displace people, but they make people change what they do and how they work, and I think that's important for robotics. One of the core beliefs we have at Chrisma X is that in the future economy, people will work new roles alongside of robots rather than being displaced by the robots. There's clearly things humans are good at, Like we have a lot of experiences we can be
responsible for own actions. We are really good at this sort of external environmental perception and decision making that's really hard for computers to do. But ultimately the robots are also great because it means that these people no longer have to be in dangerous positions lifting heavy objects. They can do what they're actually good at, which is sit in the control making controlling the robots and making sure
the robots are behaving properly. And these machines, which are you know, they're good at outputting high amounts of force and working twenty four to seven. They're the ones who are out in the field actually doing the physical labor, and we see that as a really good relationship, like it's it's it's very positive, right, Like no one actually says I enjoy cleaning sewers or hammering nails into two
by fours. But if you can make the same amount of money or even more money controlling five robots that are doing the same thing, that's a that's a huge wen for everyone.
And to your yeah, continue, I didn't mean to interest you.
Yeah, And to you to your point, I think that's where that that's similar for things like blockchain, where we are sort of at the cusp of a transition in financial systems where because information is more freely accessible, being able to have more decentralized financial systems is now a possibility and can be done safely and safely and reliably right like there were like there were problems like one hundred years ago where like you, you just didn't have
access to enough information to be able to make decisions intelligently. But now with the blockchain, with everything be one hundred percent transparent, with the Internet, with data analytics tools, it's become a lot better. And I think that the machine transactions to are a really interesting aspect. There's clearly a space for machines to be able to rent time on other machines to complete more fleet oriented tasks, and that
can only happen through blockchain. I don't think the traditional financial system is going to allow that kind of computerized transaction into the system. It's just their risk profile on how they make money is not aligned with what the system needs right now, which is speed and transparency and all that kind of stuff.
¶ Robots and Blockchain
And also is there this safety component that Look, you have these robots, they're going to be powered by a network where data is coming in and look, we've seen the dystopian movies, right with the dystopian scenario of these robots being taken control of. But blockchain and encrypted network using cryptography immutable and all these things, could that help police the network, so to speA, because you're building on a better infrastructure versus Web two infrastructure.
I think it's more that the actual humans that control the blockchain are much more diffused than with a centralized Web two infrastructure, so you can't have a single bad actor make a bad decision. Right. The whole point of blockchain is that it's unless fifty one percent of the entire community believes that an action should be done, the
chain should reject that state. And I think that's really important because it means rather than you could imagine a dystopian future, like the typical dystopian future is like the CEO of some evil ABC Mega corporation decides to go bad and commands all of his lackeys to go bad, and there's like a wash and effect down the line.
But it's much harder to attack with a decentralized system, where even for like a modern stake based validation system, you still have thousands of validators that all have to form a consensus and convincing every fifty one percent of those to follow your plans is very challenging, right that that's the whole underpinning behind modern networks.
Yeah, and you know, I personally don't view the world
¶ Robots impact on our lives
as this dystopian thing terminator I robot. I viewed more as the Jetsons. Like you said, there's going to be bad actors that has existed throughout time, right and will
continue because humans will be humans. But if my grandma could be helped by a robot where it's monitoring, doing the heavy lifting, if dangerous jobs are done by robots, and it can even free me up in my daily life to do things I want to do versus Okay, I don't need to go mow the grass, the robot can do it, or you know, as a simple example, it just seems like it's going to open up more opportunities for us to do the things we love, right right.
I think the terminator scenario was one that was unfortunately popularized like three decades ago by media that did media, a media that didn't quite understand the technology and be technology that didn't quite understand itself at the time either, like robotics was mostly a pipe tream thirty years ago when Terminator came out right, So I don't really see for variety of reasons, I don't really see robots as being a physical safety hazard, not the least of this
there there. It's actually very easy to stop a robot, especially like a modern dynamic humanoid robot, where if you cut the power, it just falls over. Uh, It's not like it would be a different scenario. I think if we were deploying like huge autonomous for cliffs rolling rolling around to go move things around there there, you would have a real real safety issue if there was a if there was like a security problem. But I think with modern, modern compliant robotics it's much less of an issue.
Now. Before we continue a conversation, I want to make sure folks know what Prisma X is. Get give us
¶ PrismaX overview
the overview to history and your mission.
