This Week in Startups - Solana Founder Anatoly Yakovenko on scaling the blockchain to warp speed | E1302

Episode Date: October 12, 2021

Anatoly from Solana joins Jason to discuss how they've managed to build the fastest blockchain (0:54). Solana's "proof of history" innovation allows for over 50,000 transactions a second, we get into Solana's value prop (25:51), use cases (40:14), developer ecosystem and more!

This week in startups is brought to you by Silicon Valley Bank. For over 35 years, Silicon Valley Bank has been providing banking and financial solutions for every stage of the startup journey. Learn more at SVB.com slash twist. Our crowd helps you invest early in pre-IPO companies alongside professional VCs. If you're interested in investing, you can join Our Crowd for free at O-U-R-C-R-O-W-D.com slash twist. And marketer hire. Need expert marketing help fast, hire vetted marketing specialists this week

from the company that's already used by Netflix, Allbirds, and more. Get $500 off your first hire at marketerhire.com slash twist. And use code, twist. All right, everybody. You asked for him. Here he is. Anatoly Yakovenko Anatoly, did I get your name correct?

Did I butcher it too bad? Nailed it. Perfect, great. So Anatoly is the CEO of a little crypto project known as Solana, S-L-A-N-A. And it's caught fire. People cannot stop talking about it at my poker game. Everybody is really excited about the potential.

And as a proxy for that, the price of Solana has gone up 90x, not 90%, 90 times since January 1st of 2020 from a buck and change, maybe $2 to $166 today, putting them as the six or seventh largest crypto project, depending on the day, blowing past Doge and XRP, and right behind the Ethereums and Bitcoin's and, and, Cardona and Tathers of the world. Tether are obviously a stable coin. We'll see how long that sticks around. Welcome to the program, Anatoly. Awesome to be here. So you've got a long history in the technology business. You worked at Dropbox for a bit, working on their distributed systems, and you were an engineer at Qualcomm, and you had your own VoIP startup at the turn of the millennium. when did you first find out about crypto and then how did you come up with the idea for Solana

and what is Solana's mission? A couple questions there. Take them as you like. Yeah. So early days for me and crypto were when BTC was just starting out and I thought it was kind of silly and it was the only way I was looking at it is well, I could probably optimize the mining code and see if I can mine these things faster than anyone else.

but things started moving so quickly by the time I thought well maybe I should do GPUs somebody was mining it, the GPUs already but the time I thought maybe I could roll an FPGA there was an ASIC already for sale people were running a Kickstarter

for sale and my true experience with crypto was at Kickstarter I forget the name of the company but they basically built the system but then mined BTC with it and didn't release the actual hardware to their users until they exhausted their advantage.

And to me, that was like the quintessential crypto experience is don't trust anybody. But I was an engineer and I looked at it strictly from kind of an engineering perspective and didn't really understand the monetary or social implications of it for a decade, I would say. Yeah. So I think at the core of so long. Vana's value proposition is that the transaction fees and the block times are going to be extremely low. Right now, I believe that your cost per transaction is, my gosh, I don't even, it's like a thousandth

of a penny or something crazy like that. Explain to people who are neophytes what this means in plain English that the block times and the transaction costs are so low, especially when compared to, say, making a transaction with Bitcoin or Ethereum? Yeah, Bitcoin can do like, I think, seven transactions per second theoretically. Ethereum, I think, it depends on the gas capacity, but it's somewhere between 11 and 20, maybe 30 at most. But with Solana, you're not limited by block space.

There isn't like a limit that the network sets. It's purely the bandwidth available to the network. and bandwidth is plentiful. It's this thing that you can get, you know, for 100 bucks at home, you can get one gigabit of it in a lot of places. Yeah. And at a data center, when you sign up and put your box there, they give you one gigabit for free.

Because they really only care about when you go over that and you're trying to get 10 gigs or 20 gigs of bandwidth. So a network with one gigabit available to it globally should be able to fit, you know, half a million of these things per second. if the software doesn't suck. So since it takes four to ten years to build one of these generic, multi-purpose, database, operating system, any kind of complex system like this,

or four years in, I think we're like a factor at 10 off that theoretical. So for people who don't understand these concepts, maybe you could help us with Solana being a proof of state network versus, say, Bitcoin's proof of work network? What does this mean for folks? So there's like a, how do you know when you're talking to somebody on the internet

that it's not a bot, right? That it's not just like a bunch of troll farms from the old country or whatever that are all pretending. Hey, that's why I was bored. They're pretending to be somebody else. So in proof of work in Bitcoin, the way that they do this is you have to prove that you burned electricity somewhere.

And that's how you prove that you are voting with something. And because electricity costs money, it's assumed that it's hard to get a lot of it. And that's where the security of the network comes from. In proof of stake, you're kind of trying to fake this by creating a virtual token, which ends up being the underlying token

that you use for fees and everything else, that is used to prove that you're not faking these votes. And you do that with a signature from, your cryptographic key. And long story short, you can think of it as no different than Veracine, this long, you know, certificate chaining that you see in old school web. That proves that you're talking to Google. Well, imagine if Google, instead of having one vote per certificate chain that had a billion

votes that they could spread out in different amounts to everybody else, that's part of their Google certificate chain. And all of a sudden, you can say, well, the two-thirds of Google agree on that. this thing. Got it. And that's really where proof of state comes from. It's just a way to assign weights to a public key. So when the Solana network was released, did people mine the coins or were they just issued

in the world? They were issued much like Veracine issues, a bunch of stuff, like a bunch of certificates. So in that initial Genesis block, you had a bunch of validators, a bunch of early investors, and, you know, foundation and labs and stuff like that. And once the network got running, part of building these totally open permissionless networks is socializing the cost of the operations. So who pays for the bandwidth?

