Stable Fees, Infinite Scale with Matt Sorg, VP of Technology at Solana Foundation

I think a lot of people still have a misunderstanding of blockchain.

Like, the actual thing that gets you product market fit is really just tackling every single 1% gain that you can for your users, which, you know, which are applications and devs.

And Trump, like literally the President of the United States, launched his, um, meme coin, then his wife did.

Like, that's, like, the basically the only time in the last, uh, year and a half where there's been any sort of median fee spike. We have multiple teams now that are doing quantum secure Solana.

Those things are going, uh, uh, underway, and we are very much ahead of the game. Welcome to another episode of 51 Insights, today with Matt Sorg. Matt, welcome to the show. Hey, thanks for having me.

Matt, it's a pleasure to have you here. You're the VP of technology at Solana Foundation. Can you give us an overview of how does that role look like and what do you do day by day? Yeah.

So we have a very large ecosystem in Solana, and the foundation kind of plays the facilitating role with all of the organizations involved.

So foundation is roughly split up in tech and growth, and I lead, um, some of the tech side of the org.

So, uh, there's probably around 10 organizations that commit-- are, are very involved in the, the core protocol itself.

Uh, we have a SIMD process, which is our Solana Improvement Documents, which is kind of similar to the EIP process in Ethereum, and we help run that process and facilitate governance and the developers tools on top.

So anytime runtime changes or just developer tools that applications use to make great apps. So it's like really that whole like end-to-end thing that we kind of facilitate across all those organizations.

So it's, it's very, it's never, never a dull moment.

A lot of different personalities and organizations that have drastically different cultures, which is great because it's very balancing, and then you kind of have this back and forth and kind of annealing to, to, to really good solutions.

I can imagine. And before we jump into more technical aspects, uh, you mentioned that Solana has a very unique technical culture. Can you describe that a little bit more? Yeah. I'm actually wearing this shirt.

Uh, it's, it's Rusconf this y- this, this week in Seattle, uh, and a bunch of us are here. So we have this, like, increased bandwidth, reduced latency shirt.

So, like, the lines get bigger, and then there's actually, like, a lot of little lines in there. Anyway, but that's, like, our culture.

So it's like you have all these different organizations and, uh, groups, and they're kind of just aligned on that, that, that ethos. Uh, we call it IBRL, increased bandwidth, reduced latency.

And there's this misso-nomer that kind of exists in crypto is that once you release your, your protocol and your spec, then it takes some grand change, a 2.0 or some branded upgrade in order to significantly improve the protocol.

Uh, and that's not really the case in Solana.

So Solana is a very efficient spec and, um, while we have some big changes coming down the line, it is much more about just good engineering of constant iteration and constant improvement.

So over the last, you know, four and a half years, you know, five and a half years maybe since, since Mainnet launched, it's been very, very consistent, uh, improvements in all parts of the stack every release.

And our releases are very frequent. You know, they're every few months, not years. So it's-- There's graphs that show this, like things like block times have, have, have decreased.

TPS has increased very steadily over the last years. Uh, l-let's take a step back for someone who's not familiar with all the chains. Solana started, you said, five years ago. Mm-hmm. They've always been about speed.

What do you think makes Solana unique, and why was Solana created? Yeah. We are the first ones to really have performance first in mind, and so, like, every single part of the stack was thought of.

So we didn't necessarily pin ourselves to any system that was out, like a Bitcoin or Ethereum or any, any other ones. It was from the ground up.

If you want to create a new layer of the internet that passes around information where everybody agrees on that, so, so typically very heavy financial use cases there.

You know, what, what kind of financial information are you passing back and forth? What markets or value transfer are you having? What is the best system to create that? It's really that simple.

And it, and it creates, a- and this-- Financial systems need accurate, up-to-date pricing data. It, it is a very, um, time-sensitive, uh, system.

Having delayed pricing means you might have, uh, bad information on, uh, you know, a buy or a sell, or delayed settlement will be a big friction to normal commerce. So that's kind of at the root of what we have built.

That's like differentiation while keeping decentralization and security. Those are always the top in these kind of financial systems on, on any level, centralized or not.

So, and, and cryptography and a lot of other systems give you that, and I, I'm su-sure we'll get into some of those weeds.

But I think there is a bit of a misnomer about both the decentralization of Solana and its, its stability. So S-Solana is a very decentralized pr-protocol. It's completely permissionless.

Anybody-- Like, you could start a validator today just by going to one of the clients that's compatible with Mainnet and downloading it and, you know, the source code, compiling the source code locally.

So there's no-nobody pushes to you. You can just permissionlessly take that code. Theoretically, you could even make your own code that me-met, met the spec if you could. You know, that's a pretty big lift.

But it-- You could. But the-- So we have multiple clients to choose from that, you know, you just run Solanas, and there's thousands of entities that do this.

In order to be profitable, you need some SOL, the native token of Solana, but there's nothing else that re- that prevents you from being a validator on, on Solana and actively build blocks.

There's that side, and then the stability side is just engineering, so re-re-architecting every single part of the stack. Uh, frankly, Solana was released

similar to the, what I was saying earlier, in, like, normal engineering practice, where you relea- you release early- Uh, and really battle test your thing live. That's why it has the beta tag on it.

Um, and, and there was definitely instability at the beginning of the release process, but that has drastically improved over the last years.

In the last two and a half years, there's one downtime of five hours, and we can kinda, kinda talk about the nuances of that, but it's been over a year and a half since any downtime and, and very little instability in that, in that entire time.

So it is, like, the most important part of these things is, you know, again, security, then stability.

So that's, like, again, part of our engineering mindset is to constantly be pushing the needle while maintaining those things. Historically, Solana has always been very quick in adapting to the latest trends.

They've been quick in adapting to NFTs, then on DeFi, Solana is really strong. They had the meme coins. Solana- Yeah... seems to be always first on those new waves of- Yeah... of innovations.

Why do you think that's the case? I think it's just, it's part of the culture, actually. So there's no entity that is like, "Okay, we're gonna become, we're gonna start a m- a meme coin ecosystem."

