Orchestrating the Swarm: Tashi’s Vision for Collective Autonomy

Rich Robinson:

And we're back. Ladies and gentlemen, boys and girls, welcome to the Intercognitive Foundation podcast. I am Rich Robinson, founding chair of the Intercognitive Foundation, where we make the physical world accessible to AI. I am joined by Intercognitive Foundation founding member Tashi. Please help me in giving a warm welcome to Amar and Jay from Tashi. Amar Bedi, Jay McCarthy, welcome.

Jay McCarthy:

Good to be here, sir.

Amar Bedi:

So excited to be here.

Rich Robinson:

Indeed. I am dialing in from the warm embrace of Mother Ubud on the Island of the Gods, and you guys are not too far in Singapore. I love the Lion City, and I ain't lion. I have known Jay for a very, very long time. And you, Amar, I've really gotten to know, and I'm blown away by your incredible turbo engine under the hood and all of your ambition and background.

Why don't you please share for our listeners a little bit about your hero's journey? Then I'll have Jay, in turn, talk about his and how you guys met.

Amar Bedi:

Sure, super excited to be here. That dialogue, "You had me at hello," that's how it was with Rich. Jay introduced us. From there, when we found out Intercognitive was going to be a thing, we were super excited, like, "Yeah, we have to be part of this." So anything Rich is doing.

A quick intro about myself: I'm from Delhi, India, and grew up there. I started my life in finance, doing investment banking and consulting. I was with KPMG for a few years and with this Swiss bank, Bank Sarasin, doing investment banking. I got into the startup world by joining a corporate venture firm, helping them acquire a few startups in Singapore and do venture building for them across Southeast Asia. We set up a live video streaming platform across Indonesia, the Philippines, Thailand, and China, back when it was new and cutting edge, almost 12 years ago.

It went pretty well, scaled to 7 million MAUs, and, as with startups, also went bust because we ran out of cash. After that, I went to a later-stage company, Uber, and set up their driver ops in the region. Then I joined Grab and set up their B2B business in the region. By then, I'd done a bunch of things on the business side, so I thought it was a good time to start my own thing. I left Grab in the middle of COVID because that's when I realized life is too short to be a cog in the wheel, and started my first startup, which was a freelancer marketplace.

I scaled that, and it went pretty well. I got into Web3 as well. That was the big cycle in 2021. I went deep down the rabbit hole and read a bunch of books to understand what's the core tech. What's the hype driving all these crazy NFT prices? Why are all these guys getting into this who aren't just the speculators? Really smart, savvy people are massively bullish in this world. There has to be something I don't understand.

That was the truth that came out of it: I understood why decentralized technology is critical, why the future is decentralized, and why blockchain and similar tech is going to be the key underpinning technology of the future. At this point, everything was very hype-driven. But typically what happens, as with the internet or mobile phones, is that all these things that were new-age tech when they started just became ubiquitous when the hype died. I wanted to be part of that movement from the hype phase to the ubiquitous phase.

We did a few different things back then. I converted my startup to a Web3 one, did a small Web3 fund with friends and family investing in NFTs and tokens, and that went okay. I've known Jay for the longest time. As he introduces himself, he has a pretty storied background and is pretty famous in Singapore as well. He's a cool guy I've known for a long time, and we were chatting about doing some interesting stuff. That's when Tashi came about, and I was like, "Heck yeah, let's build a platform that's actually going to have long-term meaningful impact."

When I spent time at Uber and Grab, the one thing I realized was what it meant to have a global vision, what it meant to build a platform that can globally scale. That's something you're building for the long term. It's not about building something in a silo, in a bubble, with small revenues, etc. You're making something that has meaningful impact that will stick around whether you're in the company or not.

When we started talking about the vision for Tashi, I was super excited. We came together and decided to start this. The idea was, "How do you take this technology, blockchain or something along similar lines, and apply it to real-world use cases?" That was the missing piece we found: blockchain is slow, and it's great for crypto, but not for a truly decentralized internet.

We essentially set out with the goal to make that happen, which was no small feat. We spent the first two years pretty much in stealth mode, just building out this core consensus protocol, optimizing the hell out of it and saying, "This is going to really be impactful in the long run." I'm super excited. I'll let Jay introduce himself and then talk a bit more about what's happening at Tashi.

