One of the main benefits that crypto protocols offer us, and certainly one of the reasons they’re valuable to humanity, is permissionless ecosystem participation. The best crypto protocols deliver new rulesets that helps thousands of individuals collaborate and produce value together without needing to trust each other. A decentralized application with a large number of stakeholders self-selecting to be in their orbit will be more productive (and therefore more valuable) than an alternative that has to coordinate those partners in an expensive centralized fashion. Instead of firms facing increasing coordination costs with each new hire, their organizing costs can actually plateau and deliver economies of scale benefits. Instead of a single organization educating the market, participants onboard each other. This is what drives crypto valuations - one cannot simply discount some cashflows; they must project how the total size of an ecosystem itself will grow, and the compounding effects which will ensue.

While overall demand for blockspace continues to spike, our narratives remain in flux. We now have well established crypto incumbents that dominate important markets like stablecoins, yield/lending, exchanges, oracles, staking, bridging, wallets, etc. Their growth continues to be impressive, and their charts show clear signs that we’re in a bull market. But what about the “network effect of crypto applications atop the world computer” that we were promised all those years ago?

There are a number of reasons why I believe the dapp ecosystem has largely floundered and has been unable to find it’s footing since the 2021 bull market. This is hardly an exhaustive list, but I think it offers a good place to start for any protocols asking how they can still try and cultivate a thriving ecosystem of dapps today.

In the early days of Ethereum, there was a very real willingness to spray capital at early stage creative ideas that wanted to leverage tokens and other crypto primitives to solve interesting problems. This was driven in part by naiveté from retail investors, many of whom had never even interacted with smart contracts before. But it was also driven by an appetite for taking risky bets to build dapps within new domains that crypto hadn’t touched before. As time went on, we witnessed funding consolidate around well-validated crypto markets like the ones I mention above, which have figured out repeatable business models for themselves.

We also saw increasing investments made in alt L1/L2 infrastructure projects, in part driven by regulatory clarity around utility tokens, which created large balance sheets to incentivize dapp developers to build there, usually via grants programs. Unsurprisingly, this kind of economic model isn’t an ideal one for coordinating a distributed group of market participants at scale, and is antithetical to how blockchain economies function in their purest form. Did the Satoshi Foundation incentivize miners to invest in hardware via grants, or did they simply create a ruleset for a game that let anyone permissionlessly join and earn based on the merits of their contributions?

In Communist system design, there is an inherent philosophical belief that the enshrined technocrats in the party are the most intelligent and are best suited to make decisions about capital allocation (and therefore market pricing). Conversely, market based system design is premised on the belief that the individuals closest to a given purchase decision are the ones who have the most accurate information, and therefore are best suited to allocate capital. The terminal state of these competing approaches in the real world should be abundantly clear by this point. It should also be clear that crypto protocol design, in its purest form, embodies the latter ideology far more than the former.

Going back to grants programs, several years have now passed and many of these programs have scaled back their funding in the face of expected challenges: sample bias of applicants, adverse project selection, and learning the true cost of successfully managing grants around subjectively valuable applications. Some of these grants programs have done their best to retroactively reward small dapp dev teams based on usage metrics, however this approach still leaves centralized coordinators in charge of the weight function determining which metrics matter, and does not grant users the ability to participate in these reward decisions themselves. Quadratic public goods funding tools have also tried to fill this gap in the Ethereum ecosystem, but again this entrusts decision making power on the select few who care enough to scroll through a long list of projects and spend some of their cash to help sustain them, which brings a similar set of biases that grants programs witness. If only they could vote on which applications they found useful, as they were using them.

When the iPhone first launched in 2007, the world was introduced to a totally new kind of device and operating system. Instead of the old model where the OS was simply a platform atop which applications could run, iOS took it one step further by including an App Store on every single device, and making it difficult to onboard without supplying a credit card up-front.

I expect most modern iPhone users don’t appreciate that this was a significant departure from the existing model in the market: the most popular operating system at the time (Windows) didn’t worry about how software was distributed or monetized. They expected applications to figure it out themselves using the help of third party retailers. This worked well for enterprises that were large enough distribute their apps via CDs in magazines and catalogs and physical stores. But it left the remainder of smaller developers to upload their software to public file directories online, without a real method of monetizing themselves.

