October 11, 2019
One Million Devs on Ethereum:
The Architecture of Web3 and How We Get There
Joseph Lubin––Devcon V, Osaka, Japan––October 10, 2019
Hi, everyone. I want to speak today about what I think will be the broad architecture of the decentralized World Wide Web and how we will get there. To give away the punchline: the critical element is you.
Today’s Web2 world consists of siloed walled gardens. There really isn’t much of an open web anymore. Google’s front end serves and monetizes everyone else’s content. Facebook, Twitter, and a bunch of other gardens have trapped us all behind their walls.
It didn’t start that way. The Internet started with the promise of a certain style of decentralization. It was about building robust network architectures such that damage to one part of the network would not render other parts unusable. On packet-switched networks, as opposed to circuit-switched networks, messages could be more easily dynamically routed around the damage.
The socially aware technologists soon framed this characteristic of decentralization in the context of political control. John Gilmore, an early famous cypherpunk, stated that “The Net interprets censorship as damage and routes around it.”
The Internet started with the promise of a certain style of decentralization.
These early cypherpunks were the godparents of all of us here today.
On today’s Web 2.0, we give away our personal information and our power in order to exist on the Internet. For most cybercitizens, it would be debilitating to be deprived of Gmail, Google search, Google docs, Twitter, Amazon’s services, Apple’s ecosystem, and the various Facebook properties.
The Web is broken — or incomplete — because it has no native constructs for identity or money.
Without a native construct for money, the web turned to advertising as its core business model.
Without a native construct for secure and private identity, the web grew exploitative before we could recognize that an advertising technology cancer was growing and killing the patient.
With deep learning techniques applied to monetizing our personal information and our attention and addicting us to clicking and scrolling incessantly on social media, the technology that should be purely empowering us has also become harmful to many individuals, and a weapon of mass manipulation targeted at our societies and nation states.
Having identified some nasty problems in the command and control architecture of the web, many of us, at this conference and elsewhere, have come up with some promising solutions. These solutions are not yet mature, but they point the way to a web and societal future that I think we can all be optimistic about.
This is why 4,000 of you flew to the kitchen of Japan to celebrate and extend the current capabilities of Ethereum.
The promising solutions involve decentralized protocol technology. These solutions will soon employ zero knowledge techniques pervasively and systematically, so only what must be disclosed in a given situation will be shared — whether you are demonstrating that you are of legal age to walk into a bar (without having to also reveal your home address), or demonstrating that you are capable of shipping a certain quantity of grain without revealing your company name.
These solutions involve tokenized, incentivized collaboration mechanisms. Instead of a world of companies that serve customers in a somewhat adversarial relationship, we will soon conduct our lives on networks for commerce and leisure. These networks will bring together many actors in diverse roles and motivate activity with tokenized mechanism design such that the collective collaborative behavior that emerges on each platform serves all the people who operate on and jointly own and govern the platform. As the platform appreciates in value, all of the owners playing different roles on the platform benefit, as well.
These solutions will see us establish our own self-sovereign identities where we control the roots of our IDs through DIDs — Decentralized Identities — as specified by the Decentralized Identity Foundation and realized in the form of uPort, Civil ID, and other products. Verifiable Claims, as specified by the W3C, will enable us to lodge cryptographic attestations from our DIDs onto DIDs representing other people regarding characteristics like their age, actions like the fact that they paid their utility bill, achievements as in educational attainment, and performance as with credit scores or developer proficiency.
Perhaps most importantly, Web3 will be more user-centric than any previous Internet platform, putting the user, finally, at the center of user experience. People must be in control of the root of their identities so that their identity can’t be stolen, and in control of their own personal information so it can’t be exploited. Even in the extreme circumstance where a refugee is deplatformed from their country, and not able to take any physical possessions with them to their new home, they should be able to easily access their decentralized identity and reclaim their persistent portable personal history and reputation in the form of verifiable claims linked to that DID.
Web3 will be more user-centric than any previous Internet platform, putting the user, finally, at the center of user experience.Joe Lubin, Co-founder of Ethereum and Founder of ConsenSys
Under the Web3 paradigm, all users will have real agency, social, political, and economic, on the many collaboration networks on which we will all soon work and live. This will be the new fabric of society, and this fabric will support far broader participation in governance and broader wealth distribution.
We are moving rapidly — though it can seem agonizingly slowly at times — towards realizing the Web3 world. It seems slow to those looking in because they hear about the potential and they expect it to materialize immediately. As though the web and Internet, mobile phones and cars were all birthed as mature products, with fully formed infrastructure, by their original inventors. As though the global electrical grid was spun by Edison, and 787s and F35s were waiting in the wings for testing, on the runways, at the international airport, at Kitty Hawk.