Rich Robinson:

I want to introduce Jay in a second, but I want to double-click a little bit. I love your Uber and Grab chapters in your life. As you may have seen, Travis has now gone deep into the robot revolution as well. I got to meet him when he first came to China to land Uber in China, which actually ended up being a big win for them because they basically got shares in Didi.

I think that hyperscaling of something along the lines of Uber and Grab is an incredible skill set utility belt to be able to bring to what you're building now. Let's put a pin in that for a second, but I want to circle back to that and maybe get an anecdote. Thank you for sharing.

And indeed, Jay McCarthy. I'd heard legend of this legend across the ocean in Singapore and would love to hear the CliffsNotes on your very long, amazing, and ongoing journey, my friend. Please.

Jay McCarthy:

Yes, Rich. We've done it together, haven't we? I've known you for some time, and that can happen in these smaller societies in Hong Kong and Singapore, but really in broader Asia. You've done a lot of things and were prominent in China.

Rich Robinson:

No, no, no, don't stop.

Jay McCarthy:

You were prominent in China as a foreigner, and there were not so many of us. I, of course, tried to grab some of that prominence by being seconded for a couple of years to CICC, China's first international investment bank, China International Capital Corp. Morgan Stanley was the 34% shareholder. I got sent up to Beijing for a few years.

I had been there previously as an investor, but it was more intermittent. I think you really embedded yourself in the ecosystem and had probably as significant a China experience as any of us can have. In my varied career, as you put it, I would say China was probably the highlight.

Being there early within a story that you knew was and is big is somewhat satisfying. Even today, as things play out and we're all getting distracted in many other markets, China looms as a big one. I saw a tweet yesterday from somebody that was shared in our politics group here in Sentosa saying, "You heard it here first, China's the next Dubai." That was it.

We were thinking about that because right now there's all this potential U.S.-China competition, and different systems, communist versus capitalist or democratic, whatever the narratives of the labeling are. But in our time, Rich, we were running up there trying to be part of every aspect of Chinese culture and enjoying every bit of it. Kids were studying Chinese. It was all happening. It's taken a little bit of a dip, but is it going to be back soon? Trump and Xi are due to meet soon. World events are playing out sometimes unsatisfactorily, but hopefully towards a good resolution. Are we all going to be running there as the next Dubai? I thought that was an interesting thought.

Finishing your opening to me, I was a New York person given the opportunity via Morgan Stanley to go work in London first and then Asia. I don't know if I grabbed it or they pushed me. Maybe I was lacking in the vision thing a little bit, but I went on what was supposed to be a one-year assignment. That's what they told me to get me on the plane. It was, within four hours...

Rich Robinson:

Ha, heard that before.

Jay McCarthy:

Clarified as a five-year engagement. That was in Japan first. I loved every minute of it. I went down to Hong Kong next, loved every minute of that, and that really became a center part of my business life and life.

I went up to Beijing for a few years, and I was working with Dr. Tay, who was head of private equity for GIC. Towards the two-year mark, he said, "What are you going to do when they maybe take you back at Morgan?" I said, "Geez, I don't know. Any ideas?" He said, "Yeah, tell them you want to come work with us in Singapore, with the government of Singapore."

I did say that to senior management in New York, which was shifting at the time from Vikram Pandit to John Mack. Or maybe that was somewhat during the period of time we're talking about. They said, "You're gone. Get down to Singapore. You've done whatever you had to do in Beijing." And now Singapore is my happy home.

Rich Robinson:

What year was that?

Jay McCarthy:

I got to Singapore 20 years ago. That was when this whole thing started. In essence, I did a banking career at Morgan Stanley. I left Morgan Stanley and co-founded SparkLabs Group with five other great partners. It has been a venture capital firm for the last 16 years, headquartered out of Seoul, Korea, but global in ambition.

I maintained my deal business. We did a lot of prominent deals, anywhere from 100 million to 300 million equity raises was our thing in Hong Kong, with the tycoons and real estate and casinos and tech. We did a lot of different things, which was exciting. It kind of had us on the map there for a little bit.

Then I settled into venture building, which, during turbulent times, I thought was a much more productive use of our time and also much more satisfying. It can be a little bit passive, waiting for the world to turn and paying it forward with the companies. Why not be the builders? Everyone had said that to me anyway, and we finally embraced it as our core ethos. Here we are as builders. I did have an exit during COVID in Equilibrium with my partner, Frank Meehan. Shout out to Frank, also a SparkLabs partner.