There's no single entity. It is an ecosystem of... So, uh, the culture of us as the protocol constantly integrating and, and putting forth tools is kind of similar culture.

It's, you attract a similar type of talent from the culture of, you know, the ecosystem. They, they, they buy into the... Oh, th- they see th- this culture of like, this team is rapidly innovating.

They're, they're serving my needs. I'm a developer. I want to make sure that I'm set up for success. I have this business idea.

And so the types of devs who are iterating and really product thinking tend to come to Solana. Like s- uh, said another way, Solana or whatever that, and there's no entity. No, no entity, uh, chose meme coins.

Meme, meme coin businesses chose Solana, and I think that's true for a lot of businesses. DePin is also largely on, on Solana with Helium and Hivemapper and Dawn and all these great DePin projects.

Uh, Geo.net, uh, Port Dover. There's, there's, there's many. And then payment providers are also doing a great, great business on S- on Solana. So it's, it's not just memes. Meme-- Like, so I think

meme coins will do something like more transactions per day than all, on Solana, than all other chains combined. Any, any transactions on all other chains.

So there's a large part of it, for sure, but there's also just a large part that's not meme coins on Solana. So I think that's why we tend to be new to a lot of things.

The other thing is the, the cost structure of Solana is such that even if something blows up and does, like, a huge, by blows up in terms of, like, growth, doesn't price out other businesses that are still exploring their product market fit because the fees stay low.

There's, there's sufficient capacity, and there is isolation of, of markets that if my application is still fledgling and that doesn't necessarily have huge margins yet, um, I'm still, my, my users are still paying small fees where they don't really notice any other spike on some other part of the network.

Whereas another network that may have, you know, small blocks or other global beacon tension, if one parts, one adjacent is- industry starts doing well, it prices out other industries at that time.

Uh, but that doesn't happen on Solana.

Uh, the median, uh, a way of measuring that is, like, kinda, like, median fees and then median fee stability in that, uh, Solana's median fees are, are essentially stable, uh, in perpetuity.

Uh, uh, there's been, been very few spikes, like, you know, when Trump, like, literally the President of the United States launched his, um, meme coin, then his wife did.

Like, that, that's, that's, like, the basically the only time in the last, uh, year and a half where there's been any sort of median fee spike.

So it's, it's very stable for, for a variety of businesses at any given time. Solana is one of the top general purpose blockchains. I would probably also count Ethereum among that, Bitcoin.

And how do you balance those different needs of those end users when you look at one corner, those super degen crypto communities- Yeah... versus institutional users who wanna use Solana for secure high-value use cases?

How do you balance that? Yeah. Uh, I think from a platform perspective, it's actually pretty simple. Again, that, you know, number one is security, then stability, then performance.

Uh, what we're building is a really deep technical system, but the KPIs are pretty universally simple. If you're building just a flexible programming, um,

then yeah, yeah, just tho- those three things and, like, you know, increased bandwidth, reduced latency is really kinda all you need to, to satisfy that general use case.

Um, maybe said another way, like, a lot of the financial infrastructure that we're disrupting is pretty archaic. You know, it's, it's ki- like, it's pretty hard with all these entrenched systems and, um,

a lot of them don't really have a clear owner on how to upgrade them. And yeah, so they end up just stagnating. And so it's, it's having that be...

Yeah, so, so having this, like, fast finality, very, very snappy system will serve both the meme coin kinda degen high-frequency trader and the, the, the security-minded, incredibly high-value RWA or institutional, institutional players.

Yeah. So it's, it's very similar. I think there are still are application-level logic that you wanna be able to customize, and we have different layers of the stack that will handle that.

There's different relayers and different other things that will help with ordering.

That gets, like, uh, into a pretty technical area, but you can do some customization on Solana, and there's applications and, uh, different parts of the stack that are, are releasing tools to do some very-- Do things that other chains would have to build, like, an entire custom stack for.

You can still do that next to the rest of Solana. You probably still have tons of colleagues that work in traditional industries, traditional finance, and if you talk to them and you had to explain them

Why did you enter blockchain, and what's the importance of blockchain, and why, why does this concept make sense? Yeah. What would you tell them? My origin story is pretty transferable, I think.

So I was working in games. I was working at Riot Games, which is a very global, international brand. They make League of Legends and Valorant and a few other really hit games.

And we were just having-- We, we do a lot of our distribution in multiple ways, so like we, we were contracting with Garena for kind of Southeast Asia, and China, uh, had its own distribution with Tencent.

So there was these kind of fragmented distribution mech- things, so we didn't really have all the data. That and you have these third-party esports organizations that would come in and offer their services.

But we didn't really have a way of certifying whether they were scamming the players that we were trying to help set up, uh, in the, the esports teams that we were trying to set up businesses for.

So there's just all these problems with like, you know, transparency, global payments. And at the same time, there were some major new regions that were tr- suffering, suffering from a bunch of charge ras- chargebacks.

And entities, uh, Visa was even thinking about delaying our, uh, you know, certifications. They were thinking about just cutting off all payments in certain, like very high-growth regions.

And so you just look down these rabbit holes more and more and more, and you're like, "Well, just accepting crypto fixes a lot of these things." I mean, this is back in 2017, um, when, uh,

I first, first started looking at this, and it was very practical. Like this solution just solves my problems. And, uh, yeah, then I started a business that was like basically related to all that. But it's really simple.

If you, if you want this, you know, you know, digital high performance, like, uh, fast finality system that, that meets the world of business and automation today, blockchain just ends up solving a lot of those problems.

As a tech leader building daily on those solutions, how actively are you following the markets and the market cycles and everything that- that's happening out here? I definitely have to be aware of it.

It's, it's, I think a lot of times it's distracting. We're on this kind of infinite game where we're trying to constantly be creating these like information systems for, for, you know, as,

you know, y- you upgrade hardware, you upgrade networking, you, you upgrade, you know, eventually we might get into neutrino communication. Like it, it really is this kinda like never-ending

economy and just like transformation that we want to facilitate. So any local thing that's very reflexive and narrative based, um, tends to just kinda be a distraction from your like, from the ultimate goal.

