Once upon a time—that is, in 2011—the company then known as Facebook launched Open Graph, a way for users to carry their Facebook identities from app to app, thereby enabling developers to give them a personalized experience wherever they went. Rather than recreating an identity each time you went to a new app, you could leverage the identity you had already built through Facebook and—at least in theory—the sum total of Open Graph developers leveraging the platform.
Sub “crypto wallet” for “Facebook account” and the goals of Web3 are shockingly similar. Both Open Graph and the decentralized web constitute attempts to address the same deep gaps around identity and memory in the internet as it was originally conceived. Both aim to enable portable identity across digital ecosystems, including data, assets and relationships. And hard as it may be to believe given the tenor of today’s discourse, even the idea of personal data sovereignty that’s so central to the crypto world was baked into the Open Graph vision.
Despite the company’s high-minded aspirations, Facebook’s Open Graph efforts didn’t work. This was for a handful of reasons, some of which could have been anticipated, others of which were more subtle, and none of which had to do with the fact that Facebook was a centralized entity. That’s not to say that the difference between a single company providing this cross-platform service versus a decentralized, blockchain-based approach isn’t consequential—just that it’s not determinative on its own. The new crop of technologists building Web3 still has a lot to learn from the problems with Facebook’s Open Graph.
1) It’s far harder to build compelling cross-app social experiences than you might think.
In the days of Open Graph, the vision held that all the applications you use would be better because they were “social.” The problem was, most developers couldn’t figure out how to make that vision a reality. Sure, knowing what your friends were reading, watching or listening to was marginally useful at times. But the prime thesis that developers could build better apps by leveraging Open Graph and making them social didn’t really pan out.
How does that apply to Web3? I think it will be harder to find compelling use cases for crypto identities and assets than you think.
Decentralized finance is a clear and specific use case that makes sense for crypto: You hold one wallet with all your assets, and through that you can access many different services and buy and exchange tokens using many different products.
But when you get into non-financial use cases, there are real questions about what will be truly captivating. The idea that non-fungible token communities like CryptoPunks will move with people between worlds is cool in theory, but will it be that great in practice? Having digital skins, tools and resources that you can bring with you from world to world is an enticing vision, but figuring out how those assets have value and interest across those worlds is still very unclear. I am an optimist and think there can be answers, but it would be a mistake to think they’re obvious. It all seems like a heavy lift.
2) There was ample incentive for apps to read data from Open Graph, but few reasons for apps to write data back.
Key to Facebook’s original idea was the notion that new apps had a cold-start problem, which would be solved if people could bring their friends, their preferences and their data with them to new apps. Programs would read that information on Open Graph and use it to create better user experiences. Then they would write new data to the graph based on the user’s in-app behavior. Other programs could use that data to improve users’ experiences, and so on and so forth, until we achieved a deeply personalized web.
The advantages of reading Open Graph information were reasonably clear: Any incremental improvement in experience you can provide is worth a shot, as is the growth potential of tapping into Facebook’s vast relationship network.
But logically, developers never had enough of an incentive to write information back to the graph. The core information economics of the system didn’t support it. If you build information about a user’s preferences, interests or relationships that you can use to make your own product better, why would you ever share that back to Facebook, let alone in a way that other developers would then be able to use to improve their competing projects? It just didn’t make sense.
For a while Facebook tried to patch this obvious issue by offering free distribution in exchange for writing data back into Open Graph: For instance, you collect data on the movies users like and share it with Facebook, and Facebook will distribute your app in newsfeeds, helping you find new users. Ultimately that trade wasn’t sustainable—it was a subsidy in an information economy that never fully worked.
What do these information-economic woes mean for crypto’s Web3 attempt? First, while platforms and metaverses will have ample reason to import some external data, such as communities, it isn’t clear they’ll want you to be able to import everything. For a long time the financial model for many immersive games has been to sell you in-game assets, so it’s pretty inconceivable that they would let you just import whatever items you want from other unaffiliated games without monetizing them in some way.
Not only that, once you’ve purchased assets in one game, it’s not at all certain that you’ll be able to take them with you when you leave. No question that, in the short term, if you can mint an NFT in one ’verse that would be useful in other ’verses, said asset might be more valuable to players and stimulate further minting. But for that to be worthwhile in the long term, other games would need to accept those items—which, as we’ve already covered, seems questionable.