Yeah. So our our our sort of long term vision is to drive mainstream adoption of robotics by helping the industry scale the things that has a lot of trouble scaling right now. More specifically, that means right now we provide things like a unified platform for robot taler operation, mostly focused on data collects right now, but potentially real world stuff as well, scalable data filters for vision for vision data collection, and things like guild ownership of robotics
which will help the hardware side of the pictures scale. Ultimately, we see Prisma access backing a lot of the man machine interactions in the future, where you would go through things like remotely operate robots to complete tasks, be the guy who ensures that the robots are safe, coordinating a lot of the ownership, leasing, access mech mechanics behind robotics.
Interesting. So, you know, we talked a bit earlier about
¶ Jobs from Robotics
certain jobs being taken out by robots that maybe are more dangerous to human beings and they can work around the clock. But do you see new jobs being created where people are going to have to fix these robots. There's a financialization, the leasing, the buying, and all that jazz. What do you see as far as jobs being created out of the robot industry.
I think like the actual hardware maintenance will be no different than existing equipment leasing, but there's going to be a lot more of it. It's going to exist at a too orders of magnitude higher scale than existing leasable equipment, which is going to require the creation of new structures in order to manage these things. Right, you can go manage a million bulldozers, but one hundred million robots is
a whole different picture. Beyond that, the real position we see being created is this idea of the human behind the robot who drives the machine, but he doesn't have to drive the machine one hundred percent of the time. The robot drives itself ninety percent of the time, but ten percent of the time someone needs to come in and make sure that the robot's doing its job or
do a spot check. And if you do the math, if there's a billion robots or even one hundred million robots, we're talking millions to tens of millions of positions being created. And ultimately we think that's where a lot of the existing sort of like gig worker demographic. I would say, like people who right now make money doing repetitive tasks
will eventually be able to occupy themselves. And it's a much better use of humans as well, because, as I said earlier, people are good at making decisions and learning about the environment. People are actually quite bad at doing things like lifting heavy items. Like we're not built for that kind of work.
Yeah, and it seems like we're in this transition period away from that, right and the old Industrial Revolution, the old way of doing things, and look, we needed to do it is what it is. We have to do
it at that point because the technology wasn't there. But it's like this brave new world where you have AI as like this virtual assistant that give you knowledge point, you give you ideas, and then you have the physical component of the robots to help execute some of it, and you as a human being, you get freed up to, like I said, do more things that you love. Maybe I don't have to spend all day doing maintenance on my yard or my house or whatever it is. You know.
I know it's a very simplistic example, but I can go fishing, or I can go work longer things like that.
Yeah, I mean, I mean driving an analogy with existing gen AI. Right, there's like a lot of concern especially like AI content generation is going to displace a lot of jobs. But the top content creators are like really jumping on the AI because it means they can spend more time being creative and less time and money doing things like correcting the color and their videos or a masking out objects like it's a it's a powerful tool for people who right now are already ahead of the
curve to do better. And what that means is that the whole curve gets pulled up right. You've displaced the bottom, and now in order to continue to compete, everyone has to move up and that ultimately, ultimately that makes people better at what they do. And it's happened in every industrial revolution. The first industrial industrial revolution took us from subsistence farming on like like planting corn and wheat, to
manufacturing goods. Then sort of like the information revolution, move people away from doing small manual tasks into doing things like more service oriented things, and overall, like in the long run, that was like we look back on that happily, like we're everyone is very happy that ten percent of the population doesn't sit in a room for fourteen hours leading socks, which is just it was not a good world, right, Like everyone agrees that was a bad world, and we're
happy to be out of it. And I think it's gonna be the same thing. You're not gonna have people going and like manually masking things out in photoshop or being robotics like like clipping burgers or mowing lawns, and they're gonna be going off and doing something new.
Absolutely. And then when you throw into the mix the
¶ Robots space exploration
next frontier, so to speak, which is space. We see SpaceX and all these companies, right Virgin Collecting, So everybody's trying to build their best rockets do space flights. And then Elon's trying to colonize Mars. I can't imagine they're gonna send humans first. They're gonna send robots, right, Yeah.
And that I think is a big hidden gem in robotics, having robots do things which are completely impossible for humans. There's like a whole lot of business considerations you have to think about, right, like what is the commercial value of being on Mars? How can we get to Mars? Launching the robot to Mars is very annoying, it's it's it's a lot of mass to put on Mars. But I think it's an active field and research which more companies should go explore.
Yeah, I mean, and you don't even have to send a ton of robots. I mean you send three. That's still because they don't have to sleep or eat and breathe. That's it's probably the equivalent of what fifty human beings.