Yeah, so I'm aware of it because it does affect the businesses on top, 'cause we, we, we, like in order to get to a place where you're actually meeting the customers, um, and, and, and solving their problems, you have to learn from them along the way, so every iteration is, is better and you can get to a fi- a final destination.

I don't claim to have some crystal ball that knows exactly what I need to build, uh, you know, for 10 years from now. And so working with our great ecosystem helps us get to a, a, a good place.

Like I- IBRL gets us like maybe 80% of the way, way there, but like the actual thing that gets you product market fit is really just tackling every single 1% gain that you can for your, your, your users, which, you know, which are applications and devs.

So def- definitely versed, but yeah, d- it's not my main focus. Do you have a thesis about how blockchains and in particular Solana and the Solana e- ecosystem is gonna play out over the next couple of years?

And do you have features or things you're building on that are aligning with that thesis? Yeah, my, I have a-- Part of it is very boring, where there's just value transfer is just way more efficient.

Even Visa, they settle, they, they, from the user perspective, they get the, they have the payment go through, and they g- I get my candy bar or whatever.

But from the merchant, uh, they only settle, Visa only settles once a day, five days a week. It's, it's, you know, like at night sometime.

And so if, that does not meet the needs of kind of the next iteration of, uh, you know, we're entering into these autonomous systems and not necessarily AGI or anything like that, just like faster moving businesses that are digital that want that capital so they can, you know, maybe purchase more compute or purchase more, uh, AI inference.

You know, like if I'm just like an, an LLM wrapper that's serving some business generating images or something, uh, or generating text stories, if I get more business, I wanna be able to pay those more to generate more stories or whatever else is making me money.

If I have to wait for the settlement of the credit card payment to come in, uh, that just slows my business down.

So I think over the next three or four years, what, you're just gonna start seeing businesses that accept cryptocurrency or systems that use cryptocurrency rails underneath, uh, are just gonna move faster than businesses that don't, and they're gonna be able to have access to global markets.

They're gonna, you know, be able to have more bit creative business strategies. So that's like boring but very, very, very functional.

I think on top of that, once you start getting these flows that you can see, you're gonna be a, your, the capital formation is gonna be much simpler. I am going to, I have some capital expenditure I need for my business.

Uh, I have some certifiable business model that now is m- more transparent that I'm getting it through. I will rev share you.

I will buy back tokens or whatever for the upfront fr- upfront capital you give me, whatever that business model. It's, it's very flexible. It's in code. It's transparent to what it is.

There will be certainly some iteration on what exactly is good for investors and in, and for the businesses themselves, but you're just gonna see different types of assets that

eases the friction of new types of businesses. And I, and you've already seen that happen. This isn't a theory. You know, Pump.fun on Solana, uh, you know, it's, I, I, it's, it's, it's highly varied.

People are still figuring out what the actual valuation is, but it is a, is a business that- You know, generates, uh, multimillion dollars a day and buys back some of their token, and everybody can kinda see when they do and don't buy back and, um, the, in the tokenomic model and everything.

So it's, it's, yeah, it's very much happening today and, and as they figure out the right model to their investors and their users, then you'll see more and more businesses take that on.

There's, uh, several others as well, and that's like a social app. Uh, it is a, uh, you know, they, they...

it's a meme coin, but al- but also like a streaming platform where the streamers can either, like basically have a representation of a token.

And so it could be something that like rivals TikTok or, or Twitch or something like that, uh, eventually. Um, there's plenty of other apps on Solana that do this. Time.fun is another.

So there's basically this, this, this capital formation for different types of businesses that y- that, um, people are gonna figure it out.

Um, DePIN is another one that did this even, even before that, which is kind of the, the Helium approach, where they raised upfront capital and incentivize internet systems, a G internet f- with, you know, their token, and now they route T-Mobile and other data off-boarding from actual carriers through their network and get paid for it.

So I think you'll see a lot of these different types of models that are enabled because you have a much more efficient, uh, capital markets.

I, I, I wanted to ask you, and you mentioned a couple of use cases now, Pump.Fun, Helium. Yeah. What are the, some of the things that are currently being built or have been built on Solana that you're excited about? I,

I think there's a-- So there's, the main product market fit with blockchain, not just Solana, but just blockchain today, are these kinda like speculative cases. So there's a bunch of,

you know, you've, you've probably heard of Polymarket, uh, integrating with Twitter X, and that's kind of a prediction marketplaces. And there's come, some coming to Solana.

Calci just integrated Solana, which I think is an interesting case. So you can start using predic- I think you'll, you'll-- Those prediction markets might even go more and more mainstream.

Just like with any sort of financial product, there's, you wanna be able to do whatever hedging and risk mitigation strategy, and I actually think it might even start entering mainstream as, as a hedge in certain cases, where something you might find something correlated with whatever financial strategy you're doing.

In the same way that like an airline is really into the

futures markets for oil or for, for, you know, jet fuel, uh, you might see some businesses getting into prediction markets for some hedging on some outcome that might be bad for their business.

So I think that's like a very interesting exploration. You'll see a bunch of different iterations on that.

Some of them are even gonna potentially try to meet institutional demand or like even like stuff like housing markets.

Like if I think like the housing market in Seattle will go up, maybe if I'm gonna buy in a year, I might like, you know, buy in somehow to like kind of risk mitigate.

Like, oh, if it does go up, at least I get like some level of the price right now and some, some level of price stability.

Or similarly, if I'm about to sell my house, you know, I don't want the Seattle hou-housing market to tank. Maybe I short the s- the house, Seattle housing market just in case it does drop.

I still get some risk mitigation when I sell my house.

So like you'll start seeing some of these things actually get, enter into the j- normal world of finance as part of people's, you know, whatever financial risk mitigation strategy.