Who pays for the hardware? And spreading out that cost is you bake in some crypto-economics is what they call it, but literally it's spam resistance. We have this token that represents a weight in the network. You want to make sure that people don't just flood the network with messages. They don't create too many validators that vote. And that token, its capacity, represents the limits of how far you can push the system. And also participating in that system gives you that

token with the hope that it's worth something for somebody to spam this and therefore this thing is value. So how many nodes, how many computers, I guess, would be a way to say it for folks out there. How many of those exist in the network? And then what does it cost to do that? And then what are they gained by being part of the network? So, on the network right now, the Solana main net beta,

there is, if I can sell correctly, I can look check right now. It's about like 1100 voting validators. Okay. On the test net is like 2,200. Who are those validators? What is their motivation? How does one become one?

So a bunch of these are volunteers, so to speak. They're people that want to earn tokens. So being a validator means you're running a box somewhere and somebody assigns you steak or you self-assign yourself some stake. And those are the tokens that represents the weight of that very signed certificate, that's a hypothetical certificate. And when you vote, you get rewards. And the more reward to accumulate, the more tokens you have.

This week in startups is brought to you by Silicon Valley Bank. What's next? What if? Are we ready? Now what? These are the questions that can keep founders up at night. And no one understands this quite like Silicon Valley Bank. For over 35 years, Silicon Valley Bank has held thousands of high-growth companies by providing scalable financial solutions, along with insights and expertise that many other banks just can't. From health care to hardware, software to infrastructure, SBB works with, with companies across the innovation landscape at all stages of the journey,

anticipating their needs even before they do. And by providing access to insights and in-depth reports, SVB can help you make more informed decisions and assist in turning your great idea into a great business. Which could be why 50% of U.S.-based, venture-back tech and life science companies bank with SVB. Learn more at SVB.com slash twist, Silicon Valley Bank, built for what's next.

If an individual wanted to put up one of these servers, they would just go to AWS, install the Solana pack, and then buy a bunch of Solana to then power it. What would it cost them to participate? Yeah. So usually you don't want to use AWS for cloud providers because AWS is very egregious cost for bandwidth. It's like 100X more expensive.

than going to like Hurricane Electric across the bay or whatever. Got it. So you want to co-locate it somewhere? Yeah. Yeah. And then it's like the cost of a box at some place, which is global average is 500 bucks.

Hudson will do it for 200 or less. So once you're up and running, then you need to go find some stake. Does it mean to find some stake? So either you buy a bunch of tokens and you stake yourself. But you need like a big, enough chunk to where you're profitable from the operations. And what we've seen right now is it's some around a million dollars with a soul. And that's largely because of the run-up

in price. You know, when we first started calibrating this, soul was much, much lower and you didn't, you know, basically divide that by like a factor at 20 or 100. So it didn't, it wasn't as like egregious. So part of the issue for Solana is that the way the network works is that when you vote, those votes are transactions. They have to pay fees in the network. And those fees are up because you're voting every four in a millisecond. So even if it's like one one thousandth of a penny to vote, you're constantly voting that

cost so you need to actually have enough soul to where your proportion of the rewards from the network cover that cost. and some of that is because of that early issue where we're just not as efficiently using bandwidth as we should be but like a factor at town. So as the software gets better, this cost will drop as well. Is this dynamic nature of commerce

and basically creating a monetary supply or the tokens, whatever we, however we want to frame those, you have this speculation that can occur, it's decentralized, people can make their bet, if they want to bet on any particular cryptocurrency. And then you have the reality of running a network.

You probably didn't anticipate you would go 100x this year. So is this now become like a problem in building out your network that it's so expensive to start a server? Starting a server is cheap, right? Right, but most staking a server, I should say. And again, that staking is cheap in the sense that the same number of tokens, now is a year ago minus inflation rewards, right? Those keep adding tokens.

But literally the same pool of people or a larger, you know, a larger pool of people, same pool of tokens exist. So when you start up a validator, there's a community of stake pools and what have you that you can participate and get enough stake, either from those pools or from the foundation, if you want to also run a Tassna node and Strasst. And like, you can kind of like duct tape it together.

So explain how these projects are architected. I know that you raised some money for the company or the project. I actually know some friends who were involved in investing in it. And there was a small venture fund that a couple of my friends were in. I guess that was one of the seed funds in Solana. I'm not sure if I'm supposed to say who. But anyway, that did really well. And now they're all swimming in Solana.

But when people invested in the company, did they buy tokens? Did they invest in the corporate? Do they do both? And now you have this foundation. How does this all architect itself? Because this is a new concept that's different than venture, correct? It's like a trial by fire. I wouldn't call it an architecture.

