Why it’s too early to get excited about Web3

By Tim O’Reilly
December 13, 2021
The Tulip Stairs and lantern at the Queen's House in Greenwich by Inigo Jones.

There’s been a lot of talk about Web3 lately, and as the person who defined “Web 2.0” 17 years ago, I’m often asked to comment. I’ve generally avoided doing so because most prognostications about the future turn out to be wrong. What we can do, though, is to ask ourselves questions that help us see more deeply into the present, the soil in which the future is rooted. As William Gibson famously said, “The future is already here. It’s just not evenly distributed yet.” We can also look at economic and social patterns and cycles, using as a lens the observation ascribed to Mark Twain that “history doesn’t repeat itself, but it rhymes.”

Using those filters, what can we say about Web3?

Learn faster. Dig deeper. See farther.

Join the O'Reilly online learning platform. Get a free trial today and find answers on the fly, or master something new and useful.

Learn more

Decentralization versus centralization

The term Web 3.0 was used in 2006 by Tim Berners-Lee, the creator of the World Wide Web, as a look forward to the next stage of the web beyond Web 2.0. He thought that the “Semantic Web” was going to be central to that evolution. It didn’t turn out that way. Now people make the case that the next generation of the web will be based on crypto.

“Web3” as we think of it today was introduced in 2014 by Gavin Wood, one of the cocreators of Ethereum. Wood’s compact definition of Web3, as he put it in a recent Wired interview, is simple: “Less trust, more truth.”

In making this assertion, Wood was contrasting Web3 with the original internet protocol, whose ethos was perhaps best summed up by Jon Postel’s “robustness principle”: “TCP implementations should follow a general principle of robustness: be conservative in what you do, be liberal in what you accept from others.” This ethos became the foundation of a global decentralized computer network in which no one need be in charge as long as everyone did their best to follow the same protocols and was tolerant of deviations. This system rapidly outcompeted all proprietary networks and changed the world. Unfortunately, time proved that the creators of this system were too idealistic, failing to take into account bad actors and, perhaps more importantly, failing to anticipate the enormous centralization of power that would be made possible by big data, even on top of a decentralized network.

Wood’s point is that the blockchain replaces trust in the good intentions of others with transparency and irrevocability built into the technology. As explained on Ethereum.org:

Cryptographic mechanisms ensure that once transactions are verified as valid and added to the blockchain, they can’t be tampered with later. The same mechanisms also ensure that all transactions are signed and executed with appropriate “permissions” (no one should be able to send digital assets from Alice’s account, except for Alice herself).

Ethereum.org’s documentation continues:

Web2 refers to the version of the internet most of us know today. An internet dominated by companies that provide services in exchange for your personal data. Web3, in the context of Ethereum, refers to decentralized apps that run on the blockchain. These are apps that allow anyone to participate without monetising their personal data.

Crypto enthusiast Sal Delle Palme puts it even more boldly:

We’re witnessing the birth of a new economic system. Its features and tenets are just now being devised and refined in transparent ways by millions of people around the world. Everyone is welcome to participate.

I love the idealism of the Web3 vision, but we’ve been there before. During my career, we have gone through several cycles of decentralization and recentralization. The personal computer decentralized computing by providing a commodity PC architecture that anyone could build and that no one controlled. But Microsoft figured out how to recentralize the industry around a proprietary operating system. Open source software, the internet, and the World Wide Web broke the stranglehold of proprietary software with free software and open protocols, but within a few decades, Google, Amazon, and others had built huge new monopolies founded on big data.

Clayton Christensen generalized this pattern as the law of conservation of attractive profits: “When attractive profits disappear at one stage in the value chain because a product becomes modular and commoditized, the opportunity to earn attractive profits with proprietary products will usually emerge at an adjacent stage.”