With venture building, we don't do it as a factory, which some people do, and God bless them if they can do it well. I just feel that can be a little bit forced. We do it when the facts collide and the personalities collide and the skill set collides. That certainly happened with myself, Amar, and our co-founders. We were probably very close to each other proximity-wise.

It's been a great run, and he's been a great partner. I feel very pleased every day to be working on an important mission, but doing it with people that are inspiring. That's it.

Rich Robinson:

Beautiful. Excellent. Thanks for sharing your backgrounds. I think it's important to be able to set the tone of where this industry is right now, which is very nascent, compared to what it's going to be. It will be the biggest industry in the history of mankind, civilization, but it's really attracting people like your good selves who can make stuff happen. I love that you talked about your tycoon whispering days.

I just want to share a quick anecdote. In that quarter of a century that I was in China, one of my adventures was working with these Nordic investors. They said, "Hey, come on this tour with us." It was my first time to Finland, my first time to the Arctic Circle, and there's a billionaire who has a ski chalet up there. There are more reindeer than people. There are going to be these tycoons from all over the world, but mostly from China. I'm so intrigued.

We fly up there, and then we fly up to this remote place. I jump out of the car first, ring the doorbell, take a look inside the chalet, and you answer the door: "Rich, what's up?" Then we had these crazy, amazing days there cavorting with reindeer and tycoons, which was fascinating.

I digress, but it's such a good story. There are not many times in my life when I'm speechless, but that was one of them. We both have mini shampoo collections from around the multiverse, so it was unsurprising to see you there. Lapland.

Jay McCarthy:

We were in Lapland.

Rich Robinson:

Indeed, the home of Frozen. I will let it go, let it go. Tell me about how you guys really collaborated around this vision and what the evolution has been since the beginning.

Jay McCarthy:

We'll maintain order, Amar, and let you go first.

Amar Bedi:

First things first, as Jay was saying, with anything you're building there are two core things that need to come together. One is the people: are these people that you like and want to work with? And two, is there a clear idea of what you're trying to build and why? A lot of times, especially in Web3, people find the next shiny object to chase after. As we've all been around the block, we weren't those guys. It wasn't like, "Hey, we're doing this. Let's see where it goes." It was, "We're in this till the end, and we know what we're building."

But a whole startup's journey is always about iterating. To your previous point about Uber and Grab, it was different at each company, so I'll add my anecdotes. The most important thing I learned at Uber, and that's why I'm quite excited with Travis coming back into the world of robotics, is that a startup is a series of experiments.

Everything that ever happened at Uber, you could be moving 1% of spend from one place to another or introducing a small button within the app, the team would do hundreds of experiments to define what's the right answer. It wasn't like, "I'm a design expert, and I know people typically like this button to be on the right side, and a black-and-white monochrome is good." There would be 100 experiments where the 100 people would try different things, and the synthesis of those answers would result in whatever the outcome would be. It was absolute facts. It was nothing to do with opinions. That thinking was instilled into everything: as long as you had facts, you were good to go.

A simple story around that was originally Uber had a clear plan not to do bikes. There was a famous meeting with Travis and one of the senior strategy guys at Uber, where he built a 50-page deck to talk about why bikes were important in Southeast Asia. Literally two slides into the deck, Travis was like, "You've given me the right stats. If you're going to Indonesia and Vietnam, you have to do bikes. Let's just go into it. I don't need to see the... Thanks for putting in the effort." But...

Rich Robinson:

Wow, love it, love it.

Amar Bedi:

The right stats were in there, right?

Rich Robinson:

Strong opinions loosely held. I feel really strongly about this. The data changed my mind. Let's go. Yes.

Amar Bedi:

Correct. Everything was very factually driven. With Grab, I'll come to that separately, but back to our story. Jay is a very impressive person who knows the world and had a very clear path of what venture building should look like.

I've been in the startup ecosystem for a long time, and I've been part of accelerators. A lot of times, the plan is spray and pray. There are so many AI venture builders right now where it's, 'Let's go build 10 things. Something will stick.' I think that's not the way to do it, especially once you have a lot of experience under the hood. You should be very clear about what you're doing and why. If you can't articulate the problem really well, then it's not the right reason to start and build a business around. There'll be a hundred things that pop up that you don't know, and that's fine. That's where the iteration comes in. That was one core ethos we had that was important.

