are we doomed to repeat web2 history?
gm moxie, and some reflections on his web3 hello-world
Moxie's toe-dipping web3 post (as usual) was open-minded, uncomfortably insightful and widely read, and his experiments are fun! If you’re either interested in or skeptical of web3, you should read it. Trying to lower my own barrier to public writing and sharing some initial thoughts in reaction.
Do centralized platforms always win due to speed?
Moxie, reflecting on the rise of web2, observes that protocols, requiring decentralized consensus, fundamentally tend towards slower development than dictatorships. And so they'll fail in the face of fast-moving dictatorships that will realize the ecosystem's potential.
Let's extend Moxie's argument that history will repeat itself (web2x2) into two potential ones: that extensive innovation at the protocol layer specifically is a slow fool's errand, and that as with web2 there will be a Slack that always outpaces Ethereum's IRC.
In practice, it isn't clear to me if the "slowness" of protocol evolution is yet holding the ecosystem back. Who's the central monster that will Slack our ETH, and how would they do it in a way users accept? Not happy with ETH? Take your pick of L2s or SOL or what have you. There seems to be a lively, fast-paced fight to build and improve these ecosystems.
More earnestly, what if that "pace" issues of decentralized development can be overcome, and team coordination evolves? What if the future of some decentralized projects looks less like a standards body meeting (failure mode) than an empowered engineering team that has architects, but where individuals and smaller pods can solve many problems without input, and sometimes votes with exits?
Why do we have distributed systems today? If you are a single amazing full-stack engineer, it's faster to manage your monolith. If you're many, it's faster to have a distributed system, because of its modularity, where many people can work on many moving parts.*
Decentralized coordination and decision-making is evolving rapidly and the improving quality of tools for coordination matters. We will certainly see a lot of abandoned ecosystems and creative destruction along the way (RIP sushi), but I am open-minded that this time it could be different, because we see much novel coordination happening.
Is it just about the money, and does that matter?
Web3 is absolutely a gold rush, as Moxie then recognizes. Not only is it a gold rush, it's one with blended incentives. Fun is underestimated. People very badly want to make money (quickly) doing things they think are fun/provide social standing -- play games, gamble, be an influencer, show off their taste, be an engineer solving interesting problems. However, the gold rush doesn't negate other incentives and effects (decentralization, ownership, community). It cloaks, confounds and fuels them.
I'd extend Moxie's you can't stop a gold rush to a gold rush leaves a mining town in its wake. Even if many entrants to crypto "do not fundamentally care about distributed trust models," an attractive alternative to the future moxie paints is one where an increasing user base grows to care, and builders grow to learn to build decentralized things. Is that possible? Purists, come pitchfork me, but what if people are incentivized to care about sovereignty and privacy by money along the way? People can learn new concepts.
And yes -- the web3 ecosystem is an engine and not a leaky bucket, both because of the scale of wealth at risk at this point, but also because of how tied it has become to so many people's social identities, which they may care to build/preserve more than wealth. Open secret about Roblox as an economy: money goes into Robux, mostly doesn’t come out. Small, medium and big earner developers spend big dollars building and preserving reputation by supporting the ecosystem of other developers.
Are we destined for an even less private web?
As Moxie points out, the "privacy" story is mixed today. Default-public writes are, at best, weird. As a security nerd myself, I am excited to see the increasing consumer interest in privacy. At the same time, we should recognize that the vast majority of consumers would still rather outsource responsibility to tech organizations than themselves (excl. meta), and will choose convenience and functionality over privacy and control. Perhaps there is generational change here that I'm missing because I'm old, or perhaps people will not have to make those choices as the decentralized alternatives become sufficiently good and easy.
However, Moxie's concern that there is "less privacy" in web3 doesn't ask the question of, for which customers? Much more privacy, identity management, and pseudonymity is possible if the user is willing to put the effort into it. That’s largely untrue if you’re participating in web2. My early experiments in crypto were all as an anon. Many people maintain split identities with crypto because they don't want to learn in public, or don't want to be cancelled, or a multitude of other reasons. Yes, for even most consumers of crypto, privacy remains a potential, but for some users it's real, and there's plenty more to be built here on the foundation of self-managed identity.
Second, Moxie is largely not commenting on the novel social dynamics of web3 -- decentralized, community ownership is a powerful parallel movement to decentralized infrastructure. It is enabling wild and weird things, from plenty of people who don't care much about managing their own keys but do care about having skin in the game.
Finally, Moxie concludes creativity within the constraints of web3 won’t change our relationship to technology. But, isn’t it already? The argument could be instead that the change is limited in audience, and fleeting.
In conclusion, he suggests two adjustments in course:
I'm all for skinning cryptography for the masses. Crypto has been a huge onramp to individuals “managing their keys” already — and we can make this much a better (transparent) experience. However, it is also useful that some people are willing to run infrastructure, and there are ecosystem-wide benefits. Not everyone needs to run a validator node.
Finally, of course a wholehearted yes to Moxie's suggestion of making more people builders. Enabling easier creation of alternatives can only be a good thing, but we also need incentives, schelling points, and technical interfaces. The blue ocean and privacy for the nerds, the wealth creation and community for many. And I would disagree that making software has become "harder" in our lifetimes. The same software has become much easier to build. We just want more.
What did I miss/get wrong? gm moxie! ☀️
*yes, we also choose these systems for scale and resilience, and distributed systems are a huge PITA to understand. let's not get buried where the analogy doesn't work