Realizing a well-formed Web3 world will involve us achieving improvements in three major dimensions: Privacy and confidentiality, Scalability, and Usability.
Privacy and confidentiality are moving fast with various explorations of zero knowledge techniques, state channels mechanisms, and layer 2 private chain architectures.
We have seen and will continue to see massive improvements in Scalability. This is also moving very fast with various kinds of layer 2 solutions being built, launched and anchored into Ethereum layer 1 for additional security. Skale Networks is one exciting example. Again, lots of use here of off-chain computation, using interactive computation games and zero knowledge techniques, state channels mechanisms, and layer 2 public and private chain architectures.
We are seeing tens of thousands of decentralized transactions per second (DTPS) added to Ethereum layer 2 this year already and very soon we will see millions. And probably before the end of 2020 we will see all of this multiplied a thousand-fold as Ethereum phase 0 will be delivered in Q1 and phases 1 and 2 will likely arrive together before the end of the year.
Usability is also seeing much improvement. Few people recall how agonizing it was to access the Internet via a dial-up on a 9600 or 14.4 baud modem. The apps and web pages were simple back then, but it wasn’t uncommon to wait 30 seconds for a page to load. And often you weren’t sure if you were still connected or just loading. In the morning you’d get to your workstation, start your email client downloading, and go for coffee to give it time to synch with the POP or IMAP server before you could read and respond. This was 1993–94. We see similar issues with some early implementations of the Web3 technology.
For Web3, many projects are exploring the “progressive onboarding technique” that enables instant onboarding of a user to a dapp, followed by a gradual handover of responsibility from the dapp to the user as investment of attention and money in the dapp grows.
New wallet architectures are taking the multisignature tooling pioneered by Gnosis Wallet and Safe, and adding social recovery (pioneered by uPort) and social acceptance (or rejection) of intended user actions to ensure that early adopters don’t get nicked by the still-sharp edges of this new technology.
And systems like ENS are gaining traction so soon we won’t have to look at long strings of hexadecimal addresses much anymore, just as we rarely look at the four octets of IP addresses in dot-decimal notation.
Realizing a well-formed Web3 world will involve us achieving improvements in three major dimensions: Privacy and confidentiality, Scalability, and Usability.Joe Lubin, Co-founder of Ethereum and Founder of ConsenSys
The architecture of Web2 doesn’t merely represent personal identity risk for people, and political risk for nation states: it also represents different kinds of platform risk.
There is platform risk to artists and others who find themselves deplatformed from a system they rely on for their livelihood. Facebook, Twitter, and YouTube now have enormous ability to shape narratives, debate, and broad morality. And there is platform risk to startups and other companies who find the rules of a platform can change from underneath them as that platform grows and evolves and finds itself cannibalizing the companies that it drew in as partners a few years earlier. The incessant drive to grow revenue once at the top of its adoption S-curve yields such anti-competitive behavior.
Google, Amazon, Facebook, LinkedIn, Apple, and Microsoft are all large platforms that have engaged in dining on their dependents. As a cybercitizen I love and need Google and Amazon, but the unfair advantages that accrue to Amazon in its e-commerce activities and to the AWS cloud are staggering. As are those that are available to Google.
A new and insidious kind of platform risk has emerged more recently. Just as platforms like Facebook, LinkedIn, and others have eaten the lunch of some startups, clouds are now cannibalizing the profits of companies built around open source projects.
- AWS has been accused for a while of re-writing, re-using, and re-branding products without supporting the underlying open source and often by eliminating economic benefits to the for-profit companies that have commercialized the open source projects.
- Many would argue that this sort of behavior is acceptable under open source licensing. Others would argue that there is an implicit social contract and that companies like Amazon are not giving back when they capture value from the open source community.
This kind of anti-competitive behavior is enabled by vertical integration. I believe the Web3 world will be largely horizontally layered with explicit interface promises between layers, or between synergistic components at each layer.
Is this likely to be a problem for our space given that our space is composed of largely open source projects?
I would argue that we have an opportunity — unique to this technology — to construct things differently. Applications running on blockchain networks are likely to be subject to different economics: cryptoeconomics, of course. The best blockchain systems are explicit about what the costs are, and these costs should be paid by the appropriate actor. In some cases it will be the end user. In some, it will be the app publisher, perhaps via the gas station network or other meta-transaction mechanism. Blockchain networks will better account for possible negative externalities and bring them inside the tent, while benefiting naturally and massively from positive externalities like network effects.
All participants on these networks will be doing so from a position of personal or corporate agency. Many will be compensated themselves for the roles they play on these networks as they, in turn, compensate other people, orgs, and dapps for the value they provide.