I also wanted to work with folks who were thinking globally. One of the things I love about Jay is, 'Let's not talk small.' Every day he reminds me, and I think I'm a big thinker, but he reminds me, 'Hey man, you're thinking small.' As we were telling you, we were on a panel last week at a conference headlined by Donald Trump Jr. One of the guys on the panel was Bob McCooey, who's vice chairman of Nasdaq. We were like, 'Let's start thinking about what our future listing plans look like.'

As a startup, typically you don't think about these things, but he always reminds me: What's the future? Where are we going? We have to have all the parts to an eventual success, and that big-vision thinking allows you to change how you perceive yourself. If you think of yourself as a tiny startup in this tiny town called Singapore, which obviously has a lot of capital here, but as a market it's small, then it's very hard to scale. You have to think about the bigger picture: what you're doing, why you're doing it, what's the one-year plan, how do you sell in two years, how are you building pipelines into it from day one?

All of that has been key to our thinking from when we started Tashi. In our first month, we were talking to the heads of Microsoft Asia and Amazon Gaming and thinking about what an exit plan could look like. That's typically not how a startup thinks about life in its first month. We've been around for three and a half years and had our fair share of ups and downs, but it's that thinking that's kept us going. We know what the light at the end of the tunnel looks like, and once you have that clear from the beginning, it's always exciting.

Amazon likes to call it day one. Over here, I'd like to always think of ourselves as being in that day-one mindset. That's something I've definitely learned a lot from Jay. When you're building something, you want someone in your corner that, A, you can rely on, and B, you're excited to work with. That's what embodies Jay, and he's a cool guy on top of that.

The last cherry on the cake is that he has the coolest house in Sentosa. Whenever we have partners and investors come into town, we do meetings at his house and they're like, 'Dude, what's going on? Who's this guy?' It's just a nice thing that we add on top of everything else that's going on.

Rich Robinson:

It is the most baller home in Southeast Asia, I would say. It's right on the water south of downtown Singapore, and it's amazing.

Jay McCarthy:

It's been renamed Tashi House, so it's obviously community-owned.

Rich Robinson:

It feels very community-owned.

I've done more events and stayed there as much as any hotel in Singapore. It's amazing what you give back to the community. But I want to give something back to you, reflect some things back. I love how we established the ethos, the character. Everything should be ethos, logos and pathos, right? Ethos is the character foundation, logos is the reasoning, and pathos is the story and the 'Let's get excited about the future.' Your foundation character is amazing.

Yet for people listening, you still have to hack your brain and say, 'I need to perceive myself as something bigger.' It's really important to show up every day and say, 'Yes, we belong in this room. Yes, we are going to have this incredible outcome, even though right now I feel like I'm in molasses.' You do that psyop to yourself every day to get moving. The mindset of an entrepreneur, no matter what age and stage, and Travis is the key guy to look at around that, he's hacked his whole mindset to keep moving no matter what. I think it's beautiful.

And day one, I've been to the headquarters of Amazon and it's called Day One. Why? Because Jeff Bezos is like, 'Every day is day one.' Once you get into day two, you start to get into day three, and then stasis sets in. You have this sunk-cost fallacy thesis, and you get stuck. You have to really think about what's the way forward. I think it's beautiful that you've brought that in. Let's get some thoughts from you, Mr. McCarthy, as we say back in my hometown of the city of Boston.

Jay McCarthy:

I will say it concisely, and I very much appreciate the comments from Amar. I think there's a little bit of politeness in there in regard to me, but well received nonetheless. There is that thing called instinct or judgment, and you can be dead right or wrong, so you can't bank it. But instinctively, when you're trying to pick your next gig, you're looking at the person next to you and saying, 'Are they commercial?'

Are their ideas resonating with the way you think about the world? Are they trustworthy? Are they dependable, not only in the good times, but in the crunch?

In Amar's case, that was obviously the case, along with the fact that he seemed to be a methodical, organized thinker, not prone to random judgments. All of us have enough excitement built in us somewhere, but to get those people who think about the world a little more rationally, scientifically, data-driven, whatever you want to term it. I thought he seemed to have a level of maturity too, not to get overexcited. As you're articulating, Rich, every day is either an up day or a down day, potentially dramatically affecting your forward trajectory.