Apple on the other hand realized that if they wanted to create network effects around their device, they would need to ensure that indie app developers could create sustainable businesses while building there. Apple was less concerned with individual apps growing as large as possible (if anything this would decrease the supplier power that Apple held), and instead focused intensely on the long tail of devs, many of these teams passing the Two Pizza Rule.

In the years that followed, a burgeoning mobile app ecosystem was suddenly born. Apple offered an improved developer experience by bundling an interface builder and simulator directly into their XCode IDE. For people trying to make money with their apps, Apple’s developer program was low friction for new developers to join at just $99 per year, compared to expensive and time consuming carrier certifications that other mobile operating systems required. Apple also offered native app distribution on every single device via their app store, consolidating the fragmented carrier app store market. They also made it hard (but not impossible) to activate a new iPhone without an Apple ID that had a credit card attached to it.

The result was that if you were in the Bay Area during the late 2000s/early 2010s, chances are high that you were getting pitched on some new mobile app that was about to take over the world. While many of these pitches sounded like fantasies, they were grounded in the very real truth that early mobile app developers were suddenly earning millions in revenue without needing to worry about the biz-ops logistics normally involved in running million-dollar-run-rate businesses.

These decisions were critical: not only to drive stickiness around the iOS platform, but more importantly to explore the full design space of mobile device capabilities that they opened up. The iPhone delivered image/video/audio input and output, powerful gyros, multitouch, a magnetometer, and a haptic motor, all while connecting to bluetooth/gps/cellular/wifi signals nearby. Apple’s decision to help devs monetize meant that if a single individual had an idea for how to combine these functionalities in a new an interesting way, they could feasibly expect to make money if it created value for users. This changed the fundamental market structure for distributing software applications, helping the plethora of long tail devs explore this green field territory. Instead of only having large companies exploring this design space, who have a certain aversion to launching risky experimental products, iOS would now have the maximum number of shots on goal, even more spaghetti noodles thrown at the wall, even more miners digging for gold. While the mobile app market has now ossified and consolidated over the years, this was driven by the market having exhaustively explored the art of the possible, not due to a risk aversion from unsustainable business models.

Looking back at the crypto space, we see that very few ecosystems have quite managed to become hubs which support a large number of small developers building businesses for themselves. If you survey the long tail of devs building on various chains today, they’ll tell you directly that an ecosystem’s ability to help them distribute their new dapp is the #1 factor driving their decision of where to build. Two ecosystems which are currently doing this well are Base and Solana, both of which are happy to confer status on small devs via social spotlighting. Nonetheless this personal process doesn’t scale nearly as well as an app store does, which can automatically surface interesting new use-cases that the chain’s users should be focusing on.

On the business model side, I also believe that by not including some kind of developer fee at the gas layer, the EVM and other execution environments have made a critical error whose costs are being borne in the market today. I believe that we’ve made it far too difficult for smart contract devs to charge fees for their usage, forcing clunky one-off solutions like taking a cut of transaction volumes or rate spreads, which works for financial apps but breaks as soon as a given transaction isn’t immediately profitable. This is why the well-validated crypto markets I mention above are all financial applications dealing with valuable transactions, not simply getting utility from the chain.

The closest we’ve come to a truly user-driven business model for dapps has come from smaller ecosystems like Fantom and EVMOS which let users earn a 15-50% share of the gas fees their dapps generated (the fee % still being fixed across the ecosystem vs varying per-application). Newer ecosystems like Sonic are now coming online, offering a similar model that pays up to 90% of network fees. These ecosystems realize that the ideal system isn’t one where dapp development rewards need to be socialized among network participants.

As a thought experiment, take for example the SafeMath solidity library built by the OpenZeppelin team, which is used in the vast majority of EVM smart contracts today. Based on the original vision for Ethereum dapps, this library would have been immutably deployed once by the team, and new smart contracts would link to the specific address where it was deployed. If the protocol had natively allowed for OpenZeppelin to add an additional line of code that sends some amount of gas units tipped to a given developer address, they could have monetized their work natively wherever it’s used. This also would have incentivized the Ethereum community to coordinate around their canonical implementation, which would have increased visibility and stickiness of their code, improving the overall security for teams which use it. Instead libraries ended up being copied in-line, and the tooling developed away from this pattern. Today we have tens of millions of duplicate copies of SafeMath being stored onchain. OpenZeppelin has thankfully been rewarded for their hard work by the aforementioned grants programs, but one can’t help but think that this is not an ideal state of the world.