Blockchain developers believe that this time they’ve found a structural answer to recentralization, but I tend to doubt it. An interesting question to ask is what the next locus for centralization and control might be. The rapid consolidation of bitcoin mining into a small number of hands by way of lower energy costs for computation indicates one kind of recentralization. There will be others.

The hype cycle

The Ethereum community’s early writings on the topic offer measured assessments of the trade-offs and challenges ahead for Web3, but most popular accounts today are suffused with hype and the glamor of financial speculation. A recent New York Times article provides a case in point:

Venture capitalists are betting billions of dollars to create what in effect is an alternative world of finance, commerce, communications and entertainment on the web that could radically transform major elements of the global economy—all built on the blockchain technology popularized by Bitcoin.

There follows a litany of investments from crypto backer Andreessen Horowitz in areas from gaming to decentralized finance to NFTs to decentralized social networks. None of the examples in the article focus on the utility of what is being created, just the possibility that they will make their investors and creators rich.

And it’s not just mainstream media that’s doing breathless reporting about the money to be made as if the creation of actual value were irrelevant. Stories from those who’ve gone down the “crypto rabbit hole” are eloquent on the subject of access to riches:

One of the great things about crypto is how it democratizes access to investments. For example, people already have easy access to 95 vetted crypto assets through Kraken. If you’re tech savvy enough, you could invest directly in over 1,150 crypto assets worldwide, each with market capitalizations above $10 million (at time of writing)….

In order to gain access to early stage startup investment deals in tech, you’ve traditionally been required to be accredited and connected in Silicon Valley. In theory, the only true barrier to entry in crypto should be awareness….

Repeat after me: neither venture capital investment nor easy access to risky, highly inflated assets predicts lasting success and impact for a particular company or technology. Remember the dot-com boom and the subsequent bust? Legendary investor Charlie Munger of Berkshire Hathaway recently noted that we’re in an “even crazier era than the dot-com era.”

Cryptocurrencies may well be the future of finance, but at the moment it’s hard to see what’s really working, given how much smoke is being blown. Yes, exchanges like Coinbase are making a lot of money, but unlike traditional financial exchanges, what’s being traded isn’t general-purpose money but a speculative asset class that may be wildly overvalued. Nor has blockchain replaced trust in the way that Gavin Wood hoped. Binance, the world’s largest crypto exchange by trading volume, is under investigation for tax fraud and money laundering. One recent headline points out that “[a] small group of insiders is reaping most of the gains on NFTs.” The interface between crypto and existing financial systems is ripe for abuse.

If Web3 is to become a general purpose financial system, or a general system for decentralized trust, it needs to develop robust interfaces with the real world, its legal systems, and the operating economy. The story of ConstitutionDAO illustrates how difficult it is to build bridges between the self-referential world of crypto assets being bought with cryptocurrencies and a working economic system where the Web3 economy is linked to actual ownership or the utility of non-Web 3 assets. If the DAO (decentralized autonomous organization) had succeeded in buying a rare copy of the US constitution at auction, its members wouldn’t have had a legal ownership stake in the actual object or even clear governance rights as to what might happen with it. It would have been owned by an LLC set up by the people who started the project. And when the DAO failed to win the bid, the LLC has had trouble even refunding the money to its backers.

The failure to think through and build interfaces to existing legal and commercial mechanisms is in stark contrast to previous generations of the web, which quickly became a digital shadow of everything in the physical world—people, objects, locations, businesses—with interconnections that made it easy to create economically valuable new services in the existing economy. The easy money to be made speculating on crypto assets seems to have distracted developers and investors from the hard work of building useful real-world services.

This isn’t to say there aren’t real opportunities for Web3 beyond financial speculation.  Crypto is well-suited for digital-only assets that can be valued and used in a self-contained world, like a computer game or the longed-for metaverse. There may be opportunities being worked out in the digital art market and for sports highlights. And as Sal delle Palme put it, “New applications for crypto, such as NFT marketplaces, DAOs, DeFi and DEXs, CeFi, charities, GameFi, DeSo, etc., are being invented, funded (often by the crowd), built, and shipped with blinding speed.” But we’re a long way from the birth of an entirely new economic system.