There's plenty of other things though too. There's, again, there's very functional, like Sphere Pay is doing international business payments.

You, you, you have just traditional, uh, kinda stock and equities coming on chain.

You have kind of pre-market stocks, like they use like these debt security, uh, things, but, you know, they get, there's SpaceX, uh, a, a bunch of, you know, private, Angel, p- a bunch of private companies that have stock one-to-one backing on Solana.

So you'll start seeing like a lot of these explorations around. They're, they're almost all-- Like, bl-blockchains are really, uh, financial products. It's value transfer. It's, um,

uh, you know, different markets and, and trading. So I, I think if you want to infuse that into other things like games or social products or stuff like that, those would be the blockchain elements, would be those.

Yeah, that, that's kinda like the ecosystem that we're seeing building.

And then the cool thing about Solana is it's all in one place, so you can actually potentially have all these things interoperable with each other, so you could have some highly liquid markets that, uh, you could execute a strategy across industries.

In 2019, 2020, we saw a, a big generational upgrade, I would say, uh, for Layer 1s. We had the switch- Yeah... from proof of, um, work to proof of stake.

And since then, all those blockchains have become a lot more cheaper and a lot more faster. Yeah.

What do you think is the next iteration or generational upgrade and, and what are you excited about from a technological perspective? So Solana is the most efficient spec of all of those.

Other people have, other, uh, blockchains have made different design decisions.

F-first of all, I wanna say that the, often what, what is cited here is, uh, is what's called a trilemma, and that was mostly a blog that, uh, Vitalik or, you know, a couple others kind of proliferated a, a while back.

I think they've actually since taken it down because it's not a proof. There, there are some fundamental, uh, engineering principles. One is called the CAP Theorem.

That is a, uh, it talks about availability and consistency. So, uh, that is a very provable thing.

The trilemma about scalability, decentralization, and, uh, security, that, that is not, it's not actually a principled argument. So Solana is certainly at some scale, there's probably some trade-offs.

There's trade-offs with engineering everywhere, but it's not like some certainty that just because one protocol is faster or cheaper than another, it doesn't necessarily mean it's, it's not as secure.

So like the lack of stability that you may have, that, that you saw in s- in Solana in the early years was just because of very, very new tech. It wasn't because of some spec issue that, that is related to S-Solana.

So I think the The thing that we're seeing is, um, just, like, again, that, that gradual, um, uh,

fruition of our, of our tech, and then we have some very big upgrades to get to, like, kinda, like, I don't know, like a local end state.

Job's never done, but we have a consensus algorithm that provably gets to the fastest, um, consensus time, uh, in a globally distributed system.

We have some of the, the best researchers, consensus researchers in the world. We hired them directly from ETH Zurich.

Nothing to do with Ethereum, but it's a, it's a premier, um, engineering university in Sweden, um, and we hired them... Or sorry, Switzerland, um, in Zurich.

So we, we hired them on to make a new consensus algorithm, and I can get into the technicals, but it's very provable that it works, and it's, it's a very simple but, but very elegant and fast algorithm that reaches finality in,

like, around 150 milliseconds. So it's, like, kinda like internet speeds. So there's that, and then long term, uh, there's even, like, bigger upgrades that make Solana competitive as just like a information system.

So right now, there's no claim that Solana can be faster than NASDAQ. N- NASDAQ works on microseconds, but it is in, like, an isolated system. It doesn't interact directly with, like, a payment layer.

It is just a trading optimization. But just for that one purpose, Solana doesn't compete. It's, uh, you know, we have this whole consensus thing that we have to bring to thousands of validators around the world.

But there is an upgrade that we are currently designing that's, like, pr- pretty well baked at this point. It's called multiple concurrent leaders, and that's where

multiple block builders around the world can be making blocks together at the same time. And that basically means that if I'm in Tokyo and I get some information, I can

have my local block builder include that transaction in, in microseconds, and I can-- Same can be true for somebody in New York. Same can be true for somebody in Dubai.

Same can be true for somebody in wherever in the world. They can guarantee, they get inclusion guarantees from their local validator in those microseconds and still have it be synced to this global market.

So that is, uh, a slightly different system than NASDAQ today, but it has different trade-offs.

So you don't have the-- You still have to wait a little bit longer for the financial, all of the financial interaction, like, to see how it actually interacted with all the other information.

But you can get your inclusion assurances near instantly anywhere in the world under that multiple concurrent leader design, where if, if I'm in Tokyo and I'm trying to trade on the New York Stock Exchange, I have to wait for New York to get back to me whether that, that, uh, trade went through or whether that, um,

yeah, my trade was gonna be executed. So it's, it's,

like, a rather huge upgrade and, and really starts making Solana have not just security and global assurances and being, like, this kind of independent financial system.

It actually gives it, uh, performance properties that are better than any centralized system can be. So it's not just about decentralization from, uh, an ethos or, like, a self-control purpose, which is very valid.

But it adds in, like, a performance and information, um, guarant- like, inclusion guarantee over, over a centralized system.

Are there any institutional use cases that are already testing things like that on Solana right now?

Uh, I mean, the, the main one is, is a lot of these, not MCPs specifically, but Solana already is a globally distributed system that entities all over the world p- participate in.

The, a lot of what you-- Global access is kinda, like, the main one.

If you want somebody in, I don't know, some, some rural place to, you know, be able to have access to the US dollar or equities, yeah, you, you start seeing institutions offer their stuff on Solana.

So a few institutional examples, uh, Galaxy just announced yesterday that they're tokenizing their stock on their actual, you know, equity stock on Solana. Another good one is, is R3 Corda.

They work with a lot of institutions, and they're usually in these, like, private blockchains, but they want access to global liquid markets, so they're executing on Solana mainnet, um, a lot of their strategy.

And, and they, they carry a lot of different types of securities or, uh, debt instruments. Uh, uh, like, this gets a little bit... They have so many things. They, they, they...

But yeah, so now they're gonna access to markets and liquidity of, of mainnet.