It was just the best we could do, given the resources we had in a really fast-moving environment. And I think we did pretty well mostly because of folks like Cosmos, where we followed a lot of the playbook that set out. So the idea is that you really want to decentralize network and decentralization really means truly independent operators of the network that can control around the nodes and maximize the cost to destroy it. Literally, the network dies if I can go and wipe out every copy of the ledger. That's what it means to kill a network. So making that as expensive as possible requires many jurisdictions, many independent providers,

many data centers, many humans, running these things. And the effective playbook is to do this via foundation. Switzerland is like a nice place to start it. The Zerg? Yeah, they have pretty clear laws on digital assets. And to still, you're still constrained by US laws. You still can't just like dump tokens on

on U.S. persons. So you still have to go through the, you know, exempt kind of style raises that you do normally for a venture back company, mostly dealing with U.S. credit investors and so forth. So let's talk about what people are building on top of Solana because it is, am I correct, that this is essentially like an operating system

where people can build apps on top of it? And how does that manifest itself? Is this a specific code base? How does one build an application? How sophisticated is that application language, if you were to compare it to, say, building, I don't know, a website or building an app on iOS? So this kind of really speaks in my background.

I spent most of my career at Qualcomm. First big project I was on was Brew, which is kind of the first mobile platform OS. So the environment for Solana is, you build an elf executable object, right? That targets Berkeley Packet Filter, which is a bytecode that's got really high security guarantees with LLVM, which is a standard compiler tool chain

that if you've been coding, you've heard of it. And you can use any front-end language, so C, C++, Rust, and we have a C-Rust that we built to help you target the thing, and we prefer Rust. So it's kind of like building for an embedded system, and you should think of it as an embedded system because it is one shared giant computer for the entire world. And the world is big, right? So if you don't treat it as an embedded system, it becomes very expensive to be able to host arbitrary applications because you just can't use a lot of arbitrary resources.

Would somebody be able to build an application equivalent to, I'll just give two examples here, and we'll walk through it. let's say Twitter, a basic social network, micro blogging platform, or Spotify on Solana currently. And when they did, does that mean when they deploy it, they have zero server costs? They're just deploying to the network or they're paying. So when somebody would resolve to run Spotify or to run Twitter, they would be running on the Solana network.

Is that the concept here? So it's deceptful. If I said yes, it's possible, but I'd be lying. because what the network is scaling is rights. It's simply scaling writing, writing the state and confirming that these rights happened. Got it. The database component.

All the actual hard work in Twitter is on scaling the reads, right? Twitter has like 5,000 tweets per second. We can handle 5,000 transactions per second. Not a big deal. But how do you globally scale 100 million concurrent reads and structure this data? It's the same work that Twitter does. It's a giant, yeah.

What do they call those non-SQL databases, I guess, that let you pull up pre-constructed lists of your feed, and that is another type of technical challenge. So what we would be saying here in the Twitter example is, yes, your rights to your feed could be handled by the Solana Network, but the reads would have to be done in some other place, correct? Correct. And then if you did deploy in this fashion, you basically no longer require a database at, you know, AWS or somewhere, you're relying on the Salana network to be your giant database to confirm all of these tweets in this example. But you would still have to figure out a way to have people read it and construct those feeds, which you would do through a normal whatever lamp stack or whatever you used. Yep, exactly. And then the people running the network would get paid a fraction of a penny every time you wrote a tweet to the network. And you could socialize it from like your single application server that kind of feeds, takes in the funnel of human tweets and then pays for them, right?

And you can send them ads instead. Or you could use a truly self-custody decentralized way where your user expected to have a wallet, they have some soul in it, and they just automatically. pay every time they tweet. This would be a way to mitigate against spam and deal with identity on these networks because if you were running a bot farm, you would all of a sudden

have a cost to everything you publish to the network. And so if Twitter wanted to get rid of spam and bots, if they charged a penny every time you created an account and, you know, it was a hundredth of a penny for your first tweet of the day,

a tenth of a penny for your second, and a penny for your 10th, whatever it is, you could actually model out a de minimis amount of cost like something people would never even notice a dollar a year maybe, a dollar every 10 years, but it would be enough to kill all the bots, correct? Correct.

That is exactly civil resistance, the whole proof of work reason. The reason for proof of work to exist is to kill bots. It's time for another R crowd deal of the week. Right now you can join R Crowds, investment in Orient. According to the deal, Malmelt, Orients, software-only indoor GPS is 20 times more accurate and scalable than current solutions. And they've landed contracts with some of the largest retailers of the world. So why join Our Crowd, you ask? Well, Our Crowd investors were in early

on some of the best performing IPOs of 2019 and 2020, and now you can join them. With Our Crowd, accredited investors can invest directly easily and most importantly early. Our Crowd investors have benefited from companies IPOing like Beyond Me, or being bought by companies like Intel, Nike, Microsoft, Oracle, and Uber. So here's your old CTA, the call to action. Our Crowd's accredited investors have already invested over $1 billion in growing tech companies. If you're an accredited investor, you can join Our Crowd for free at OUR, CROWD.com slash twist and review the current deals. There is no payment involved until you decide to invest. That's OurCrowd, com slash twist to sign up for free.

Now, let's take a more robust example, like a Spotify, where you have a ton of media that needs to be stored somewhere. How do you think that works in a decentralized environment? Obviously, you could keep the database of all the songs if people want to publish songs, but there's no way to store the files on the Solana network, correct? Yeah, the files would be somewhat expensive. Salon is really designed for like the metadata layer, the execution, the relationship between tokens, who owns what. And Audius is basically that thing that you're talking about.