In the absence of simple business models for dapps, we’ve seen devs fall back and rely on another kind of business model that crypto offers us: launching a token. No matter how easy launchpad UIs make creating a new token, anyone who’s ever done it can attest it’s a high-fixed-cost activity that is not for the faint of heart. Moreover, having tokens be the primary business model for dapps has meant that otherwise simple modules are forced to imagine some grander vision, where they’re suddenly evolving to become a fractal ecosystem of their own. Not everything has to be that serious!

I think any ecosystem trying to cultivate a diverse garden of applications should think critically about how they can lower the barriers to monetizing their dapps to be as low as humanly possible. Young teams feel like they can’t win building dapps in crypto without raising large sums of money up front to fund development and distribution costs. Things do not need to be this way — this is why there are so many success stories of outsiders building successful companies in the mobile app era. I believe that ecosystems which lower those costs to near-zero for teams going zero-to-one will win in the long run, because they will offer a sustainable way for a long tail of small devs to come online and join the dapp dev market. The terminal state of “play to earn” will be “dev to earn” where instead of clicking a button every day to claim tokens, users click a few extra buttons, prompting their favorite software-generating LLM to deploy a new game they’ve just imagined, which immediately gets distributed to a global audience.

By this point we must recognize that market forces clearly incentivize large applications which fill blockspace to eventually graduate from their native chain and become an ecosystem of their own. If ecosystems want to fight these forces without verticalizing and operating applications of their own, they must create systems which allow small dapp devs to stay lean and build cool stuff without the overhead that forces them on a growth trajectory which many don’t want in the first place. Ecosystems can bear some of the high fixed-costs of distribution and education, leaving dapps themselves to specialize on product delivery. There is symbiotic relationship waiting to be formed, a ruleset which figures out how to get devs paid for the dapps they launch, in a way that’s frictionless enough for the end user that they’re happy to do it. Time will tell which ecosystem will truly get this right.

One of the main benefits that crypto protocols offer us, and certainly one of the reasons they’re valuable to humanity, is permissionless ecosystem participation. The best crypto protocols deliver new rulesets that helps thousands of individuals collaborate and produce value together without needing to trust each other. A decentralized application with a large number of stakeholders self-selecting to be in their orbit will be more productive (and therefore more valuable) than an alternative that has to coordinate those partners in an expensive centralized fashion. Instead of firms facing increasing coordination costs with each new hire, their organizing costs can actually plateau and deliver economies of scale benefits. Instead of a single organization educating the market, participants onboard each other. This is what drives crypto valuations - one cannot simply discount some cashflows; they must project how the total size of an ecosystem itself will grow, and the compounding effects which will ensue.

While overall demand for blockspace continues to spike, our narratives remain in flux. We now have well established crypto incumbents that dominate important markets like stablecoins, yield/lending, exchanges, oracles, staking, bridging, wallets, etc. Their growth continues to be impressive, and their charts show clear signs that we’re in a bull market. But what about the “network effect of crypto applications atop the world computer” that we were promised all those years ago?

There are a number of reasons why I believe the dapp ecosystem has largely floundered and has been unable to find it’s footing since the 2021 bull market. This is hardly an exhaustive list, but I think it offers a good place to start for any protocols asking how they can still try and cultivate a thriving ecosystem of dapps today.

In the early days of Ethereum, there was a very real willingness to spray capital at early stage creative ideas that wanted to leverage tokens and other crypto primitives to solve interesting problems. This was driven in part by naiveté from retail investors, many of whom had never even interacted with smart contracts before. But it was also driven by an appetite for taking risky bets to build dapps within new domains that crypto hadn’t touched before. As time went on, we witnessed funding consolidate around well-validated crypto markets like the ones I mention above, which have figured out repeatable business models for themselves.

We also saw increasing investments made in alt L1/L2 infrastructure projects, in part driven by regulatory clarity around utility tokens, which created large balance sheets to incentivize dapp developers to build there, usually via grants programs. Unsurprisingly, this kind of economic model isn’t an ideal one for coordinating a distributed group of market participants at scale, and is antithetical to how blockchain economies function in their purest form. Did the Satoshi Foundation incentivize miners to invest in hardware via grants, or did they simply create a ruleset for a game that let anyone permissionlessly join and earn based on the merits of their contributions?