And the really interesting thing about this, maybe I haven't mentioned this quite yet, like, one of the, uh, other benefits of decentralization that's not just for the protocol, is that any entrepreneur now can see the assets that they're issuing and see how to interact with them.

They don't need R3 Corda's approval.

They just can formulate a business and add value to R3 Corda's offering just by knowing how to interact with those assets, so if they wanna make a marketplace or some derivatives market, whatever they can, and yeah.

And if users like it, they can use it. So it's, like, right now, if I want a financial product and I'm using Fidelity, I'm kinda beholden to everything that Fidelity services me as a user.

Whereas on a blockchain, anybody can kinda give that service for, for a user, and it really-- And, like, institutions certainly, uh, have some hesitancy there 'cause they, they don't want, um, their assets to be used in a way they don't want or, you know, security risks and, and, and otherwise.

But it does make a much more open ecosystem where they don't have to build everything.

So, and they, and they, and they, there's obvious benefits in terms of liquidity and global access, because the other thing that, that they don't have to worry about is on-ramp, off-ramps everywhere in the world.

Other people are doing that for them so that everybody has access.

So that's, like, one of the easiest examples of what I'm talking about, is I, as a, as an R3 Corda, I don't need to be onboarding to every single continent and every single region.

People are doing that for me, and I still get the benefit and my, and my users still get the benefit for me 'cause they get to buy in or whatever from, from globally.

All I have to do is make sure I'm issuing the assets in the, the best way possible or whatever, you know, that, that business is. Um, and then you can get into all the other, other pieces.

But That, that access and that liquidity is very attractive to, to institutions. And from an institutional perspective, how do you think about privacy?

We, we, we've seen the first blockchains now positioning themselves as a kind of private blockchain. We see- Mm-hmm... uh, blockchains like Canton, blockchains like Avalanche, who- Yep...

allow you to create Avalanche sub chains, Polkadot para chains. How do you think about privacy? First, this is gonna go back to what I was, I was saying earlier with increased bandwidth, reduced latency.

Privacy is just math. It's like cryptography and other-- We don't have to get into, like, all the mumbo jumbo, but it just requires a little bit more advanced computing.

So the more and more efficient you make the underlying system that is Solana, the more and more there can be privacy layers that just make use of that in an inefficient way.

So we actually view it as one and the same, and you're seeing that already.

There are-- We have a confidential balances, uh, feature baked in directly with some of the token standards, and that's basically-- It doesn't off- Like, if I send token to you, um, it shows up on the ledger that I'm still to- I'm transferring to you, but it doesn't, it makes the balances confidential.

So it doesn't know if I'm saying five USDC or 20 or 40. You, I- You just know there's a, uh, there's a transfer. There's other protocols that are

obfuscating the entire thing, the, the entire transfer, but still allows for value transfer. Uh, and then there are privacy layers that are, we call them network extensions.

One ag- example that was just announced this summer was called, is called Magic Block, and what they do is they kinda, um, allow you to either mint or, or, or, or, or, or, or issue some token and kinda delegate it to a private trusted execution environment, and then you can still run any code that you want, uh, that's, you know, on Solana.

You can run markets. You can run whatever, but it's done in a, an environment that is hidden from, from public view.

And then if you want to access Solana liquidity, you still have all of the proofs and, uh, the state right there that you can trade on the public layer and still interact with some of, some of these others.

So you, you're starting to s- see, like, different approaches to privacy, either like a network extension where you're happening in this trusted execution environment.

You're getting it directly on main net with cryptography. We see a-- There's-- I can't quite name all of them. They're still in stealth, but there's basically just a bunch of different privacy solutions.

One of the ones I can talk about, 'cause it's just kinda known, is there's basically ways of doing kinda ZK at- attestations. So as a user, I can-- or like a credentialer, you can certify something.

So, like, like, the, the, the-- This isn't gonna happen anytime soon. Maybe, hopefully it will. But, so, like, say, like, the United States of America wants to start issuing passports, uh, credentials on-chain.

Uh, they could do so in a way that just has a proof of all the information on my passport, like my nationality, my, my age or whatever, but I wouldn't-- It could be private what that information a- actually was, and I could just as a user just issue statements like, "I am an American," and it would show, it would say, "Yes."

"I am over the age of 21," and it would say, "Yes." Uh, and so I wouldn't have to give away my, you know, all of my personal information about my exact birth date or my address or anything like that.

It would just be in this cryptographic proof on-chain, and so it's a sense of privacy of that. So it's like, th-that's just, you know.

And that could, that same thing could be used for arbitrary information, like my income or any accreditation, uh, I have achieved, like a pilot's license or a- or anything.

So it c- And it can-- And this is actually the way I think any sort of credentialing should be done because a lot of credentialing, you really only need one part of it in order to certify something, a loan or some access to something.

You don't need to know every single thing about the service or the, uh, credential that I, that I, that I obtained.

So we're, we're gonna see that more and more, and there's-- We already have this, uh, Solana attestation service that helps issuers with this process.

And yeah, adding privacy on that is, yeah, it's, it is a, it's a natural extension of that, that feature that, that people are utilizing.

So you- Again, just to summarize, you're seeing it in terms of, like, balances and transfers, but also just, just raw information or just, like, compute entirely.

It's really like-- And, and that's part of being-- Like, part of Solana's ethos is being that this new layer of the internet is that it's connecting all of these different types of privacy layers, and then it's just up to the users and the businesses to figure out what, um, yeah, what people, like, care about using on any given day.

Mm-hmm. And Matt, I, I feel like I have to ask you that because I hear it over and over again. Many people are talking about quantum proof of blockchains. What, what do you think about that?

Are blockchains quantum proof? So quantum security is, is definitely something that we are, are actively tracking. I'm actually a mentor to a quantum accelerator here in Seattle called, uh, Barco.

It's-- So we're, we're very much, like, in, in the conversation about the timelines and also the quantum secure cryptography, 'cause there's, there's always, like, that, that zero day or the Q day, as some people call it, where quantum can start-- quantum computers can break common cryptography that's used across the inter-internet.