I don't know, and you don't. You don't need spacesuits. They can grow more freely like a person in a spacesuit is not a very dexterous thing. Yeah, so the problem, the problem with robots on Mars has always been the few minutes of latency between the Earth and Mars. Like the speed of light is finite. So it's a little bit troublesome, and they're trying to figure out how to deal with that.
Oh. Absolutely, Now tell us how Prisma X is executing on robotics. Are you do you actually are you building the actual physical robots? I saw on the website the fleet tell us a bit about that.
Do not build physical robots beyond what's absolutely necessary to kickstart the platform. We partner with external robotics companies. Ultimately, one of our short term midterm visions is to drive a little bit more standardization in the robotics industry in order to make it easier to build applications. When you have three dozen different types of robot, each with their
own quirks, it's hard to decide what to do. But by pushing the platform and the data and the operators to be skilled on a few embodiments which we think are we think are the best embodiments of that class. We can make things a little more efficient. So we do offer some robot access right now, and as the platform grows, there will be more access and more community owned robots which will be made available through the platform
for everything from data collection to task completion. There's like, if you like in not too long, I think if you want a human oild robot to come mow your lawn, there will be potential to go on the platform and try to hire a humanoid and find an operator to mow your lawn. Whether the early platform, in the early hardware be efficient at actually getting a lawn mode is a different question, but it's it's right, it's a it's that's that's a nice thing about building something decentralized. The
community gets to grow and discover what's valuable. Like we offer the tools to hire a robot, hire an operator and make them do things, and what you want to build on top of that is up to you. Like it could be anything from collect data to mow your lawn to who knows what else, put on a song, put put on a show, which is probably going to
be actually a big use case. Or you go hire a humanoid to stand in front of your restaurant and like wave around like people will come either just look at the robot.
Yeah, it's it's man like I said. A brave new world is so fascinating. It's exciting. Which blockchains or blockchain
¶ Which blockchains are being used?
or blockchains are you using to power your network?
It's still a bit TBD right now. We do a little bit of collaboration with some of the more prominent projects. The nice thing about the way the current ecosystem is set up is that you do have the potential to have different deployments of robots and different communities on different chains, depending on depending on where the communities are, what they're interested in, and then use technological means that tire reading together.
And you know, I know the future is interoperable. We're seeing a lot of folks issue stable coins and doing different things on multichain right because you have these chains that maybe have different swim lanes, like an example be etherorem is more on the enterprise corporate level salon and more in the retail meme coin n FT level Bitcoin in its own lane, and you know some folks are targeting sports, the sports industry, so maybe it makes sense
to be on multiple blockchains. Until you said targeting different communities.
Yeahl L wonder ferentiation right now is a very interesting thing. Like I've seen the entire industry well before the concept of L one L two was a thing, and that was bad because you needed to become a node to have a wallet, and uh right, like coinbases, original product was a wallet where you didn't have to go download the entire chain in order to have a bitcoin, and
it moved on from there. So I think as the industry moves forwards, we're going to see more clarity and what the what the chains differentiate as right now, there there were there was Bitcoin, which is slow as balls, then if the ethereum, which was still sort of slow but offered like the ability to become programmable, and then sort of moved on for there we got we got L twos to go solve the ethereum like transaction speed problem, and then we have all kinds of novel novel novel
systems which are less and less decentralized but more and more practical. So we'll we'll see, we'll see where that grows in the end.
Do you envision Let's say I buy a robot right five years so, now, just an example, that robot would have a crypto wallet built in. It's powered on maybe multiple blockchains where there's a primary blockchain, but it can
do interoperable things with different blockchains. I tell it, hey, robot, whatever or whatever name I give it, can you go to home depot get me this part, and it will go do that and it will pay because it has maybe a wallet that has stable coins, and because stable coined legislation and crypto legendis at the cusp of passing, it will all be you know, into the payment rails, the credit card companies and so forth, and it will that robot will come back with the product that I
asked it to buy and able to have executed a payment and everything.
Yeah, so that's actually interesting. I think that you will be able to buy a robot which contains some sort of credits, and those credits can be used for a variety of tasks machine to machine. Obviously, the robot completing tasks, the robot making some sort of external transactions, and it's up to the world and the regulators to decide whether those credits will be stable coins, which can be used
to also pay for most real world services. I think there's a lot of potential with the new legislation because people are there's a lot of new services which MasterCard, Visa discover refuse to touch. Like fundamentally, the value of the stable coin outsid Like obviously one thing is international transactions, which are just sort of bad right now, like the financial system was not set up for high speed international transactions.