In Communist system design, there is an inherent philosophical belief that the enshrined technocrats in the party are the most intelligent and are best suited to make decisions about capital allocation (and therefore market pricing). Conversely, market based system design is premised on the belief that the individuals closest to a given purchase decision are the ones who have the most accurate information, and therefore are best suited to allocate capital. The terminal state of these competing approaches in the real world should be abundantly clear by this point. It should also be clear that crypto protocol design, in its purest form, embodies the latter ideology far more than the former.

Going back to grants programs, several years have now passed and many of these programs have scaled back their funding in the face of expected challenges: sample bias of applicants, adverse project selection, and learning the true cost of successfully managing grants around subjectively valuable applications. Some of these grants programs have done their best to retroactively reward small dapp dev teams based on usage metrics, however this approach still leaves centralized coordinators in charge of the weight function determining which metrics matter, and does not grant users the ability to participate in these reward decisions themselves. Quadratic public goods funding tools have also tried to fill this gap in the Ethereum ecosystem, but again this entrusts decision making power on the select few who care enough to scroll through a long list of projects and spend some of their cash to help sustain them, which brings a similar set of biases that grants programs witness. If only they could vote on which applications they found useful, as they were using them.

When the iPhone first launched in 2007, the world was introduced to a totally new kind of device and operating system. Instead of the old model where the OS was simply a platform atop which applications could run, iOS took it one step further by including an App Store on every single device, and making it difficult to onboard without supplying a credit card up-front.

I expect most modern iPhone users don’t appreciate that this was a significant departure from the existing model in the market: the most popular operating system at the time (Windows) didn’t worry about how software was distributed or monetized. They expected applications to figure it out themselves using the help of third party retailers. This worked well for enterprises that were large enough distribute their apps via CDs in magazines and catalogs and physical stores. But it left the remainder of smaller developers to upload their software to public file directories online, without a real method of monetizing themselves.

Apple on the other hand realized that if they wanted to create network effects around their device, they would need to ensure that indie app developers could create sustainable businesses while building there. Apple was less concerned with individual apps growing as large as possible (if anything this would decrease the supplier power that Apple held), and instead focused intensely on the long tail of devs, many of these teams passing the Two Pizza Rule.

In the years that followed, a burgeoning mobile app ecosystem was suddenly born. Apple offered an improved developer experience by bundling an interface builder and simulator directly into their XCode IDE. For people trying to make money with their apps, Apple’s developer program was low friction for new developers to join at just $99 per year, compared to expensive and time consuming carrier certifications that other mobile operating systems required. Apple also offered native app distribution on every single device via their app store, consolidating the fragmented carrier app store market. They also made it hard (but not impossible) to activate a new iPhone without an Apple ID that had a credit card attached to it.

The result was that if you were in the Bay Area during the late 2000s/early 2010s, chances are high that you were getting pitched on some new mobile app that was about to take over the world. While many of these pitches sounded like fantasies, they were grounded in the very real truth that early mobile app developers were suddenly earning millions in revenue without needing to worry about the biz-ops logistics normally involved in running million-dollar-run-rate businesses.

These decisions were critical: not only to drive stickiness around the iOS platform, but more importantly to explore the full design space of mobile device capabilities that they opened up. The iPhone delivered image/video/audio input and output, powerful gyros, multitouch, a magnetometer, and a haptic motor, all while connecting to bluetooth/gps/cellular/wifi signals nearby. Apple’s decision to help devs monetize meant that if a single individual had an idea for how to combine these functionalities in a new an interesting way, they could feasibly expect to make money if it created value for users. This changed the fundamental market structure for distributing software applications, helping the plethora of long tail devs explore this green field territory. Instead of only having large companies exploring this design space, who have a certain aversion to launching risky experimental products, iOS would now have the maximum number of shots on goal, even more spaghetti noodles thrown at the wall, even more miners digging for gold. While the mobile app market has now ossified and consolidated over the years, this was driven by the market having exhaustively explored the art of the possible, not due to a risk aversion from unsustainable business models.