There's an algorithm called Shor's algorithm. So basically, anytime you sign something, it's like a factorization mechanism, and Shor's algorithm can re- reverse engineer that. So that's the attack m- mechanism.

That's why you actually see a lot of people just moving their funds to a new wallet that has never signed anything before, 'cause the initial attack case is only on wallets that have signed something before.

So you're actually pretty safe if you haven't.

But there's already, uh, on Solana quantum secure vaults, so it would require both a signature from Ed25519, which Solana uses, but it also require a signature from a quantum secure algorithm, and you can execute strategies from those quantum secure vaults, and all they take is more compute.

So again, it goes back to- The more performant you make the underlying layer, the more room you have to execute more advanced cryptography. That's really it.

So I, I know it's a very common theme of this entire thing, but you just, it, it's, it's just such a core principle in engineering is that if you make the underlying system more efficient, you can do more things, and you can have more efficient markets and apps.

So yeah. So eventually we, um, like even today, we have enough capacity on Solana to do way more advanced cryptography.

The challenge is just going to be the, the crossover, 'cause eventually we're gonna have to make it so all signing mechanisms on Solana use quantum secure cryptography, and so, like, the addresses will have to change and other things.

But we are actually very much underway on that.

We have a team that is doing, or actually multiple teams now that are doing quantum secure Solana, both analysis of different signing mechanisms and also to tr- um, you know, trans- tr- uh, transformation plans.

So how do, how do you get the ecosystem to, and the wallets to get people to move over to different, different addresses and things?

Those, both of those things are going u- uh, underway, and we are very much ahead of the game. I think that's, that's the main thing here, keeping track of all the advancements, and so there's not gonna be surprises.

Mm-hmm. I think... Okay, actually, maybe, maybe I'll be able to say this another way. I think when people hear these, like, very deeply secure things, it's, it comes off as, like, very scary, right?

Like, because it's, it...

But, like, from my point of view, like, a normal engineer will just tell you that just, like, there's always gonna be some risk of just technology flourishing or, you know, some outage here or there.

Like, this is all, these are all, like, pretty standard engineering problems, and they, and these are v- very standard engineering solutions. So it's not scary for us.

It's just what is the, the trade-offs in improving to, to meet the needs of the engineer, engineering problem.

A- and so I understand why people are scared because it does change a lot of things, but yeah, it's, this is very much within our, our wheelhouse to, to, to upgrade. Super fascinating.

Matt, a last topic I wanna touch on is AI. Yeah. There was a time about six months ago where we had that AI boom, and I would say almost the bust, and everyone- Hmm...

was trying to figure out how can they position their chain more actively in AI. Yeah, yeah. What are you doing in, in AI, and how do you see that coming together with blockchain and Solana? Yeah.

Uh, I actually have a background in AI. I was the head of machine learning engineering at Unity, which is a game.

So we were, but we were doing, um, like, simulated training and simulated compute and a bunch of other stuff for AI training, as well as making, like, AI bots for robotics, like, kinda simulating warehouse use cases and other, other things.

So anyway, so I, uh, am very in tune to kind of the needs of AI.

I think you're right that a lot of people kinda really got ahead of themselves, and they went right to, like, human o- or, you know, provably owned AI and all these things that we're don't re- we don't even really have

AI assistants yet other than just kinda, like, voice prompt things that, uh, I wouldn't say somebody's... There's not many AI that are, like, actively running our lives yet, so, like, owning them is, is kind of

cart be-before the horse. There are some very functional things happening, though.

Data collect- I, I think, I think what you'll first start seeing is kind of the, the training process, so from, like, data collection, annotation, uh, and training and a lot, a lot of that, like, infrastructure side, you'll start seeing, like, really start getting some adoption, and that's already happening today.

There's, there's, you know, a lot of the DePIN companies I mentioned earlier are, are essentially da-data collection companies, and, um, AI is a, is a prime use case for that.

Like, Hivemapper has partnerships with Lyft and, um, Volkswagen, I think, and a few others, and a lot of that data gets used for AI purposes.

They do, they do mapping, so they take live, like, a, a more up-to-date v- street views, and so people can use those in, in whatever AI models they have, um, to update, you know, traffic information or lane closures, that kind of stuff.

Other side of it, so that's like the inference and the, or sorry, the, the training side of it.

There's also inference side, which is also just infrastructure, uh, you know, buying GPUs and, and, and obtaining GPUs for much cheaper, and so you don't have to always go through very expensive data centers.

And yeah, and then there's also then using AI on our side. So this is where, like, so up to th- to now, I've talked about AI using blockchain. I think it's just gonna be very functional for the next couple years.

Uh, the other side of it is, like, how the blockchains use AI, and I think that one actually gets pretty interesting 'cause it's one of the early product market fits of AI is helping coders.

'Cause we're all, like, we all, we, people making the AI know the coding problem because they're coders themselves, so it actually makes a lot of sense that it's one of the more refined things.

Some people kinda joke that, like, I think there's this weird view that AI was gonna replace, like, all the manual labor. That's actually the hardest thing for AI to solve.

Like, translating an AI algorithm to something that works with all, like, the nuance of, like, little motors and, like, robotics and all the variations that can happen with wind and dew points and, like, literally anything.

Like, but when you have just completely digital systems, it actually is much easier for AI to just learn exactly what to pattern match there and do that.

So things like radiology and, uh, like, a lot of things that we actually think are professionals degrees, which includes, um, yeah, like radiology, lawyers, and stuff like that, but also coding.

So w- we wanna be very in tune to how to help the AI help coders. We don't think it's gonna completely replace coders anytime soon. It's just going to allow you to build applications in a much, much easier way.

Like, maybe, you know, five years ago and prior, there's, like, this huge SEO push, like, like making good businesses, can you efficiently appear top on, on search engines and get, like...

You know, there's always, like, golden gooses of, like, can you find a highly searched thing that doesn't cost a lot and form a business off of it?