Yes, we had them on episode 1273. So explain what Audius built on top of Solana and why people are excited about it. I am, so first of all, it is like this true multi-chain project. The governance token is on Ethereum. It's bridged over to Solana. On Solana, they have smart contracts that track artists and users that follow that artists. and the music likes and music follows and plays and all this metadata, the relationship between artists and their fans. And I don't know if this is up and running yet, but their goal is to have the audio content to be running on R-Weave,

which is a persistent storage-optimized network. Once you post it on R-Weave, it's there forever. So you kind of pay once. Oh, wow. And that's kind of like, you know, you can think of it like two decentralized computers, right? One is your disc, one is your like, I think of Solana's a GPU. What is it called? It's R-Wave?

R-Wave. R-Weave. A-R-Weed. So we had an explosive launch of NFTs on Solana, partly because Solana built this auction mechanism, everything's on chain, but by default, we had all the content on R-Wave. And Sam, the founder, was like, staying up at night, figuring out how to decentralize and scale their CDN as like people

are minting, you know, 10,000 NFTs an hour. So we had a quite a fun like web two in the trenches. Things are scaling faster than we can keep up experience. Yeah, so let's talk about NFTs as well.

But I wanted to stop on you said multi-chain solution. What does it mean to be a multi-chain solution? We understand that means use Ethereum and so on, but what does it practically mean? And what's the benefit of being multi-chain? So it's that you're thinking of the, I think these things are backend technologies.

And they're not the web client. There's not like a single thing with 3,000, or I mean, 3 billion users that are all of the same virtual machine running locally. So these are back-end systems. And with back-ends, you can pick and choose whatever you need for whatever solution you want. And the thing that really glues all this stuff together is the cryptography. I can have cryptographic proofs that I own something in Ethereum.

I can move that stuff over to Solana through a bridge and vice versa. All this can be automated through cryptography. And the users shouldn't really care at the end of the day where this information is synchronized or managed. And so you can kind of pick the system that you want for the use case that you care about. And that's kind of how I think of the multi-chain world, is that this stuff is complicated, just like a database,

there's perroto efficient tradeoffs. And when you end up with perrote efficient tradeoffs, somebody's going to maximize one aspect of it, and somebody else is going to do something, maximize another. So Solana picked a very clear different perado efficient spot than Ethereum or Ethereum 2 or Bitcoin or any, you know, any of the even competitors like near Avalanche, etc.

Which is speed and low transaction cost? it is so that is a side effect to the proto efficient spot we picked we picked um a single shard so a single network that fully replicates information as fast as possible so this this again speaks to my 12 years of qualcomm what how i look at this as a unconsorable broadcast radio how do i make sure that when you transmit a bit it's propagated everywhere around the world as fast as possible and that's guaranteed what is that going to look like if the adoption right now you have how many parties and how many transactions occurring that are real transactions a day? Are we talking about billions, millions? Like 40 million a day. But a lot of that is coming from like a central limit order book called zero. So market makers, all they do is some cancel orders.

Got it. Cancel the sword. So let's say if the network went a thousand, thousand X and all of a sudden, you know, we're not 40 million. You know, you're at 400 billion and then 4 trillion. What would the infrastructure need to be in terms of the storage, the bandwidth, etc? And, you know, because that's eventually where this is heading is that these things are going to hit scale. Would that look like, you know, the footprint of Google? Would it look like the footprint of Facebook's data centers or would it look like something different? So 1,000x in humans?

thousand X in transactions maybe to start you know or well so the thing is like Twitter right let's say we went from our current humans you know which is maybe a million like multi-active humans maybe half a million to Twitter's 200 million multi-active users we're going from you know like 30 transactions per second that are human to 5,000 that's not a big deal

but the interest the read infrastructure to support structuring all this data displaying to them, that's going to have the same footprint as Twitter. There's no magic there. But the chain itself, like the piece that we're working on, that is scalable today to that level easily. What happened on September 17th? I know you guys had a really rough day.

There was a blackout for, I don't know, 12 hours. 17 hours. 17 hours. 17 hour block. Don't want to bring up any PTSD. I know this was just two weeks ago. You probably didn't sleep. But, you know, this was a, you know,

like this is what happens when you get popular. We all remember at the early days of Twitter, the farewell. It was like, you know, Twitter had an existential crisis here. Why did you go down? What did you learn from it? And how do you prevent that in the future? So this was really not the first time we went down. It was kind of the first public one.

The first time people noticed because nobody knew who you were previously. Exactly. So the thing is that like what is decentralization? It's to say it's in my mind, I think of everything objectively. it's the cost to destroy every copy of the state. So if the network goes down, what it means is that it stopped confirming blocks, but the ledger in the state is still there. So there's multiple reasons why it could stop confirming blocks.

Everywhere from like every government in the world orders their telecoms to start dropping your data and your kind of host to major AWS one down. And for some reason, you had more than a third of the stake in AWS and no longer. you can no longer get to super majority. A lot of networks are susceptible to this. Or in our case, there was a radium launch for a product, an IDO. Kind of think of it like a Binance Exchange launch or something like this. And they're already bots optimized to try to snipe orders.