Of course, crypto and Web3 are only a tiny fraction of today’s speculative excess. Valuations of the startup du jour are also sky-high, and it’s not at all clear that the valuations are accurate measures of actual value being created. They may well just be a con game that benefits a small number of insiders, much like the financial instruments that made so many Wall Streeters rich before nearly crashing the world economy in 2009. So, as Matt Stoller wrote recently, “Web3 is a bunch of bullshit. The problem is, compared to what?” The current economic system is rife with fraud, and is also rigged in favor of insiders! Web3 dreamers like those behind the Celo project are right. We do need a new economic system.

Two kinds of bubble

The Dutch tulip mania from 1634 to 1637 is the classic example of a wild difference between the nominal financial value of an asset class and its intrinsic value. When the bubble popped, tulips went back to being flowers, beautiful but no longer worth a fortune, with no lasting impact on the prosperous Dutch economy. There have been many speculative bubbles since, and most of them have faded into the background noise of history.

There is another kind of bubble, though, identified by economist Carlota Perez in her book Technological Revolutions and Financial Capital. She notes that virtually every past major industrial transformation—the first Industrial Revolution; the age of steam power; the age of steel, electricity, and heavy machinery; the age of automobiles, oil, and mass production; and the internet—was accompanied by a financial bubble.

Perez identifies four stages in each of these 50–60-year innovation cycles. In the first stage, there’s foundational investment in new technology. This gives way to speculative frenzy in which financial capital seeks continued outsized returns in a rapidly evolving market that is beginning to consolidate. After the speculative bubble pops, there’s a period of more-sustained consolidation and market correction (including regulation of excess market power), followed by a mature “golden age” of integration of the new technology into society. Eventually, the technology is sufficiently mature that capital moves elsewhere, funding the next nascent technology revolution, and the cycle repeats.

An important conclusion of Perez’s analysis is that a true technology revolution must be accompanied by the development of substantial new infrastructure. For the first Industrial Revolution, this included canal and road networks; for the second, railways, ports, and postal services; for the third, electrical, water, and distribution networks; for the oil age, interstate highways, airports, refining and distribution capacity, and hotels and motels; for the information age, chip fabs, ubiquitous telecommunications, and data centers.

Much of this infrastructure build-out is funded during the bubble phase. As Perez puts it:

What is perhaps the crucial role of the financial bubble is to facilitate the unavoidable over-investment in the new infrastructures. The nature of these networks is such that they cannot provide enough service to be profitable unless they reach enough coverage for widespread usage. The bubble provides the necessary asset inflation for investors to expect capital gains, even if there are no profits or dividends yet.

And so, there was a canal bubble, a railroad bubble, and, of course, the dot-com bubble, which ended just as Perez was finishing her book. A frenzy of inefficient investment had left behind dark fiber, empty data centers, and a whole lot of talent and know-how that was ready to be reused during the consolidation phase.

In Perez’s narrative, many smaller technology cycles are rolled into one. Consider the history of modern digital computing. It has had several phases, each dominated by a new generation of technology: the mainframe, the PC, the internet and the World Wide Web, the smartphone, and now, perhaps, cryptocurrencies and the metaverse. Each of these has had its own cycle of innovation, speculation, bust, and maturity.

So is what we’re calling Web3 the foundational investment period of a new subcycle, or the bubble period of the preceding one? It seems to me that one way to tell is the nature of the investment. Is abundant financial capital building out useful infrastructure in the way that we saw for the previous cycles?