But the other one is being able to do things which the three big credit card processors and the merchant banks are uncomfortable doing because of chargebacks and risk, and I think I think robotics is one of them. But maybe if things don't go well, you're just going to prepay for API credits which will be usable. They'll be like uber bucks or whatever, which will be usable and selects words. But I think I and everyone else really hope that the it will be a lot more universal
in terms of adoption. And I think Stripe moving in that direction really helps because they're a big back end processor for a lot of SMBs. I like stripe board process for big shops like home depot, but for stuff like hey, can you go buy me breakfast from this my favorite restaurant. That guy probably uses a stripe terminal to go do their processing. So that kind of integration
will really really help. The key I think there is to go hide the crypto side from the from the merchants, Like the merchants have no business learning or understanding that there's cryptography. Right, it's just another payment system and it's secure. Like no one knows how swift for credit cards work, or people who build the systems do. But like in general, like when you process a credit card payment, you don't know what happens behind the scenes, but you don't need to.
It's just another way to eventually you get dollars exactly.
And you know, often people ask me, okay, what's what's the purpose of blockchaining, crypto and stable coins right, because they don't know, like to your point, how money is settled, what happens behind the scenes. And look, a lot of people don't want to know how the sausage is made. But once they understand the crux of the issue of swift and the days it takes to settle and messaging, this technology solves all of that, makes it more efficient, faster, and you can lower fees as a result, right.
Right, Its exactly right. There's like a lot of moving bits when you swipe your card in a restaurant. It like there's a merchant bank and ah settlement system your bank, and there's just a lot of moving parts that are really inefficient, and the different parts of the stack all disagree as to what they want to do.
Yeah, for sure, does your platform allow for the creation of tokens or is that not necessary?
Uh? Yeah, yeah, so right right now, we're going through a point system for some of the some of the early early early interactions. But eventually we're going to have a native token.
And I know you mentioned you're still trying to figure out, uh, you know what, what which direction you go? Are you gonna build your own layer one? Are you gonna build a layer two like on Ether or something like that.
It will probably start as an L two probably probably not probably probably Solana or some some one of the newer one of the newer chains with a bigger community, like the Youth community, has a very well defined, well defined set of features right now, which is mostly focused on technology development and as you said, enterprise use cases, which is not not a not a super good fit I think for the more community oriented transactional stuff that
we're doing. I love etherory and like, if you're there, they're great guys. They really got us wherever we are. They do a lot of amazing research, but they have their Every chain has its own personality to your point, right, like we really that's what differentiates the chains. It's not the technology or the settlement speeds, or the validation scheme or the security. It really is the people who have gathered around each chain and that that that that that's
that that gives them different personalities. And for us, I think it's because we're doing for for now, we're doing a lot of more offline transaction settlement. Things like transaction processing speeds are a bit less important dust than if we were, say doing real time payments where you actually need like some second settlement worst case for everything.
Man, it's just I go into deep thoughts sometimes thinking about all these things with robots and blockchain and how it's all going to work. Now. You guys recently raise
¶ Raising $11 million
eleven million dollars from a sixteen Z and we know Andres and Howitz they are. They are big time believers in tech. They have been since the nineties. Crypto as well, tell us about that fundraise.
Yeah, yeah, so, and A sixteen Z is a great is a great fund for a project like ours because they're very plugged into the frontier of every piece of technology and that includes robotics as well. Like they but by virtue of being a sixteen Z, they've interacted with many robotics startups. They have a good they have a good pulse on good pulse on what the industry is going to grow, and they like they just really immediately
saw value of what we're doing. Like behind robotics, there's a huge late set of human coordination and human services that are needed in order to make the industry take off. Uh then hence hence the investment mm hmm.
Yeah, And I often watched their portfolio to see what they're investing in because it's also uh it's it's kind of like showing where the direction the puck is heading in. So, yeah, they were early to obviously the web one point oh, Web two point zero and now three point h and then you got these other technologies. So Billy I wanted to circle back to you know, the impact of robotics
on society and the economy and so forth. You know what if someone came up to you and said, okay, Bill, I have seen all these movies Terminator and so forth, and if you wanted to play Devil's Advocate, what would
¶ Risks of Robots
you say are some of the potential risks? And once again, every technology brings risk with it, so I want to make sure you know, we give the nuance there.