And they got pretty good at their optimizations. And they flooded the network with about 400,000 at peak request to make a transaction, not actual transactions and blocks. But they basically flooded it. until the network could not handle that load and started running out of memory. Got it. So that would be like a denial of service attack

or a memory overflow kind of attack, which has been solved in the DNS space, you know, through Dynne and Cloudflare and others. So I assume you've figured out some way to if, I don't know, I say a bad actor, but, you know, if somebody does something outrageous, is it more resilient to the outrageous moments like this? were edge cases?

Quality of service is not a, it's not like computer science problem. It's just a pain in the ass problem. And, you know, we test for this. We flood the network all the time. There's just a couple corner cases. So when I get blame that code, it was me who wrote it two years ago, where I had a couple unbounded cues that were growing. So can only get blame myself.

But this is part of why we still want to call it Maynod Beta, is that we have extremely high confidence in safety. You only need one of these replicas to survive, one of these replicas to be valid. And the network is already, as censorship is, like, decentralized from that physical perspective, in my mind, as Ethereum 2 or Ethereum.

There isn't that significant difference between 1,000 replicas and 5,000 and Ethereum BTC. Like, we're at that, we're like a few more, like, you know, jumps behind, but we're basically pretty much there. But liveliness and like resilience to arbitrary attack vectors while guaranteeing low cost, low fees to all the actual users, that's a probably will take a couple iterations

and probably will take some iterations from now until we get to 100 million users, right? And then from 100 million users to a billion. Don't you wish you could hire a ringer to help scale out your marketing team? I know you do. Well, with marketer hire, you can do it right now. Marketer hire gives you access to expert freelancers on demand with no long-term contracts and no risk.

You can hire experienced specialists across the most valuable marketing disciplines from paid social and search to growth, SEO content, and even get a fractional CMO. If you don't need a full-time one, maybe a fractional one would work great for you. Again, no long-term contracts. You can cancel at any time. and if it's your first time working with freelance talent, you'll start with a no-risk trial. Only hire what you need and stay on budget with hourly, part-time, full-time arrangements.

Every freelancer on marketer hire goes through a rigorous vetting process with industry experts. Freelancers for marketer hire have been hired at over 1,500 companies, including top brands like Netflix, Allbirds, and Lambda School. You're going to get $500 off your first hire at marketerhire.com slash twist. Again, marketer, M-A-R-K-E-T-E-R-H-I-R-E dot com slash twist.

Marketorhire.com slash twist. Get that $5-Hundee right now, and you'll also get a free consultation on who to hire based on your needs and your goals. Again, 500 right now, marketer-hire.com slash twist. Is this Solana project, it's an open source project, the codebase? Yep. How many people contributed to that codebase in September? In September? Ballpark. I'd say maybe 20, 20 people like churning through the code.

So there are 20 people. How many of them work for you or the foundations? I would say vast majority, probably 15. 15 of them. So there's five outside contributors. And the reason I ask this is the project's worth 50 billion, which doesn't mean that there's $50 billion sitting somewhere. You sold the tokens when they were worth pennies, I assume.

So you were able to raise some amount for your foundation, 100 million, 200 million? What was the initial sale? What did that generate? Up until the launch, we were, we never had more than 20 months of runway. And our total raises through the years added up to like $25 million. Okay. So you raised $25. That was stuff.

Go ahead. Yeah. So that was a year and a half ago. And then due to the growth in the ecosystem and kind of the developer adoption, the labs did effectively erase, but it sold tokens to investors that want a chunk of the ecosystem. Got it. And that included like A16C, polychain. How many tokens were sold in that?

It was to $314 million worth. So pie times $100 million, exactly. Exactly. So you take the pie, the big pie, and you put that in a foundation. Now the foundation has essentially unlimited resources to pay, hire developers and to stake people who need tokens to put up servers, correct? Is that the concept behind the foundation? Or what happens to all that money?

Foundation is a separate entity. What it's focusing on is censorship resistance. So it's got a bunch of tokens left over. And it is 100% focusing on bringing on more validators, more physical layer participants. You can think of it. We actually have to build the physical network. otherwise there's nothing there. It's just a bunch of software.

Yeah, exactly. And then the next time this person floods the network with a bunch of orders, so you could go down again. So the $314 million then goes into the company or the labs. And then what do you do with that $314 million? Where does that go? That is paying for engineers to do engineering. And that's kind of all over the place because, you know,

the kernel itself, that's just a small part, right? Like building out the actual Linux kernel is a very tiny part compared to Android or even Ubuntu, right, or RedHap. There's a whole lot of other work that needs to be done at the application level, at the just developer education, like how the how people even know that they need to do this. And now these are not shares in the company, to be clear. These are utility tokens people are buying because they want to have access to those tokens, correct?

Correct. So does that mean when you sell those as opposed to an investment when you sell shares, you don't pay tax on it? You have to pay tax on selling those utilities? Yeah. So the government gets like some huge tax benefit from you doing this as opposed to you selling shares in the company? Correct. Which would be 30, 40% of the proceeds or something? The US is pretty good at capital, like, taxing corporation.

So I think it's closer to 20 or something. So this is an interesting point. You know, you look at it and there's a debate. Is it a security? Is it a utility token? You've made your decision. This is a utility token. And people are actually using them to do NFTs and this kind of stuff.