Platform risk will be reduced by decentralized governance and the ability to fork open source codebases to launch competing platforms, and even to fork operating platforms when they cease to serve their entire constituency. We will all come to understand that forks are a very healthy and powerful economic tool. Eventually the more advanced platforms will make it easy to fork themselves, copying over the entire state (or building stateless systems that make forking easy), to ensure that those in governance of the system are incentivized to be, and remain, good stewards of the system. Forks effectively keep those “exiting” within the broader tent, enabling the extended community to define itself in broader and more varied terms. John Gilmore would be a huge fan of the forking mechanism.
I am powerfully focused on maximal decentralization, permissionlessness, the base trust layer that is Ethereum, and its potential to be a global settlement layer — largely devoid of platform risk — for natively digital assets or digitized real world assets. But I think it is short-sighted when some people believe that all blockchain systems must be fully permissionless and maximally decentralized.
When you are issuing valuable assets or anchoring into a base trust layer, that layer must be maximally decentralized and that requires it to be completely permissionless. But adding some decentralization to use cases that are enacted among a small set of counterparties is enormously valuable, and this is best done on a private permissioned system.
A commodity trade finance platform, for instance, linking banks and commodities traders, does not need radical decentralization and permissionlessness in 2019 in order to be an order of magnitude or more better platform for conducting trade amongst those parties.
Come 2020, it may make sense to link that commodity trade finance platform laterally into a commodities trading platform, and into a payments platform, and into a corporate identity platform, and into an insurance platform, etc. These networks can be linked to one another laterally using bridges or using the Liquality technology which can enable unintermediated trustless atomic swaps or message passing across network boundaries. Liquality already links Ethereum and Bitcoin trustlessly.
And by 2021 it may make sense to anchor all of those platforms down into the Ethereum base trust layer for increased security without sacrificing scalability, and also so that base layer can serve as a trustworthy message and transaction bus between a wide variety of applications.
This suggests that the architecture of the decentralized world wide web will look like a patchwork of blockchains — some possibly private and permissioned — linked to one another and anchored down into the permissionless and maximally decentralized base trust layer using a variety of techniques. Ethereum 2.0 is currently the only candidate to be that base trust layer as it is the only blockchain project that is assiduously focused on maximal decentralization in all dimensions.
The architecture of the decentralized world wide web will look like a patchwork of blockchains — some possibly private and permissioned — linked to one another and anchored down into the permissionless and maximally decentralized base trust layer.Joe Lubin, Co-founder of Ethereum and Founder of ConsenSys
This Internet of blockchain architecture rhymes nicely with the fact that the Internet is an Internet of networks. And finally our planet will have an automated, objective trust foundation to replace the subjective trust infrastructure we have built on for millenia.
To continually address scalability concerns, we must practice minimum viable decentralization as we stand up new permissioned Ethereum networks. Going forward we will always have scalability concerns and there is no need to over-secure every use case. We will get used to designing systems with appropriate levels of decentralization.
We should start to think of layer 2 blockchain networks as logical constructs, because in time we will want to lift these and set them back down on more capable physical architectures. We need to start designing for such upgrades. Eventually most of the world’s private permissioned systems will run on public permissionless infrastructure, just as corporate and government IT infrastructures today run largely on cloud, just because it will be the best substrate on which to run these systems.
It is critical that this base layer be permissionless and maximally decentralized because that is where we derive trust. And that trust layer must be global so that we can build international financial and trade systems that all countries can collaborate on.
What makes Ethereum the current best or only option for a base trust layer for the planet?
On a recent “Into the Ether” podcast, Matteo Leibowitz outlined what he called the “Distribution Quadrilemma.” His argument was that since regulators have woken up, any post-2015 protocol that wishes to achieve widespread adoption has to succeed with respect to the following four pillars:
- The token must be issued broadly and equitably.
- The community must remain engaged, vibrant, and grow large over time. One way to achieve this is by enabling the tokens held by the community to appreciate in value. As a side note: the community must also attract the best and the brightest developers and entrepreneurs, in large quantities.
- The project must bring in sufficient capital so it can be delivered, maintained, and continuously improved.
- And finally, and this is the new wrinkle: The project must be regulatorily compliant.
Among all the competitors, the Ethereum-killers, to have emerged in the past few years after Bitcoin and Ethereum, none show convincing signs of meeting all four requirements of the quadrilemma. The best way to build and grow a strong community is to sell them a token and promise that it will appreciate and they work to make that so. This is the definition of a security. Since you are now selling a security, your token won’t be issued broadly and widely, and your community won’t be robust. There will be novel mechanisms tried over time, but so far no projects have gained great traction, and it is hard to imagine building the world’s trust foundation and global settlement layer on a VC Coin project.