Yeah, yeah, I think the I think an interesting thing which I expect to see a bit of case law soon is if a tele operated robot makes a mistake, who's liable the operator or the operator or the manufacturer of the robot. To me, it's clearly the operator. But it gets a little more nuanced if the operators are being like hired through a platform service, right like say you hired an operator through task grab it, which is what a lot of companies do. Now, who's responsible for
for for the operator's actions? And you know as well as I do that once robots take off, there's going to be like some incident where a robot is teleor operated and goes and like club someone in the head and there's like a big ruck us around it. It's not going to be a huge issue, I think, because in general society is not full of people clubbing each other on the head, Like even in the absence of punishment, It's is not beneficial for you to go like get enough,
get enough fight with people. It's a lot of risk and there's no guarantee you're gonna win that fight. So it's I don't think it's it's I don't think it's natural for people to go do that kind of stuff, but there will be isolated incidents and we're gonna have some sort of legislation legislation around that, I think, And it's it's Decentralization helps there because it makes people individually
responsible for their own actions. Like it's it's there's no like central hiring agency, so there's less of a target and you can build up the network obviously, like if people misbehave, they're no longer well come on them on the network.
And then with these different robots, right, and you have different ais, you get your chat gpt is and much more. And all these AI agents that are going to be created, what category if I'm phrasing that right, what of AI would power these robots to help them do what they need to do?
There are a new emergent class model. They internally look a little bit like the large language models the power chat EPT, but they learn how to plan a factions in a very abstract mathematical space. They like understand the concept of moving your arm to the left and then picking up an object, but they do so without explicitly writing that down. They're just they're just numbers which describe
the concepts of leftness, armness, and picking upness. So the model will go learn those and then finally, different hardware vendors will go adapt those backbones to their own hardware. Is the current is the current paradigm? And I think it's here. I think it's here to I think it's here to stay. I think I think that's a good approach.
And in addition to the physical aspect of helping human beings to execute different tasks and saving us time and like I said, freeing us up to do what we want,
¶ Robot companionship
can these robots also be a companion and smart enough to have a conversation like we are right now where once again I give the example my grandma that the robots there, and she can ask it questions, You can ask it to make a phone call, so you can say, hey, can you remind me when is this happening. Can you put this on the calendar and it can have that interaction.
Yeah, I think the helpful the helpful companion has been a vision of a lot of even the software only companies without the hardware embodiment. And I would say the biggest challenge there is this idea of talking like building an AI system where you can interrupt it halfway and
it knows how to respond properly. If you've ever tried to like speak to an LLM, like not like chat with one, but like actually talk with one, it's very unnatural because you have to finish, pause and wait for the model to respond, and if you then interrupt the model as it responds, it just like really goes south from there, like it doesn't really know how to handle injecting outputs, as injecting additional input as it generates output.
And I think it's a problem that it's a whole can of worms that I think is also right for decentralization, where you could go collect the training data needed to train this kind of interaction. So like, if anyone's listening, I think it would be a helpful thing for the industry to be able to really train that model because right now just isn't enough conversational data to do the training right now, Like the technology is there, but the training data isn't.
This is a bit dystopian and gets into kind of
¶ Robots and laws
the liability aspect. Let's say these robots have the guardrails in place. You cannot attack human, you can't hurt a human. What if I tell my robot to go attack my neighbor's robot? How do you how do we stop that?
It'll be resolved through the traditional legal system where if you go, like break your neighbor's possessions, you have to pay for them, Like like what stops you from like taking a hammer to your neighbor's car, right, the answer is you have to pay for his car plus like triple damages, So you're not going to do it. And as I said, I don't think I don't think humans
are like society. Seems that I have read a human which is not like naturally conflict seeking at least in like ninety nine percent of humans are not confluence seeking, and the bad actors can be dealt with through the traditional means, Like fundamentally, a robot is a tool, and using a robot to hurt someone or using a robot to damage someone is like, if you're in control, that's no different at all from using like using a hammer
to go beat your labor on the head. Like it's punished by the same mechanics, and the thing which guards you from doing that is like society and penalty is not the fact that clubbing your neighbor with a hammer is difficult. Right.
And also I'm assuming the robots are going to have kind of what a black box like airplanes do. It's going to record all of these things, so the evidence is there and it's actually be something you don't want to do if you're trying to do something nefarious and keep it as sea.