So you have an up and running network. Seems like it's pretty safe to say. And then, you know, the government's concern, I think, at least here in the United States, is are people paying their taxes? And here, you're writing a big check to the government. You wrote like a $60 or $70 million check as opposed to raising money to just sell shares, correct? Correct. But, you know, this is just the way that U.S. laws are written, right? We have to comply by every regulation. So even when we do the sale, we're compliant with the way the sale is run as if it is a security. But we still pay the IRS because the IRS doesn't give a shit about. So you're getting the worst of, wait, hold a second. You're getting the worst of both worlds. You're making sure people are accredited so you're covered.

Correct. And you know your customer and you're paying the taxes. You chose to do the worst of both to cover yourself. So you took the most conservative position. Of course. So what else are you going to do? So you probably heard Gensler's sort of discussion about crypto and how he believes people are buying these things for appreciation and that makes them stocks.

What do you think the right framework for the United States is I know you're based in this erg, so you have some protections there. Foundation is in the Zerri. Salana Labs, myself, and the labs executives are in the U.S. The Delaware C Corp. Okay. So you're a legitimate American company.

So what do you think the right approach is? Because let's face it, the ICOs and all these Bitcoin's and stuff like that really poisoned the well for real projects like yours, real companies that are playing by the rules that are doing accredited investors that are paying their taxes, etc.

So you're kind of standing on this mud of all these ICOs and you know, detrius and the and the and the and all the broken glass. Now you've got to try to rebuild this in a legitimate way. What do you think the SEC and the, the, the, the, the regulators approach should be to not stifle this incredible innovation while protecting consumers. What's reasonable in your mind? What's reasonable, I think is some, I mean, a lot of the, forget the person's name who wrote some of the stuff, even on GitHub from the

SEC, just creating some safe harbors for a project to be able to go from open source, open built project in those few years to go from nothing to decentralized protocol. And having some

former safe harbors there. That's really all that's necessary, I think. Because, like, reality is that there's some projects that have management and CEOs, and then there's projects that are piece of software. Right.

So, Solana, the protocol and the tokens, it's a piece of software. All the token does is spam resistance. There isn't like a network CEO that makes decisions for which spam to prevent, which not. Right. Right. Like the physical network is run by other parties. It is just a giant piece of code, right? That's it.

Are you, um, concern that the price went up so dramatically. Would you rather it be at like 10 bucks and you know, not, you know, at this early stage have grown so quickly? Listen, I'm sure it's great for your personal finances or your investors. But it also is a little bit like, okay, wait a second. Now we got to catch up to this. I mean, 50 billion as a market cap is, I mean, you're basically like Uber or Airbnb or DoorDash now. So, you know, and without the, you know, the scale yet or the customer base. Well, you got to compare to like, the crypto customer base, right?

Got it. So if you think of Ethereum with 10 million multi-active users and Salonat 1, 1 million monthly-a- So each active user is worth $50,000 right now. In crypto, you would say that. Wow. Because it's a different market, right? Like where else have you seen digital marketplace to get to a billion dollars in value in two months? Yeah, are you talking about NFTs, I guess?

Right. So, like, there's opportunities in crypto that are much, much bigger than anywhere else in the world. So people, yeah, are people taking the NFTs that they forged on Ethereum and then moving them over to Solana because it's cheaper? Is that what's happening? These are natively minted on Solano. Natively mined on Solana. And so are the NFT, is the NFT crowd saying it's too expensive on Ethereum? Let's do it on Solana because it's going to, what does it cost to trade a million dollar?

let's take one of the crypto punks or something like a million dollar NFT, what would it cost to trade the million dollar NFT on Salana versus Ethereum. The fees to trade, like while Ethereum gas is expensive, like $100 when you're trading a million dollar NFT, you really care.

Yeah. On a $10 NFT, now you get a problem. Yeah, so that is the thing that is blocked entry-level artists to really jump into the space. and what we've seen is like on Solana there's this project called Metaplex that mince you know a 10,000 set for a $1.66 worth of R weave and sole combined

mostly R weave to store the images. So like you meant 10,000 NFTs and you sell them and the artist makes like you know a hundredk to a million dollars depending on how successful they are and that would not have been possible for them as an independent artist that has never seen more than then like 30K a year and like their income, right, to be able to do that on Ethereum. It is kind of crazy. What, at its core, what is happening with NFTs? Is it that there are a bunch of crypto millionaires who this is just a fun pastime?

It's like their version of poker or their version of gambling. Or is this actually the art community has now recognized this as a, you know, the future of art. And that's why these things are getting priced at $200,000 or a million. you know, people we understand as an existing artist, but it doesn't really make sense that 10,000, you know. I don't know. You don't know. Okay, good.

I love that answer. I mean, yeah, we're all sitting here trying to figure this out. What is actually happening that people would pay $200,000 for? Why is Ethereum worth what it's worth? Why is Bitcoin worth $50K? Yeah, that's a great question. Yeah. If you, if.

I mean, most people would say limited supply and too much money in the world and speculation. And people are speculating and one ahead. you know, I guess against the dollar or their native currency? So for whatever reason, if Bitcoin is worth a trillion dollars, it's some cryptographic system, right, that allows global settlement of a trillion dollars with the funds. Everything that's being built is built on top of that. Yeah.

It is an interesting moment in time. I had Bologi on the All In podcast, and he was talking about, I don't know if you heard the episode 48 with him, He was talking a little bit about using blockchains to store facts and news, and maybe this is some way to find the truth, et cetera. Of course, whatever facts are put on the blockchain could be wrong. But this idea of permanence and immutability, is it really immutable and permanent?

