In many ways, each Uniswap release has been a landmark and a signal for the progress of the application layer.

But, of course, crypto applications have been around long before Uniswap. They first started as application-specific L1s - the first useful one I can recall is Namecoin in 2011. The most interesting ones were the Graphene cousins - BitShares and Steem. Before Ethereum went live, BitShares offered a DEX, user-issued assets which included everything from memecoins and algostables, and introduced infrastructure innovations like proof-of-stake with delegations, low latency and high throughput. Its spiritual successor, Steem, used much of the same tech, but added decentralized social networking where all content was stored on chain. Needless to say, these came with significant compromises, requiring very powerful systems and leading to centralization, with very few unsubsidized full nodes. Without economic sustainability, these app-specific chains become easy and cheap to attack. Indeed, Steem underwent a hostile takeover by Justin Sun and co, and remains to this day, 3 years later, under a 67% attack - with no users running nodes there was no recourse. The only option was forking to a new chain, which happened with Hive - but the community was forever fragmented and splintered, and remains a shell of its former self. Steem was #3 in mid-2016, after only Bitcoin and Ethereum, while BitShares was a top 10 project for many years. Both of these, and pretty much all app-specific L1s, are irrelevant today.

I digress, but there’s an important lesson in how this relates to applications - it’s extremely difficult to run a financial-focused app-specific L1 long term, over years and decades, while remaining economically secure, sustainable, and liquid. Very easy to run L1s in a bull market, or even a bear market following it. But to do it over multiple bear markets, and the inevitable secular bear market - nigh impossible. There may certainly be specific non-financial apps that may be suited to a modern ZK-L1, that can do without economic sustainability, but we know the vast majority of economic value in this space is driven by financial apps.

Beyond inheriting economic security and decentralization armour, the other key benefits of smart contract L1 is as a wellspring of liquidity. The early app-specific L1s showed a glimpse of what’s possible, which no doubt inspired the list of potential applications discussed in the 2014 Ethereum whitepaper. But it was Ethereum that introduced the first wave of application layer innovation.

Thus far, we have seen two key waves in crypto - the first one with Bitcoin & Bitcoin killers, the second one as app-specific L1s that experiment beyond P2P money. The biggest wave came with the 2016-18 ICO bubble, where the application layer threw the kitchen sink and more at blockchains.

Yet, the original 2014 Ethereum whitepaper proved prescient - 99% of the ICO projects proved to be pointless, and once the dust settled, the building continued on applications that actually made sense.

While Uniswap did not go the ICO route - a lot of the initial work was funded by the EF AFAIK - it was very much a product of that wave. Uniswap V1 was the prime example, alongside the likes of Maker, Aave, ENS, Cryptopunks, Compound etc, emerging with an awesome MVP. It proved that the 2014 Ethereum whitepaper, and all the discussions in 2014-16 were correct - certain classes of applications did make sense and will see product-market fit.

The next wave - fourth wave by my counting - was about getting to a rounded, finished product that was ready to settle billions of dollars. This is what we saw in 2020. Uniswap V2 was once again a flagbearer that set things in motion. The $COMP airdop in June 2020 was the catalyst that led to the explosion that we know as DeFi Summer, and broad adoption of DeFi apps. I’ll note that by “broad adoption”, I mean by economic value. So even though there weren’t that many users in this wave, significant value started using these protocols. You don’t need the masses for a DeFi protocol (or crypto’s dominant usecase - alternative reserve asset) to be successful - just the top 1% entities/institutions/HNWIs who are responsible for 99% of the economic activity. Of course, the dynamic is very different for non-financial applications, but more on that later.

Having proved product-market fit, the fifth wave kind of took two paths - efficiency and experimentation. Protocols like Uniswap V3 and Aave V3 are the perfect example of refining the product for greater efficiencies. Parallel to these, we saw experiments that built on top of these now efficient DeFi staples, as well as offer niche alternatives. As with experiments, most of these did not work, but some did. (I’m not naming them because they have highly volatile tokens that I do not want to mention.)