Yeah. Yeah, Like if I think it's pretty clear that if if you personally own own it's your robot, and it goes and does something bad, it's it's your problem, regardless of how it happen, Like it's you're the It's the same way like if you if you own a knife and it gets something it gets used for something bad, or you own a gun and it gets used for something bad, you'll get looked at and especially if you are the one behind the knife or the the or
the trigger. Right, And I think it's the same thing for machines, Like the robots are just tools in the end, and they have a little more autonomy, but you're still like the people are still responsible for the actions the machines take.
Billy hard question, what's your best guess as to the timeline when we're going to see robots in our homes? Is it ten years from now? More than that?
Yeah, I think ten ten years is a good estimate for when you'll really see them in your homes as something that you can go buy off the shelf and be reliable enough and comfortable off to work with. There is a bunch of software problems that need to be solved along the way, and probably some some hardware problems as well. Dev definitely is. I would say like social acceptance is also also one that we need to step into. We can't just drop robots and everyone who everyone's houses
call me maybe a pessimist. There's at least one prominently venture funded company that said that we'd have robots in your home last December. The robots currently aren't. You aren't in your house right now, So that's how it's shaping up.
Do you think it might actually start in the business sector? First, right Amazon's warehouse and whether before it even makes it to our.
Homes, absolutely, and for for deep like deep scalable stuff like Amazon's warehouses, I think we're gonna see automation happen very soon because they can go rebuild the entire warehouse in order to make the machines fit and reap the ten year returns on that investment, like they have the
money to do so. I think where you're going to see the first deployments of what I call advanced robots, like modern robots with complicated actuators, is in small, small to medium businesses like things like corner stores, convenience stores, maybe grocery, maybe maybe restaurant back end to some extent, but there's a lot of like human interaction there, so
maybe maybe not as much. I think that's where you're going to see the first deployments where there's it's it's complicated enough of a scenario to justify a modern robot, but straightforward enough and sort of enclaves enough that you can deal with some of the reliability issues early on and then we're going we're going to move on in there once Once again, there are a few highly opinionated companies that will definitely tell you differently, But that's sort
of my my take on the situation right now, and I think I think it's the realistic way for things to grow for sure.
The other component that this is energy, right uh, And maybe that's kind of solved because we have car batteries and the alternators that keep powering it and you could charge your vacuum. You can charge this device and that device now, But what do you think the initial lifespan of a battery in the robot would last? And is it just a home charger that it just goes and stands and plugs itself in.
I think it's going to be around two hours, maybe three hours for the for the first robots, So you're going to need some sort of fast charging technology to really get at the takeoff. Batteries are a trick Batteries
are a tricky thing. I did a stint as like a sort of traction like electrification engineer some some time ago, and there's really a lot of laws of physics that make it hard to make the battery two times better than what it is now, and that puts a lot of like there's a lot of hard physical laws and we're like getting pretty close to the optimum I think for actuation and battery and power draw and I think we can recover some of the battery life problems by
changing the form factor, Like if you accept that the robot doesn't have legs, which I think is a great idea in the in the short to medium term, you can pack a lot more power, it becomes a lot more efficient in moving itself around, and you can maybe squeeze a day of battery life out of a reasonable machine.
But it's engineering is a series of trade offs, right, And who knows what will happen in the future, like maybe probably batteries in the un foreseeable future, fuel cells if you're still that kind of guy, Like if you're still if you're still interested in hydrogen in twenty twenty five, which you get like a startup a year that tries to play that game.
Yeah, man, it's just fascinating. I'm just thinking, like I have a seven year old daughter. I my wife, of course, and I'm just thinking, how would that robot fit in our home? Obviously it will get things done, it will be productive, but how would my daughter interact with it? And would that be you know, helpful for her where she can have some sort of companionship, not replacing us as parents of course.
Yeah.
Yeah, but you know, she can ask it random questions and I can ask it random questions kind of like what I do with the Amazon Alexa or to Google Home and things like that, but get even more value.
Yeah. I think there's a lot of human interaction problems that people will need to think about once we're past the baseline of getting the robots to actually complete physical tasks the whole, Like h CI aspect and the product experience are things which no one's thinking about right now, and I think there's a lot of work to be done there, especially for domestic robots.
And then I'm assuming you can choose a personality, right if it's a male, female and different things like that.