Rich Robinson:

Sometimes before breakfast, it's up and down four times.

Jay McCarthy:

You have to be able to ride through some stuff. I felt he had that kind of vigor, which was very value-added. We had some family connectivity, and we knew each other in a deep way. When we really got around the idea of Tashi, it was nice to think, 'Wow, I can immediately work with this guy.' It would be very natural, and that's exactly what happened.

It's a very short answer, but the humans. That old adage: you read someone in the first 30 seconds after you've met them. Of course, I've known Amar for some time, but you do that screen scan and print of good, bad, great co-founder, not-optimal co-founder. It was certainly that way with Amar, and here we are, and some others who have so far gone unnamed. Hopefully we can get their names in shortly. The journey has been everything we thought it would be, a bucking bronco, but we're backstopped by an enormously important and growing space.

There are so many incredible influences out there. I have to say that somehow my life has zeroed in many times on Elon Musk, just because he's so quotable. I say it a lot of times, and it's getting old. It's like, 'Do you know anyone else in the industry?' But he lays these things out there which really drive you, as Amar was saying, to think bigger, think dramatically, more aggressively, ambitiously. To be reminded that the robotic space is a 30 trillion revenue space playing out. I'm borrowing all his math and all this, but it really is that kind of fuel. Michael Jordan used to get fueled by anyone who said anything.

When we get these high-level quotes, and Dario laid one out the other day which has resonated with me, thanks to all these guys for being such visionaries. He said, 'I've determined the new limit of AI.' You zero in, it's like a headline, and he says it's called 'human imagination.' Whatever we can think to ask for and prompt. That extends beyond just AI. It's the motto of our lives. We're limited by our imagination and what we think we have the right to go for.

When you start to responsibly take the cuffs off and go for some stuff, you're helped greatly by the collaborative mindset of tech. I worked in banking, which wasn't as collaborative in my day, but tech is amazing. It really propels all of us forward. There's that phrase, 'A rising tide lifts all boats,' or tech-industry collaboration propels all good entrepreneurs, ambitious entrepreneurs, forward. We're trying to be those guys, and Amar was, in my judgment, certainly one of those guys.

Rich Robinson:

That's beautiful. I love it. Only limited by your imagination. Your brother is a prominent guy in Hollywood, and most movies in Hollywood are derivative because it's so difficult to make a good movie from scratch. It's supposed to be an Oscar winner, but most definitely are not. Not every play in American football is a touchdown. Every song is supposed to be a hit, and every startup has this thesis that it's really going to work.

But the challenge with a startup is that you don't know until you launch. You have to launch, and then who knows is the end user. The end user is telling you, 'Yeah, I really want that,' and then they contort it or twist it or do something different with it. I don't think, from day one, you were thinking about the robot revolution, but it's really converged very beautifully and elegantly on what you've built. I'd love to hear about that evolution, where you've converged today, and the kind of orchestration capabilities you have.

Jay McCarthy:

Amar, let me go first with a lead-in sentence, because it was my original thing, but then you can expand on this, please. I was thinking, Rich, we're a solution in a compute-starved world. Even our friends at Microsoft in Asia at a high level, CEO level and below, were telling us they couldn't get enough of this and that, GPUs, CPUs, whatever. And it's going to affect the stock price. I kept thinking, 'Wow, here we go.'

It was in the earlier days of this AI ramp, where now it's very clear: we need more energy, more capacity, more everything. You could feel that we were entering this period. For us to be furthering the decentralized quest, living out on the edge and trying to help, bringing back peer-to-peer capabilities, but performant and secure, whatever the adage was, it felt like a very good place for us to be.

Originating this fantastic, well-invested, well-developed consensus engine, which we're optimizing like crazy, made me feel like we could use that to find a place in the industry and pick our spots. Of course, what we always remind ourselves is, 'We're not moving fast enough. We're not moving fast enough.' To this day, we've never moved fast enough, and we're still trying to perfect that, using a few AI agents to juice ourselves up instead of breakfast cereal. We may get there. But that was how I was thinking about it. Amar, maybe you can expand or speak differently.

Amar Bedi:

For sure. I think that's the perfect platform. That vision has always remained the same. As a quick background for the average person, there are two popular ways the internet's backbone of coordination runs. One is called the client-server model, where everyone is a client and there's a server in the middle. That's what all the cloud companies have been: AWS, Azure, Google Cloud. They're all on the client-server model.