And what if people put things on a blockchain? Because I can tell you're strongly in favor of, you know, stopping censorship. What if people put things on a block? What do people put things on a blockchain that should in fact be censored? Like, I don't know, your phone number or doxing you or, you know, horrible things that shouldn't be shared? So that read layer, right, is where like that really, you know, there's a bunch of bits somewhere on some hard drive. You connect to a blockchain, you get those bits, and you have to reassemble them into something that humans understand. Sure.

Right. And then you have to have a domain. name that says, hey, I claim that the stuff that I've reassembled is actual USDC, actual dollars. And if you use my service, then I am handling US dollars for you. I'm claiming that like all this stuff is going to work. And this is where that regulation and like local jurisdictional authority, I think, plays. This is where you have DNS, you have, you're running a business, you're getting users, you're telling them that this stuff is true.

To me, that's really kind of not that much different than connecting to the internet. If you really think about it, what does the layer one do? Especially one like Salon. So you're okay with the information beyond the blockchain,

but then you could block the reading of it. So it's there, but not accessible or hard to access. What's the difference between that and it being on the internet? Well, if it was on the internet, it has to reside on a server somewhere. the feds can come or whatever legal organization can come and take that server off the internet physically.

If it's distributed on thousands of servers, they cannot. If it's on the blockchain, they cannot because they would have to take the whole network down. So what if it's sure, but like there's plenty of places where you can encode it. Like I encode every bit of this bad data in a separate tweet. Yeah. I mean, I guess the question would be the scale of it. I guess, let me ask it another way, you are always thinking about the edge cases as this master architect and it's truly impressive what you've built. If somebody were to do a massive leak of

everybody's social security number, email, password, phone number, and put it on Salana and on the network and wrote it and you had 10 million people's, you know, most important information compromise, their password to their bank account, their phone number, et cetera, and it could lead to chaos. What do you do? Well, what will you do? Because at some point this will happen. Some human has to reassemble that data into something that displays as words and images, right?

Okay. Right. So that reassembly is the, you can think of it as the key. We could just as easily encode the stuff in some encrypted way, right, with a key that you can go and, like, acquire because it's a publicly known key, right? just some silly thing that the data scrambled on the servers. It doesn't actually mean it's not there, right? Somebody still has to go through the steps, reassembling it, and how easy that is,

is really the threshold is this like a fire? Or do they just go after that one person that's sharing that information? Yeah, so what if the person said, I hear, because we assume that wallets and access, you know, people being to access the blockchain is going to get easier and easier, right? It's getting, you know, Twitter will have wallets built into it. Your Facebook account will. be a wallet eventually.

So backing that data up on BitTorrent, sharing those like links, torrent links on Pirate Bay. If that data is available there, what does it matter if there's some way to reassemble it from some set of bits on chain? Fair enough. So since there are other ways to do it currently, this new way of doing it is no more harmful in your mind. I think that censorship, that actual like government enforcement of the old finding these folks and shutting them down, that is the layer that is kind of behind because we don't have

deep understanding how technology works outside of engineers, like outside of our small niche of people. There isn't like a CTO at every like police station or something like that. For sure, for sure. I mean, it just is interesting to think about the edge cases and, you know, the stuff will eventually happen. and so how do we want to defend against it? And, you know, it's...

Like, the validators could vote to just go kill that part of the Merkel tree. Yeah, I guess that becomes where it becomes pragmatic, because you might have somebody do something where all the validators or the majority of them say, we really don't want to have this out there. We're going to take it down. And I guess they could vote to do that, yeah? And reverse it.

So this idea that it's immutable, maybe... It's immutable unless all the server hosts decide they want it to not be immutable. But if storing the history, if somebody's somewhere storing that history, yes. Is it, I mean, no, I mean, if you take down a copy of the pirated movie, and a thousand people have downloaded already, it's, it already exists, you know, so, I think it's always interesting. I mean, I don't want to seem like a boomer, but, you know, these are things that we have

dealt with, like even with Pirate Bay BitTorrent, like, yes, it's impossible to stop, but you have, you do have government agencies. So do you worry that government agencies are going to look at this and just take a sledgehammer approach if this kind of stuff starts happening and say, you know what, we just, this is too much trouble. We've got to take it down. They don't come after like AT&T switches for the intermediate data stored in the buffers, right? No. And they don't. Certainly if people were planning some sort of horrible attack or trading illegal things

over the phone talking to each other, they would not turn down the AT&T network because of it. And if somebody robs a bank with a Volvo and drives away in the getaway car, they do not blame the Volvo and they blame the criminal. So I think that's largely direction correct. That's going to

be a while, right, for that to propagate. Like, for that like a human mental model to get to that level. So, you know, to me, I'm just as uncertain about the stuff as you. I don't know what's going to happen, right? Well, I mean, it's great to have an informed conversation with somebody who's building

the future and I appreciate you coming on the show to have that open discussion because we do blame technologies disproportionately for things that occur in the real world. As but one example, when Airbnb came out, everybody said, oh, my God, what if a bunch of meth heads go to your house and rent it and destroy it? And it was like, by the way, that's happening in 10,000 hotel rooms at this very moment. there's a meth party in 10,000 hotel rooms and rock stars are destroying and throwing televisions like this is happening in every hotel. It happens every day in probably every hotel a room gets trashed. Why would Airbnb be any different? The hosts understand they're taking

some level of risk for a certain amount of reward and there's insurance. The end. We're done. But when that happened, you probably remember when that woman had her apartment trashed by a drug party, you know, the first thing Airbnb did was they just gave her $50 or $100,000 and she got a brand new apartment. And it was like, okay, sorry for the trauma, but. So the challenge here is there isn't like anybody to really do that, right, to like yell at or to help you recover. Yes.