The SEC recently gave the EOS project a slap on the wrist for their token sale, but this does not mean a project structured like EOS could have a successful raise at this point. The SEC is tooled up now and very well aware. Today, it would shut down a sale structured like the EOS sale before it was able to turn on its first billboard in Times Square.
Currently only Ethereum is sufficiently mature, decentralized, and programmable to accommodate the needs of today’s infrastructure, industries, and individuals, with the momentum and roadmap to support the growth of all these longer term visions.
So, instead of the vertical walled gardens of web 2.0, we will have horizontal decentralized collaboration layers.
At the base we will have Ethereum for trusted transactions and guaranteed execution of agreements, linked to protocols for decentralized storage, bandwidth, heavy compute, identity, proof of location and others.
On top of that, we will have the open decentralized financial plumbing layer for the emerging global economy. This will consist of price stable tokens, borrowing and lending networks, decentralized exchanges for liquidity, tokenized assets in all forms, other financial instruments like derivatives and synthetics, payments systems, subscription systems, and staking and validation systems. Even governance systems will partially lie in this financial plumbing layer for certain kinds of networks and tokens.
Many of the blockchain systems we will build going forward will be enabled by, and enhanced by an increasingly capable and sophisticated financial layer. For instance, the IT industry is becoming more granular and commoditized. Trusted compute, heavy compute, storage, and bandwidth, will all be distributed on cloud and fog infrastructure. We will negotiate, provision, and pay for these services with tokens in real time. Providers and users will need natively digital financial instruments to hedge their pricing and other risks.
The base trust layer enables the financial layer. And upon that financial plumbing layer, the rest of the emerging global digital economy will be built.
It is an amazing time to be active in the technology space, and great to be a developer at a time when the nerds are about to inherit the earth to an extent far greater than in previous technological eras.
There will be a bit of a tug-of-war over the next few years. Web2 antipatterns won’t die rapidly and easily and many valuable Web2 constructs should be evolved to benefit Web3. And of course there will be Web2 wolves in Web3 sheep’s clothing.
The nerds are about to inherit the earth.
To achieve rapid technological evolution and societal advancement, we simply need more Web3 devs. We need the best and the brightest from the Web2 world to understand what is now possible under this new paradigm.
There are somewhere north of 30 million software developers globally. That number includes, as rough estimates:
- 600k open source developers
- 12 million Java developers
- 8 million Python developers
- 20 million Apple/iOS developers
- 3 million Microsoft developers
- 15k Linux developers
So, here’s my ask — or rather a challenge — to all of us for the next year: let’s get one million of those developers building with the Ethereum technology.
We can estimate based on Truffle downloads and other metrics that there are as many as 500,000 software developers engaging with Ethereum regularly in some capacity already, and probably 200,000 or more who have full time Ethereum developer jobs or are otherwise fully engaged.
At first, 1 million ETH devs might sound overly ambitious, but I think that with a concerted effort we can get there in a year or so. Ethereum has solid developer mindshare and momentum already. We have an incredibly talented range of people and teams contributing their time and efforts to improving tools and resources. And we have some of the most fascinating, bleeding edge software technology in the world. We’ve all decided to redirect our careers, our days, our nights and weekends, to building this technology together, and we’ve come all this way in just a few short years. From the start of the Ethereum project this has all seemed inevitable to me, though it is astonishing how fast it has moved. But for all of those looking in and waiting on us, let’s get this thing in a higher gear.
This is my challenge to all of us: one million Ethereum developers. By Devcon VI perhaps? Maybe that is a little too ambitious, but we will give it a shot.
Over the next days and weeks, please reach out to your developer friends who are not yet in the Ethereum space. We are putting together several projects to support this thrust. The website onemilliondevs.com will house multiple ways to introduce your dev friends to many of the “holy shit!” moments that blew your minds when you learned about the space.
Kevin Owocki and the Gitcoin team have put together a game that will be available extremely soon to highlight many of these insights.
Billy Luedtke, Kames Cox-Geraghty, and Joe Burnitt of the RAPID team have put together a quests system to help your dev friends set out on holy-shit-moment learning odysseys. There will be badges, kudos, and a scoring system. If you bring in 20 devs and they start racking up kudos and points, you will find yourself rising on the leaderboard, as they will all represent members of your team.
By the way, staking out a position in the top ten on the leaderboard will make you super-hirable in a developer relations — devrel — role in our ecosystem. ConsenSys and others find people with those skills to be enormously valuable. The onemilliondevs.com site will reveal much more fun mechanisms over the next few days. And we will grow it over time. Please check it out and send your friends.