Yeah, yeah, And a lot of it is like like companion chat and like interactive AI systems which talk like humans are a that they're a It's an interesting space, right, There's like a lot of things you can do. There's some good, some bad, and there's a lot of social
risk of as well. Uh. There there are a lot of people who actually don't understand they're talking to a machine, like people who aren't tech savvy, people who aren't so tech savvy and aren't plugged in, And I think it's a it's going to be a bit of an issue once we have these machines to teach people that there isn't a human at the other end, like it literally is a computer talking to you. Uh, Like context hallucinations,
long term memory. These are all things that like that they align well with like the traditional text AI space.
I think you may have answered this, but I want to make sure I ask you this question. There's always the question of something becoming self aware AI and so forth. Is it blockchain that will once again police that if something becomes self aware and truly becomes autonomous. I don't know.
I think, oh, man, the whole consciousness in AI systems is an interesting question. The way AI works right now is it's a statistical model of everything that's seen in the past. So it's very difficult for these systems to become cautious in the traditional sense. They do exhibit a
lot of emergent behavior. And I think you're right that blockchain or some variant of it should be the tool that controls what behavior is acceptable and what behavior is not because it fundamentally aligns with the community and it's its values, rather than with the risk profile of a trillion dollar technology company, which is obviously very limited in
what it can do. And I think once you start, uh, once you start understanding the fact that these systems are going to be interacting with different cultures, some of which have vastly different beliefs from yours, it makes sense that the policing is done in a like more more local way,
the decisions are done more locally and more specifically. I think it's it's the the community will be able to select what data goes into the training splits of these robots, of these models, and and that really determines, to a very large extent, the personality of the resulting model and what it can do.
Mm Billy, have you seen the movie Chappy've any chance
¶ AI avatar in Robots
which which movie Chappy?
I have not?
So in that movie, it's it's it's a world where robots exist, and that one of the robot engineers he uploads his consciousness to the robot. That super sci fi.
Stuff here, right, Yeah. Yeah.
However, I'm currently exploring, and some other creators are as well, creating an AI avatar of ourselves where it has learned my speaking patterns, my you know, my quirks, uh and and a lot of my knowledge. So I like, I upload a lot of my transcripts from my podcast and so forth. Right, So let's say I create that avatar, and eventually I upload that into a robot. And let's
say I pass away one day. You know, I'm of old age, but my grandkids want to have a verse this is this is kind of weird, but that AI avatar is in the robot and they can have me around in that way.
It has been explored in the in the like non embodied sense, this idea of a magic box that carries on your uh personality when you're when you're gone, when you're not available. Uh. There's probably a solid mark it for everyone, you know, everyone. People do two things with me. Certainly, we're born and we die, so everyone that lives dies, right, so your your market's like the entire world. The social impact of that technology is curious, and right now it's
not widespread enough to go beyond science fiction. But I think there's gonna be a lot of like uh like teaching people that it's really just an avatar and not something it's not like the actual person, and it's not as capable of learning from new experiences as a human can. And I mean there are ways of there are ways around this you can go learn from you can go like literally make the model learn from its new experiences. But no one's shipped that product yet. It's kind of expensive to ship.
Yeah, I mean, it's it's like weird, but it's also interesting at the same time, because if you love somebody and you kind of want something of them around, let's say they pass away at a very early age or whatever it may be, and you've seen bits and pieces of this in movies, you know, that could be very interesting. But also, like I said, socially, I don't know if we're ready for stuff like that.
Hopefully very strange, very very strange.
Yeah, okay, what's on your roadmap? But what can we expect for the remainder to year.
Yeah, so we're going to launch more on our platform very soon. There's going to be robots that people will be able to remotely drive. We're planning on scaling up the standard platform and a few standard embodiments that we really will be able to collect a lot of data on and hopefully be able to show real progress on
the frontier technology front through our partners. That being able to scale the data to rather than dozens of operators, hundreds or thousands of operators really gives you like new new magic and what these machines can do and help move the needle forwards and as we say, move robots when we're into the mainstream by giving them the same kind of flexible lily that the language models have.
¶ Crypto outlook
You know, you were mining bitcoin very early, so here in OG you've seen this asset class grow significantly multiple cycles. What are you expecting, as you know, crypto legislation comes into play, expecting a lot of innovation this industry to grow significantly.