Now, that's actually translating into how can you make your updates to your platform as AI readable as possible, so when somebody asks about making an application, you know, you get recommended.

So, like, that's like A thing that a lot of, like, different, different providers are, are, are, are, are pushing for.

So, like, you wanna make payments, eventually you want Solana just to be like, "Hey," AI to be like, "This is a good solution, is you-- if you wanna issue an asset or do payments, here's, like, the way to do that on Solana."

And then there's even, like, the, the lower level system than that is, you know, how to make your docs whenever you update something, 'cause we're-- again, we're constantly iterating and, and improving the chain, and sometime- sometimes we change or add a new feature.

How do we make sure the AI is reading that correctly live, knows to-- it has a discoverability issue?

And so that's, like, the next generation of, like, SEO type of, of thinking that, that, you know, that we're very much in tu- in tune to.

So we put out a rag system, we put out our own, yeah, like, multi, uh, MCP servers, and yeah, and are working with several different AI systems to read our docs better.

So it's, it's really, uh, kinda, like, all-encompassing thing to kinda, like, be involved in this, in this, um, evolution.

I will say, though, I, I heard this, I think this was on, like, the All-In podcast or something, uh, a little while ago, but they, they described what's happening in AI as an evolution, not a revolution, and I think that's kind of a good framework.

So I understand why people think, like, like, a bubble popped or whatever.

I think it-- people just expected this, like, this idea of a singularity where AI just starts taking over to happen, uh, right away, and I think there's a, just a, f- frankly, just, like, a lot of, um,

engineer energy and infrastructure and other just, like, limitations that make it, like, hard for that to do right away.

What I think we'll see is, is just evolution, where businesses will just adopt certain parts more and more, like, individual employees will just use it to make more efficient workflows.

And yeah, and, uh, frankly, part of that is if, if an AI finding the most efficient workflow, we just wanna be there to be the most efficient payment workflow, uh, for that, and, and, and use the most liquid markets.

So, like, when I say things like finance tends to flow to the least friction, uh, AI is just gonna find the thing with least friction, and the user may not even necessarily know how it was able to pay the person halfway across the world and access some market that they didn't know how to access themselves, but the A- just, AI just figured it out and it, it just happened to be on Solana.

So that's like, yeah, that's, uh, kinda where we're going. And I, I fully agree with what you, you were saying about the progress.

Uh, it definitely feels like, at least from a user side, on the models themselves, the progress is slowing down. So the jump from- Yeah...

uh, GPT-3 to GPT-4 was bigger than what we now saw with G- GPT-4 to 5, for example. Yeah. Yeah. And also- Yeah... a funny anecdote to, to what you said before,

um, I think Brian Armstrong, founder of Coinbase, tweeted that yesterday, that they are already coding 50% of their total- Yep. Exactly. Yeah... code base with, with AI, and they wanna bring it to 60% in the next months.

Yep. Yeah. And that's-- And we wanna allow people to do that for Solana too. It's a little bit harder for...

So I'm, I'm sure that includes, like, their website and other-- Like, they have a lot of features that aren't just their financial features.

Um, but yeah, like, w- when you're making really, really low-level systems, the-- and not, that, not just, like, the front-end JavaScript stuff, it's, it's a little bit harder for us to use it f- on the core protocol.

But yeah, I-- We wanna make app development on Solana as simple as that. We want people- Yeah... just to be able to say, "I wanna accept payments. I wanna make a trading bot.

I wanna, you know, execute this strategy," and then just have it, have the AI be able to take, take off. So I think that's- Yeah... kinda where things are going, for sure. Yeah.

And actually, the, the highest rate that he mentioned was Base. So they used- Mm-hmm... the highest rate of AI, and the lowest rate- Oh, that's crazy. I didn't r- [laughs]... was, was... Yeah. Never mind.

For, for their institutional product, that was the lowest rate, I think about 40%. Base was a- Okay... around 60. Yeah. Wow. I-- Okay, that's surprising, I'll be honest. 'Cause, like, the, the

way it works is it's just pattern matching. Like, what, what are the, uh, systems that have already been built that you can kinda

save a lot of time by, uh, especially, like, cross-integrations, like mixing this API with this, this, you know, front-end element and stuff like that.

It would just-- It's really good at just knowing how to do that thing.

There may not be like-- It's really hard for a human to write individual docs on how every single API should interact with every single API, but there is some code for it living on GitHub somewhere that, uh, it's much easier for the AI to just read all that and, and know how to do it.

It's really-- The typical use cases typically don't get into core protocols that is, like, very custom and novel systems, but I'm sure Base probably has some, uh, yeah, so some systems that are, you know, maybe cryptography-based or something that's, like, kinda, like, known systems that they're kinda reproducing.

But it i- yeah, I would, I would not have guessed that, obviously. As a last word, last question, is there anything you would like people to know about Solana, about blockchain, that you think is important? Um, yeah.

I think it's important to know that-- I think a lot of people still have a misunderstanding of blockchain, I think. It's because it's very deep tech.

It, uh, r- involves these crazy terms, cryptography, we've talked about quantum, consensus, blah, blah, blah, blah.

But really what you need to know is, like, there really wasn't a w- a way to have a digital, own a digital thing.

And, um, so, like, you can have this digital thing that o- only you can choose how, how, where, where it gets used and, you know, how to, how to transfer to somebody else, all that stuff.

Like, there was IP rights, and there were individual platforms that would broker things to you that were kind of digital, but there was no real way for me to just independently own a digital thing until blockchain.

And that's really kinda where we're going, 'cause more and more things are becoming digital, and if you can also make settlement times fast and, you know, that, that, you know- Uh, and that's why people call it sovereign money.

There's always different ways of saying it, but really it's that simple. And that's why, like, the CLARITY Act is really kind of trying to define what digital assets are.

And I think lawmakers seem to really understand it actually, 'cause they don't really know how to write law around digital things yet.

And so they see, like, oh, we have this blockchain thing where you have a person who owns the ability to sign whatever kind of thing they wanna do with this asset. This is, this is, this is functionally new.