If you think back as recently as 15 years ago, half the internet was peer-to-peer. If you think about BitTorrent, Napster, everything was peer-to-peer. You and I communicated directly. There was no one in the middle. For enterprise-focused reasons, where you add in SLAs, peer-to-peer becomes less performant. The biggest gap it has is a lack of trust, because as compute capabilities have improved, so have hacking capabilities. If I'm interacting with you, am I opening myself up to a Pandora's box where something can go wrong?

Our thesis was: how about we combine the best of both worlds? You're running a peer-to-peer environment where, say, three of us connect directly, but we run this consensus on top that ensures no one can break the security or hack into the other guy's system. That's the core thesis around what we built, which is what we spent the first two years on. We knew it would apply to multiple industries. From day one, we knew this was an agnostic technology that applies to a bunch of use cases.

Because we were building core tech, we were like, 'Let's prove it to the world.' At the end of the day, when you do something unique, people need to see something tangible to understand it. The average guy isn't esoteric enough to understand backend deep-tech technology. Even when I speak with CTOs today, they sometimes take a few calls to fully understand this. That's when you understand that these aren't things people are thinking about on a daily basis.

Or think about it in a simple way: OpenAI existed from 2015 till 2022, until the ChatGPT moment happened in November '22. People started going deep into large language models. You needed that thing on top, the large shiny object, for people to say, 'Whoa, that's why LLMs are cool now.' Until then, it was just a random foundation, and people were just talking about the Elon, Sam Altman kind of issues and nothing else. When we spent the first two years, we had the thesis on why this tech was critical.

After that, we started building and chose gaming as our beachhead. We chose gaming for a few reasons. One, because we had a pretty good understanding of the world of gaming. Jay has been our first point of entry when we go into any industry because he has a massive network, and he knew the founders at Supercell and Rovio. We were talking to the senior folks at Unity, the who's who in the gaming world, and they were all guiding us on what that could look like. You want to have a very clear path when you're building in an ecosystem, and gaming offered that for us.

Rich Robinson:

What a beautiful tip of the spear, just like NVIDIA. They started in gaming, right?

Amar Bedi:

Exactly.

Rich Robinson:

That's a great way to get reps and do it at the highest level. That's great. Keep going. Sorry, I just wanted to emphasize that.

Amar Bedi:

So two core factors played out there. Another reason we chose gaming was that it's one of the most latency-sensitive environments in the world. Now robots are getting into that world. But before this, gaming was a mature industry already at $200 billion, right? When you're playing a game, you're not going to want to wait. Fifty milliseconds is the max threshold for your first-person shooters.

So we gave ourselves a constraint to build this technology, and we forced that upon ourselves because we were like, if we can prove this in gaming, every other industry will become easier. IoT is not that latency-sensitive, AI is not that latency-sensitive, DePIN is not that latency-sensitive, and so on. So we were like, let's pick one where we force ourselves to build this technology in the ecosystem that makes it the most performant for everything else.

That allowed us to get to where we were able to tune our consensus down to 30 milliseconds. That's what really stood out for us and makes us stand out. But as we went down that rabbit hole, and these are things you learn along the way, because our tech comes from the Web3 world, we got bucketed into Web3 gaming, which we weren't. We were focused on Web2.

But you don't decide how the world perceives you. You can tweak yourself along the way, but the world decides. So when we got put into that category, it created a bunch of challenges for us because Web3 gaming as an experiment completely failed, and people lost tons of money in that process. So we found ourselves in a room that was not given the most amount of love.

That was the moment for us to look at it and say, sure, this is a meaningful challenge for us to solve. We're cutting server costs for all these large games by half. But if we're not in a scenario where we're growing along with the industry, it'll always halt our future opportunities. So we hit pause in that moment and said that gaming is still key, and we'll still work with a few of our partners. We work with a lot of large players as well. But let's look at where else this is critically needed today. Where are people looking at this as a challenge immediately?

That's when the watershed Intercognitive moment coincidentally happened for us as well, and we started going down the robotics rabbit hole. We're not hype guys by design. We're all builders at heart first. So when we started going down this path, it wasn't like, "Hey, let's be robotics guys now." We spent the first few months doing tons of internal simulations. We hired a few folks and went deep down the hole of understanding: what are the different hardware specs? What are the different operating systems on top? How do they get deployed in the marketplace?