Yeah, I think it's going to be really good. The generally like crypto friendly, and I would say that the current administration is just more friendly to advance financial instruments than the previous administration, and that means you're going to see a lot more exposure in marketing and education about this new asset class. Stable Coins for sure are really
useful because they do like every want to read. Stable coins do have utility, and there's going to be a lot of investment around building new stable goin driven platforms but beyond that, we're going to get a lot of clarity on what differentiates securities from tokens, and I think that's going to encourage a lot of building in that
space as well. So sort of like one of my it's very much an opinion, but I think a lot of us would agree, right, It's like we think it's good for people to be able to invest in new technologies more directly than they do now. Like there's a lot of frustration that open ai and Nvidia and these
open Ai especially right, they're a big private company. The people who work there are making billions of dollars off of their work, but you don't have any way to engage with this progress other than losing your job to the AI. So being able to let people participate earlier through through tokens, or through network incentives, or through decentralization, I think is going to be a really good thing for sort of equalizing things and getting people more engaged
in frontier technologies. Like once you have skin in the game, everything changes. Like I can guarantee you that if you had open ai shares, you would not say a single thing bad about open ai. But there's like a lot of concern that like ohai is going to displace jobs, and just like being able to interact with the development and the growth of the industry as a whole is
is really cool. I'm really healthy for everyone. And as I said at the beginning, with in this modern information age where information is a lot more accessible, I think the world is ready for people to be put like ordinary people to be able to participate in the frontier of technology development in whatever way they want, and like
that that's what we do. That's a lot of what the AI AI plus crypto platforms do right, Like you can contribute, not necessarily through like writing ithon code or building models, but like through data, through contributing your experiences or your chat logs, And that's your way to get a little bit of skin in the game. And when everything takes off, you'll benefit and you'll feel a lot better about it, like you're part of the project. And
I think that's really good. And I think ultimately the legislation that's shaping up is going to be really helpful in guiding companies, especially in the US, to offer more access to be able to participate.
Yeah, absolutely great perspective and it's just fascinating man, I'm so excited to see how all these technologies roll out. And that have a seven year old daughter, so I'm trying to educate her about all these things and right right, just prep her so to speak, like this, the world is going to be very different, and obviously she'll find her a way. But if I can, you know, teach her a bit about robotics earlier, AI and all these things to give her Yeah, yeah, exactly, Yeah awesome, Uh, Billy,
I have some wrap up questions here for you. First,
¶ Wrap up questions
if you could create your own metaverse, what would the theme be.
My own metaverse? That's interesting.
It'd be robot themes.
Probably probably robot theme, probably sort of frontier tech themed. Metaverse? Is a metaverse? Is an interesting question right now, and I think there's a lot of potential for that space to come roaring back to life once people build more applications. The first attempt to build the metaverse was it didn't go very well. People were selling plots of land and video games that didn't exist yet. But I think the
second time around, we're gonna see it. Honest would probably be like gaming the like recreating, recreating the incredible virtual worlds of the early MMO era, like there were real communities where that people really enjoyed based on in the world of Warcraft and RuneScape and all of these big decentralized MMOs, and by and large they've they've gone by the wayside as people have switched to like the sports type games. But I think there's potential for these things
to make a comeback. Like people lived their lives and like were happier because of it in these virtual worlds.
Oh for sure. Yeah. I mean the first iteration not great, but you know, version one point zero of anything is not great. But yeah, yeah, as again later version.
I think there is potential, Like there's already precedent, like there were in the early MMOs, like the the in game money was monetized, just not officially for like legal reasons, but like you could go buying game credits using real money, and there was like a whole economy of in game items and communities, like people built cities and there were guilds, and like replicating all of that into the modern era I think would be really freaking cool.
Mm hmm, yeah, that would be awesome. All right, I got some rapid fire questions here for you.
Favorite food, favorite food, probably some sort of sushi.
Favorite musician or band.
Uh none at the moment. It used to be Taylor Swift a few Taylor Swifts ago.
My daughter would be happy to hear that. Favorite movie, Avengers Endgame.
Favorite book. That's a good question. I don't really have one at the moment. Found Asimov's Foundation, if You're familiar, is a good one. It has a lot of interesting concepts. It's a It's a sci fi piece from the late sixties. I think I'll.
Check that one out. I haven't. I haven't heard about it.
I think there was an Apple TV series based on Foundation, but it wasn't wasn't particularly true to the true to the story.
Ah, when you're not working at prismat X, what are you doing for fun?
I'm still building robots, I think.
Bailly, absolute pleasure chatting with you, and I love to have you back on because I'm so fascinated by all this technology and the direction the world is heading in. But thank you so much for joining me.
Yeah, yeah, absolutely absolutely. I love discussing frontier technology and how it will change the world, and I'm happy it'd always beyond