This is not something that was able to be done before.

So I think at the end of the day, any institution or individual should start thinking of, like, what are the digital parts of my business or what things would I like to digitize, make digital, more global, and how does that fit in?

And so it really is that simple. It's, it's not like the c- the core, like, delivery to the user is pr- is pretty simple. I can now own digital things. Thank you, Matt. That was a wonderful ending word.

Now a, a quick lightning round. Those are short questions, short answers. A book or podcast that influenced your thinking on technology. Oh, wow. I don't know.

I mean, I, I, I've, I've been going through this quantum book actually. [laughs] Uh, it's kinda like the, the intro to, uh, quantum, um... I think is, is very influential in how I think about the world.

The world is way different. Like, we, we, we, in our eyes and our ears, like our senses kinda like distill the world down to something that seems way more deterministic than it actually is.

We actually live in a world that's, has to deal with way more probabilistic behaviors.

And, you know, even when light shines, it's kinda like exploring all possible paths, and we just kind of observe and, like, our mind con- coalesces it to one path.

But yeah, there's just a lot more actually going on with our universe that, um, yeah, is just not something that we would interact with day to day.

And I think it kinda opens your mind to not necessarily being stuck in whatever bubble that you might be in, your problems of the day or whatever.

Uh, you know, these are kind of like other things that we as humans have evolved to kind of really, really care about because... Sorry, this is a lightning round.

But, like, w- we've evolved to really care about, like, things that won't kill us, right? Like, so, so we, we've kind of distilled our world to these very simple terms.

Um, but I think, like, learning much more about, uh, quantum and these other, like, base physical properties of our world, uh, or physics properties of our world really shows you that, uh, yeah, there's a lot more nuance and intri- intricacies, intricacies, eh, that we can't, like, always perceive every day.

Mm-hmm. Good. Next one, proof of work or proof of stake? Uh, I mean, I don't think it's, you need to have a choice. I mean, and d- different ones work for different things.

Like, proof of work is a, you know, like hard money, like, uh, a distribution thing.

Proof of stake is like a very efficient, very efficient at securing a network that is financial and does arbitrary things for, like, arbitrary businesses. It's, like, just slightly different trade-offs.

Also, the book is called Introduct- I- Introduction to Quantum Mechanics. Perfect. So I, I knew it was, like, Intro to Quantum. I just didn't remember the... Yeah.

It's by, uh, uh, David Griffiths and Schroeder, so like Schroeder's cat. Yeah, I mean, I don't think that... I, I, I know that these, like...

There's, like, spicy Twitter debates about proof of work versus proof of stake. I just think that, you know, engineering trade-offs with everything.

I mean, I can go into, like, the, the details, but, like, proof of work, like, if you're trying to be a, a, a super fast system, proof of work is kinda like a varied mechanism.

So, so some variance on, based on difficulty rates and stuff, so block times can't be as syncopated. So you're trying to build the most high performance systems that runs financial applications that are arbitrary.

Proof of stake has a much more consistent leader schedule. That's actually one of the main innovi- innovations of Solana with proof of history. It will change slightly with Alpenglow.

It's like having-- But it's, it's, it's having this leader schedule that is very tight.

I go, and then you go, and then the next leader goes, and, like, there's very little, there's very, as minimal gaps between all the leaders as possible. Everybody knows when they, when they, um, it's their turn to go.

With the proof of work, the next leader is the, just the winner of some contest, so it's v- both varied on the block lengths and who goes next.

So it's, so they have to, like, re-hypothecate, uh, the network just when the next block builder is building their block, who's going.

And then it's also just because of the lack of determinism of the next block, the finality tends to be a little bit longer. So it works really, really well.

Proof of work works really, really well if you're trying to have this, like, distribution mechanism across blocks that is just based on, like, h- hard work. And, like, I understand. Like, I, I believe in...

Bitcoin's great. Like, it's, it's hard money, and, uh, it's, like, globally liq- like the most, probably the most globally liquid asset in the world because of this, like, access.

But if you're trying to build, like, a highly performant system that's still decentralized, proof of stake tends to be better. The first cryptocurrency you've ever bought and the last you'll ever sell. That's a good one.

I think the first, like, kinda-- I mean, my first one was Bitcoin. But the, the first one that was kinda weird, uh, I actually bought some Qtum at the time.

Uh, and it was like, I kinda like, they liked the idea of kinda like a hybrid of, uh, both a, like, UTXO model and a, like, general purpose, like, Turing complete language.

And so, uh, you know, in fact, like, yeah, there's like a few blockchains that have done various forms of that over time, and I no longer, like... Whatever.

I don't mean to talk trash about any given token, but, um, yeah, that was, like, kinda like my first exploration to something that's not, uh, Bitcoin. The last token I'll ever s-sell, I mean, probably Solana.

Like, I, I think... I mean, I'll always be a user of Solana, so I'll just... So it's-- And I wanna be involved in the governance, so I just don't see a reason why.

No matter what the token will trade at, like, um, being just involved, like being a voter, active voter, and participant in the, in the network, I will, I don't see why I would ever sell all of it. Cool.

And, and last one, if you had to describe Solana's competitive edge to a CTO in one sentence, what would you tell him or her? Oh, wow.

Uh, I mean, other people are probably way better at, like, these marketing things than me. But, I mean, I think it's just internet scale capital markets where you can...

Yeah, that's, uh, yeah, internet scale capital markets. Like, that's what Solana is, is hyper-focused on delivering. Cool. Matt, where can people learn more about you, about Solana?

Solana's, hopefully, hopefully solana.com. Like, a lot of this communication also happens on Twitter through our handle, like @Solana. I am actually hilariously more on LinkedIn.

I try to go where other people aren't to, you know, establish my lane. But that, uh, yeah, I'm just Matt Sorg on LinkedIn. Matt, thanks a lot for coming on the show.

It was a really wonderful discussion with you, super interesting, and I wish you all the best. Yeah, yeah. Thanks for, thanks for having me. It was a great time. [outro music]