We were getting insights from a lot of players. Who's looking at what? Where do we fit in the stack? What will the product look like? Doing that groundwork was essentially what the second half of last year was, where we felt confident making phone calls to senior folks. Jay helped us catch up with the head of SoftBank Robotics, and we're talking to folks at NVIDIA. We know what we're talking about. You don't want to go speak to those guys without having proven something.

Now we're in that spot where we have tons of simulations already done that showcase why Tashi's tech is better than this centrally orchestrated model that everyone is using right now. In that process, it helped us identify three core use cases that people are now chasing us for.

One is multi-vendor robots in warehouses. If you have a Tesla robot and a Unitree robot operating in different data lakes and having their own sovereignty, how do they interact? It's important that they do, because that's when they become collectively autonomous.

The second one that's been getting us the most amount of attention at this point is drone swarms. We're peace people, not big fans of war, but everything that's happening right now has gotten the world's attention towards drone swarms. Defense is only one use case. If you think about drones, the use cases are crazy, right? Uber and Grab will use them for delivery at some point. Agriculture has a massive use case, and so on.

Exactly. All of those drones are individually autonomous right now, and as these drones flood the sky, they should be speaking to each other. That does not exist right now.

Rich Robinson:

Even entertainment. There are so many use cases. That's beautiful. Okay, great, excellent. So robots dancing together from different companies, figuratively speaking, maybe literally. Number two, drone swarms, I'm sorry. And number three?

Amar Bedi:

Number three is the most exciting one. We're still in the nascent stages as an industry, but it will probably become one of the most important ones in the coming years: the agentic space. So AI agents, as OpenClaw started the popularity around it, and everyone is launching their own variation that's safer, more secure, and better, people are starting to run their own agents. You'll start seeing: how do these agents discover each other and coordinate with each other without compromising trust?

That's where we fit. You have your Intercognitive or Animoca agent, and we have, say, my Tashi agent. How do we discover each other? How do we coordinate with each other? Whose systems do we live on? Whose data does our coordination live on? All of that is what we're trying to solve for, and that's been pretty exciting.

We're actually kicking off a hackathon today. I don't know when this gets released, but it's the 18th of March. We've partnered with a couple of interesting large universities: NTU, NUS, Washington universities. We have people from Caltech signing up, people from HKU signing up, all because they're very excited by the use cases that Tashi is offering for them to build on top of.

Rich Robinson:

That is fantastic. I love it. I saw Elon Musk on a podcast yesterday, and they were like, well, you build SpaceX, and then as you build SpaceX, how are you going to make that work? Then, of course, the Starlink opportunity unfolds. From there, he's like, well, now we're going to put satellites in space that are going to have five times the efficiency because it's always sunny in space, and then we're going to do compute in space.

He's like, I think we're living in a simulation because the most interesting simulations get to continue, and all the other ones get shut off. So in some ways, you're an overnight success after years of chipping away at it. Then you show up and you're like, wow, this is actually where we should have been all along. It's just a very foggy, very winding road along the way.

Now you have some level of height that you've reached, and you're able to see the horizon. You're like, "Wow, there are incredible opportunities that are going to unfold in this space." There's almost unlimited appetite, not just for compute and energy, but also for these kinds of solutions as well.

Amar Bedi:

100%. It's like that old Zen proverb: every step you take makes the path clear. Our job is to keep taking those steps.

Rich Robinson:

Fantastic. So a hackathon right now, lots of movement around the drone space. Between now and, say, 2030, where do you see some interesting anecdotes in real-world deployment on your roadmap, where you can see Tashi behind the scenes? Just make people smarter and expand their minds so they'll be like, well, by 2029, there's going to be this, this, and this. What are some of the things that you are anticipating?

Amar Bedi:

To go forward, I would say, if you go back 10 years, one of the things that everyone was talking about was smart cities. Devices have become smart, but we're not getting the benefit of them being interconnected. That's the missing piece right now.

You can think about it for two reasons. It can either be enforced from the top down or from the bottom up. It's hard to enforce this from the top down because if you think about large enterprises, everyone has their own ecosystem that they prefer. Apple lives in its own world, Tesla lives in its own world, and so on. You can't go tell Apple, "Work with Tesla because I tell you to," and vice versa.