It’s possible that the Web3 world will have the same information economy problem as Open Graph, but in reverse. Platforms and ’verses will be more than happy to allow users to export assets, but will have a far harder time allowing users to import whatever they want, thereby stymying the digital asset supply chain.
Ultimately, Web3 platforms are going to look a lot like nation states, which means they’ll have to be subject to free trade agreements governing data and assets moving between them. Part of the dream of Web3 is to create systems that enforce their own rules—but history tells us that free trade is a delicate, difficult-to-balance equilibrium, especially without a hegemon like the U.S. or England to police it.
3) User trust is very difficult to maintain.
One of the core realities of building internet properties at scale is that, by definition, the average user IQ is 100—that is to say, average. Within that user base, though, are people of vastly different ages, temperaments and levels of digital fluency.
There is a narrative that has long persisted about Facebook (both the company now known as Meta Platforms and the app specifically) that it has a big trust problem. To my mind, at least, the reason for that is largely because issues of identity, data and portability are tricky for average people to wrap their heads around. No matter how straightforward you make the platform or how transparent you make the rules, they still have a hard time understanding how data moves around the internet.
The core ethos of crypto is supposed to fix this problem by giving users clear access to and sovereignty over their data. I hope it works and scales in a way that billions of users can trust, but it’s going to be a challenge because the fundamental user problem persists—understanding how identity and data portability work in digital space is complicated and sometimes surprising.
If anything, one of Facebook’s biggest mistakes over the years—with Open Graph and beyond—was taking a paternalistic stance, bending over backward to try to protect users. In taking this approach, I fear the product teams wrote a series of checks that were effectively impossible to cash.
Crypto’s approach is the opposite—completely rules-based laissez faire, which fits with my philosophy of how to build sustainable good platforms. But at the same time, I don’t think you can overlook the fact that an open platform, even with these properties, will be used in unintended ways. Users will have surprises and bad experiences. People will be hurt and alarmed by the data about them that winds up circulating on blockchains.
Hard-core crypto people will object to this framing and say that zk-snarks (aka zero-knowledge succinct non-interactive arguments of knowledge, a way to prove that you know something without anyone else having to confirm it) and next-generation privacy-preserving technologies will fix all this. Those might help in some Web3 niches, though they also carry their own challenges. But it’s unrealistic to believe that in the vast technological universe, the Web3 world won’t face similar trust challenges with average users.
The Twist: What About Decentralization?
The many-trillion-dollar question is whether decentralization is really the secret ingredient that will make Web3 rise where Open Graph fell flat.
The argument in favor of decentralization as the key is the internet itself. It took a huge group of people to all agree to certain protocols and rules of engagement to bring the internet to life. Then the fledgling thing had to endure repeated assaults from large corporations like AOL and Apple marshaling huge resources to subsume and kill it. Governments looking to control their own domestic information space are still trying to restrict its spread.
Yet the internet abides, and for two reasons. First, its open framework has allowed for innovation and growth to proceed at maximum speed, superseding all other priorities. Closed systems might be higher quality and safer, but they just wouldn’t be able to keep up. Second, the internet worked because the U.S. and other world powers defended it at key moments out of a deep belief in free speech and democracy. If Web3 (which is really internet 2.0) succeeds, it will be because it followed a new and improved but fundamentally similar path.
If Web3 turns out not to be the solution, it will be because it doesn’t actually solve any of the core user-oriented challenges of an open web platform. Consumers don’t understand or care about decentralization and data sovereignty. Technologists like those things, but typical people care about functionality and experience. If closed systems are able to provide those better than an open platform, we’ll have a series of siloed immersive digital experiences that basically looks like the app world of today.
I bank on Web3 making it. For one, it’s the future I personally want to see. But I also hope that the U.S. and other democratic powers will wake up quickly to the fact that China has captured the closed future. Our best strategy against the dominance of highly centralized systems is to once again take up the mantle of freedom and openness as we have done in the past.
That said, if we want Web3 to work, we have to learn from prior attempts at similar ends and understand the true challenges that lie ahead.