That brings us to what I’ll call the sixth wave - extensibility and maturity. Uniswap V3 is a tremendous protocol and covers >90% of the value (once again, value, not people) for asset exchange. However, there are always niches, and that’s where V4 comes in. Now, I’m not going to talk about V4 - you can read all about it on their blog. The gist of it, though, is that V4 can be pretty much the last major upgrade to the core asset exchange usecase, and the niches can be satisfied by extensions around it with Hooks.

Once these applications reach a mature and extensible state, I armchair-speculate there’ll be a final seventh wave. This will be focused on user experience and adapting to new infrastructure. For example, with Uniswap V5 I see a protocol which is mostly the same as V4 under-the-hood, but with new mechanisms to share liquidity across rollups and make cross-rollup swaps seamless. It’ll also support new UX paradigms like account abstraction and smart wallets natively. At the end of this wave, I expect applications to have “good enough” UX and complete functionality. To be clear, whether there’ll be “mass adoption” remains to be seen. Current evidence suggests tradfi fintech UX is at least a decade ahead of crypto and improving at a faster pace than crypto, but we shall see. The best example is payments itself - a decade ago, there was a genuine usecase for stablecoin payments. But fintech has significantly out-innovated crypto in this time, and today we have inter-compatible payment apps in most of Asia that have instant, free transactions easily accessible to all, with basically near-perfect UX, with the rest of the world not far behind.

To be clear, these waves are clearly illustrative, and for entertainment purposes. In reality, we’ll continue to have app-specific L1s, innovation with new apps, new efficiencies discovered, and above all, even the mature protocols like Uniswap V4 will continue seeing incremental upgrades. But we’re today at a stage where the 2014 Ethereum whitepaper, and all of its usecases have been delivered, and we are entering a maturity stage. Particularly the core financial applications like Uniswap, Aave, Maker, alternative store-of-value (BTC, ETH) etc. that’ll likely inherently account for most of the economic value of the crypto space.

But there’s one last thing that could lead to a new, parallel wave of innovation - fractal scaling. So far, we have basically had just two types of infrastructure. Standard chains like Ethereum or Bitcoin which target accessibility of running unsubsidized full nodes. Or fast chains which are variants of standard chains with higher system requirements - like BSC, Solana or Arbitrum One. These fast chains have their limits, though, offering roughly only a 10x increase in throughput, or at most 100x in their endgame states. We’ve had fast chains for a decade now, and even fast smart contract chains since 2017. They have certainly improved over time, but yet, in the last 6 years (an absolute eternity in the internet age - e.g. ChatGPT gained hundreds of millions of users and thousands of developers within months), all we have seen are variations and remixes of the same applications available on the standard chains since 2019/20.

Fractal scaling (or whatever you want to call it) finally brings a new paradigm when you can have multiple fast chains. So, now, you can have 1,000 fast chains that can interoperate and intercompose. With fractal scaling, the infrastructure layer will finally offer a new paradigm for the first time since Ethereum introduced smart contracts in 2015. For certain types of applications, like at-scale onchain games and social networks, thousands of chains is basically a pre-requisite - just like their web2 counterparts need thousands (or more) of servers. Whether they achieve enough product-market fit to saturate thousands of rollups remains to be seen, and based on available evidence skepticism is well warranted. But we must try! As a final word, this is certainly a long game that’ll take several years, but the pieces are all in place. Fractal scaling also offers a glimpse of novel applications beyond those mentioned in the 2014 Ethereum whitepaper. Of course, it’s not just novel applications, but also existing applications adopting the new infrastructure - as I mentioned with my Uniswap V5 speculation. I’ll note, though, that financial applications do not need this fractal scaling type of throughput - most usecases will be adequately satisfied by a few fast chains, and because of their financial nature and benefiting greatly from economic security and liquidity I’d expect these to be mostly Ethereum L2s.

Tl;dr: Uniswap and the application layer has made steady progress over the last decade or so, and is approaching its maturity stage. As infrastructure matures, so will the user experience on these applications. We also have the potential for innovation returning to the app layer with fractal scaling - remains to be seen whether said potential will be fulfilled.