Even in a place like Singapore, which is the most unique bubble in terms of what a smart city could look like, we've seen some experiments. When I was at Grab, we were trying to do smart cities in Indonesia and Malaysia, in small places. The technology still operates in a silo, right? Each device with IoT capabilities, all that exists today. What we're trying to think about, say, five years or 10 years down the line is the bottom-up part of the stack.

Can you have something that's truly open source, truly decentralized, that anyone can run and benefit from everyone else around you? When we think about that smart-city world and the future, one of Travis's goals at Uber was autonomous cars, right? Autonomous cars need a bunch of things to become true. But one of the key parts will be: how do all the different cars work with each other? How does the Waymo car work with the Tesla car, with the future GM and Ford car, and with the BYD car?

Right now, because they're individual, they all have to put everything in that same model. But our vision down the line, as we're building Tashi out, is that you have these lightweight Tashi nodes on every device, and everyone's sharing information with each other that makes everyone more collectively autonomous and more intelligent. That's what the 2030, 2035 version of us looks like down the line, where everyone is creating collective autonomy without losing their own sovereignty. That's what the long-term vision for us looks like.

Jay McCarthy:

Let me add to that real quick, Rich, before you move topics. Just trying to be a little bit forward-looking here, so it shows that we can all think about the future of robotics. I still think we're thinking about physical hardware robotics too many times. Amar did mention AI agents, but there's a rapidly growing field within this space which we should all be thinking about at the same time that we're thinking about the hardware stuff.

That is what they call soft robotics, right? Think about that for a minute. What about soft robotics for things like delicate handling in health applications at home, environmental monitoring, soft-item handling, groceries, this kind of stuff? A lot of this could be considered flexible, bio-inspired even. We've all read about nanobots going into our bodies, capable of operating in unstructured environments.

I think we're in the clunky hardware stuff, which is still impressive as hell. Whenever I see a Unitree robot dance at a function and do flips, we're all blown away. But take this to its natural end, where stuff's getting smaller and smaller and more and more flexible. Even that phrase I used, bio-inspired stuff circulating through our bodies.

All of that at least has the commonality of communication, right? Being able to communicate, being able to take guidance, being able to learn in this new environment of Jay's bloodstream, like, what the hell? Who's done a YouTube video of that to train a Waymo? Nobody, right? That's where this is headed, which is why we always felt that if we could be good at assisting communication or data sharing or learning on the fly, we're going to have a use case in this world.

There's a very well-known, well-watched video that we've referred to occasionally. I'll mention it now. Waymo is a self-driving car. Bless them that they're leading the industry. They're driving people around San Francisco. They've been trained on millions of hours of YouTube, and whether you're using cameras or just using software, Elon's point, it's all happening.

There was a video a few months back where the Waymo, with all its fantastic training to date, cruises down the road, knows the road, knows the lights, knows the intersection, goes to take that left. Unfortunately, there happened to be a drug bust, and the cop's out on the street, he's got his gun drawn, and he's telling the guy to get out. Then there are two elderly people in the Waymo. They're at risk, the guy's at risk, the cop's at risk. What does the Waymo do? It just goes down the road, takes the left, ignores the human, but doesn't hit him. It does its job.

If there is any kind of real-time or more modern communication, a little bit IoT-ish, shout out to all the IoT thinking over these years, something should have said, the traffic light or the cop's badge or the other car should have said, "Hey man, 100 yards, everybody stay away. Everybody's at risk here." If you could read that kind of situation. So soft robotics versus hard robots get us closer to that. We see it all coming at whatever time period. We're just trying to build in advance to be able to facilitate all this stuff.

Rich Robinson:

That's beautiful. I love that you bring that up because I think most of the time, the most interesting things are the convergence of technology and services that come together that we can't necessarily conceive of right now. I think smart cities are going to be a completely smart ecosystem, moving faster and coordinating more deeply than I think we can even imagine.

Thank you, two smart cookies, talking about smart cities, the machine economy, the robot revolution, and everything in between. Amar Bedi and Jay McCarthy from Tashi. Gentlemen, thanks so much, and hope to see you soon at the Network School. To be continued.

Jay McCarthy:

Thank you, Rich. Thanks for the opportunity.

Amar Bedi:

Thank you, sir.

Rich Robinson:

All the best.