Looking back at the crypto space, we see that very few ecosystems have quite managed to become hubs which support a large number of small developers building businesses for themselves. If you survey the long tail of devs building on various chains today, they’ll tell you directly that an ecosystem’s ability to help them distribute their new dapp is the #1 factor driving their decision of where to build. Two ecosystems which are currently doing this well are Base and Solana, both of which are happy to confer status on small devs via social spotlighting. Nonetheless this personal process doesn’t scale nearly as well as an app store does, which can automatically surface interesting new use-cases that the chain’s users should be focusing on.

On the business model side, I also believe that by not including some kind of developer fee at the gas layer, the EVM and other execution environments have made a critical error whose costs are being borne in the market today. I believe that we’ve made it far too difficult for smart contract devs to charge fees for their usage, forcing clunky one-off solutions like taking a cut of transaction volumes or rate spreads, which works for financial apps but breaks as soon as a given transaction isn’t immediately profitable. This is why the well-validated crypto markets I mention above are all financial applications dealing with valuable transactions, not simply getting utility from the chain.

The closest we’ve come to a truly user-driven business model for dapps has come from smaller ecosystems like Fantom and EVMOS which let users earn a 15-50% share of the gas fees their dapps generated (the fee % still being fixed across the ecosystem vs varying per-application). Newer ecosystems like Sonic are now coming online, offering a similar model that pays up to 90% of network fees. These ecosystems realize that the ideal system isn’t one where dapp development rewards need to be socialized among network participants.

As a thought experiment, take for example the SafeMath solidity library built by the OpenZeppelin team, which is used in the vast majority of EVM smart contracts today. Based on the original vision for Ethereum dapps, this library would have been immutably deployed once by the team, and new smart contracts would link to the specific address where it was deployed. If the protocol had natively allowed for OpenZeppelin to add an additional line of code that sends some amount of gas units tipped to a given developer address, they could have monetized their work natively wherever it’s used. This also would have incentivized the Ethereum community to coordinate around their canonical implementation, which would have increased visibility and stickiness of their code, improving the overall security for teams which use it. Instead libraries ended up being copied in-line, and the tooling developed away from this pattern. Today we have tens of millions of duplicate copies of SafeMath being stored onchain. OpenZeppelin has thankfully been rewarded for their hard work by the aforementioned grants programs, but one can’t help but think that this is not an ideal state of the world.

In the absence of simple business models for dapps, we’ve seen devs fall back and rely on another kind of business model that crypto offers us: launching a token. No matter how easy launchpad UIs make creating a new token, anyone who’s ever done it can attest it’s a high-fixed-cost activity that is not for the faint of heart. Moreover, having tokens be the primary business model for dapps has meant that otherwise simple modules are forced to imagine some grander vision, where they’re suddenly evolving to become a fractal ecosystem of their own. Not everything has to be that serious!

I think any ecosystem trying to cultivate a diverse garden of applications should think critically about how they can lower the barriers to monetizing their dapps to be as low as humanly possible. Young teams feel like they can’t win building dapps in crypto without raising large sums of money up front to fund development and distribution costs. Things do not need to be this way — this is why there are so many success stories of outsiders building successful companies in the mobile app era. I believe that ecosystems which lower those costs to near-zero for teams going zero-to-one will win in the long run, because they will offer a sustainable way for a long tail of small devs to come online and join the dapp dev market. The terminal state of “play to earn” will be “dev to earn” where instead of clicking a button every day to claim tokens, users click a few extra buttons, prompting their favorite software-generating LLM to deploy a new game they’ve just imagined, which immediately gets distributed to a global audience.

By this point we must recognize that market forces clearly incentivize large applications which fill blockspace to eventually graduate from their native chain and become an ecosystem of their own. If ecosystems want to fight these forces without verticalizing and operating applications of their own, they must create systems which allow small dapp devs to stay lean and build cool stuff without the overhead that forces them on a growth trajectory which many don’t want in the first place. Ecosystems can bear some of the high fixed-costs of distribution and education, leaving dapps themselves to specialize on product delivery. There is symbiotic relationship waiting to be formed, a ruleset which figures out how to get devs paid for the dapps they launch, in a way that’s frictionless enough for the end user that they’re happy to do it. Time will tell which ecosystem will truly get this right.