All of like Labs does is it publishes open source code to GitHub. So the only power we have is just a bunch of words, really. Yeah. And the government has the ability to take measures to find criminals. And so people are doing criminal things and they're using various tools to do them, the government has other ways to find them. What's the most challenging thing about the job right now for you, Anatoly? Infinite Zooms. Just talking to developers on Zooms?

Or is it a bunch of investors who are trying to glom on now and get a little bit of the pixie dust of the $50 billion market cap? I've been able to hand off the investor stuff to underlings. You're post needing any funding, right? You're done. You'll never have to fund this project again. and you could always just issue, I don't know, 1% more tokens? Is that correct?

You can't. You'd have to get votes to do that? Well, like the network. It's a fixed supply. Yeah, the network has the control over the issuance. We don't. All we do is publish source code to GitHub. And so that network, if they all voted to issue another, to double the amount of supply,

the concept there is if they decide to be money printing machine go burr, and they decide to be like the Fed and print up, you know, a trillion dollars, or let's just say they try to print $50 billion, they would have to sell it to somebody, right? And how would that work if they voted to issue another billion or, you know, 100 million tokens? What would happen? So those bits, right, all they're doing is synchronizing bits. If you're, let's say, in exchange, you personally run an exchange, you're on a node and you're on software that, let's say, labs publishes open source, but there may be different clients in

the future that interprets those bits and continues voting. And you and your exchange that has real users, you tell them that these are soul. And if you give me dollars, there's some other rails, I will give you soul and vice versa. And when you click this button, you can actually get soul on the network. Right? Yep. But you are the one that's actually manifesting those bits into soul. Not the validators, not anyone else. You're running that node yourself and you're saying that these bits

are actual soul and you're the one that's promising that to your customer. Not, right? Like, you're actually, like, responsible for this. The custody of your node, you're doing it. Yeah. So people can create their own currencies underneath Solana?

Yeah. So let's say, like, all the validators, like, double the supply. If you accepted that change, then you're kind of accepting their vote. But if you don't, then you effectively end up running your own fort.

That's really fascinating. So are there enough developers out there to support these projects? How hard is it to get developers to leave what's going on out there in the world in terms of working for Googles and Facebooks and getting paid huge amounts

of money? And then how long does to take to train them to contribute to this project? Because it seems incredibly complex and there's a small number of people who actually understand this distributed computing plus cryptography. Is this just only available to the top one or two percent of developers? I mean, Google, Facebook level, if those folks can all do it, the challenge is getting them to quit their jobs and convincing them that this is important enough. So our hackathons, we have 13,000 registrations in the Salana Summer Hackathon. This ignition one had 5,000 registrations,

but we think higher quality doves. So people are there to start coding. It takes a weekend to learn Rust if you've had a career in like dotnet, Java, C++ or C or any like modern programming language. It's really a question of like, do you want to be a CEO? Writing the code, the smart contract? You might not even have to because there's already enough reference implementations for whatever permutation of lending protocol, Defi, NFT marketplace that you want to do.

Somebody already wrote something that's pretty close to what you want. How do you convince somebody to go be a founder? Yeah, it's fair. Yeah, I mean, it's, if they know how hard it's going to be, you just basically have to lie to them. and just tell them like, it's going to, you know what, there'll be tough days, but you'll get through it. Yeah. And if you actually tell them nine out of ten days are literally putting your hand on the desk and smacking it with a hammer and, you know, or a brick. And it's basically like, yeah, here's a brick.

And every day, flip a coin. And one day, if it's heads, you smack yourself in the head with the brick. And if it tells you don't. That's basically what being a founder is. Be careful what you wish for. because it's just a nonstop headache. The second you solve some fucking problem, another one pops up,

the second you get some great team member, they get 10 offers to go quit the company, somebody has personal problems, some technical problem is insurmountable. Whatever problem, the smartest people you hire can't solve, becomes your problem. That's being the founder.

Yeah. Congratulations. So that's basically like what we're focusing on. And I feel like everyone else, that's not really competing for that, for those people, is not focusing in the right thing. So all the hackathons, all the tooling, all of these podcasts, this is me just like trying to find,

who are these people that are willing to eat glass? Exactly. As Billy would say, yeah, eat glass and look over the abyss. All right, listen, thanks for coming on the pod. You've been awesome, so candid. You're hiring developers and you want people to contribute to the open source, whichever they prefer? Yeah, yeah, or just join a hackathon.

Perfect. Where can they find more? Yeah. Salana.com slash ignition. Okay, great. Or Salana.com. Yeah, just go check it out. Follow me in Twitter. Perfect. All right, follow him on Twitter. He's A-E-Y-A-K-O-V-E-N-K-O. Anatoly, I really appreciate you coming on. Continued success with the project. And we'll see you all next time on this week and startups. Bye-bye.