It’s not clear to me that NFTs fit the bill. There’s no question, though, that the disruption of finance—in the same way that the internet has already disrupted media and commerce—would represent an essential next stage in the current cycle of technological revolution. In particular, if it were possible for capital to be allocated effectively without the trust and authority of large centralized capital providers (“Wall Street” so to speak), that would be a foundational advance. In that regard, what I’d be looking for is evidence of capital allocation via cryptocurrencies toward productive investment in the operating economy rather than capital allocation toward imaginary assets. Let me know of any good instances that you hear about.

To make clearer what I’m talking about, let me take an aside from crypto and Web3 to look at another technology revolution: the green energy revolution. There, it is completely obvious that bubble valuations are already financing the development of lasting infrastructure. Elon Musk has been a master at taking the outsized speculative price of Tesla stock (which at one point a year or two ago was valued at 1,500 years of the company’s profits!) and turning it into a nationwide electric vehicle charging grid, battery gigafactories, and autonomous vehicle capabilities, all the while catalyzing entire industries to chase him into the future. So too has Jeff Bezos used Amazon’s outsized valuation to build a new infrastructure of just-in-time commerce. And both of them are investing in the infrastructure of the commercial space industry.

In assessing the progress toward Web3 as advertised, I’d also compare the adoption of cryptocurrency for other functions of financial systems—purchasing, remittances, and so on—not only with traditional banking networks but also with other emerging technologies. For example, are Ripple and Stellar more successful platforms for cross-border remittances than bank transfers, credit cards, or PayPal, in the same way that Google Maps was better than Rand McNally or first-generation GPS pioneers like Garmin? There’s some evidence that crypto is becoming a meaningful player in this market, though regulatory hurdles are slowing adoption. Never mind remittances, though—what about payments more generally? How does growth compare with that of a non-crypto payment startup like Melio, which is focused on building against small business use cases? Given the interest in crypto from companies like Square (now Block) and Stripe, they are well positioned to tell us of progress for crypto relative to more traditional payment mechanisms.

Similarly, if Web3 is going to be the future of identity, or of social media, we need to ask ourselves what evidence is there of adoption—is it really a better mousetrap in the way that previous generations of internet technology proved to be? I fault the reporting on the field for its almost complete lack of coverage of this kind of information.

Where are we in the cycle?

The question might be asked whether the current stage of Web3 is more equivalent to 1995 or 1999—the early stage of the bubble or its end? Given the current valuation of crypto assets (and tech startups in general), it’s hard to argue for the earlier date.

I like to remind people that I wrote “What Is Web 2.0?” five years after the dot-com bust with the explicit goal of explaining why some companies survived and others did not. So too, I suspect that it won’t be till after the next bust that we’ll really understand what, if anything, Web3 consists of.

From the last bubble go round, I can offer several pragmatic observations in addition to the technology and business-model changes I had tried to capture in “What Is Web 2.0?”

  1. All of the companies that survived were making money—a lot of it. (In the case of Amazon, it was free cash flow, not profit, but the numbers were huge, as was the business and economic insight behind it.) Their valuations, while high, were supported by plausible models of future earnings and cash flow.
  2. None of them needed to raise enormous sums of money by today’s standards. (Yahoo’s total pre-IPO investment was $6.8 million, Google’s $36 million, and Amazon’s $108 million.) When you see companies go back again and again to investors for funding without ever reaching a profit, they may not really be businesses; they may better be thought of as financial instruments.
  3. They all had millions, then tens of millions, then hundreds of millions (and eventually billions) of daily active users for world-changing new services.
  4. They had all built unique, substantial, and lasting assets in the form of data, infrastructure, and differentiated business models.
  5. The companies that came to dominate the tech landscape during the next generation were not all up-and-comers. Apple and Microsoft handily made the transition to the next generation, and in the case of Apple, even led it.

Keep in mind that it was still early when the dot-com bubble popped. Google Maps hadn’t been invented yet, nor had the iPhone and Android. Online payments were in their infancy. No Twitter or Facebook. No AWS and cloud computing. Most of what we rely on today didn’t yet exist.

I suspect it will be the same for crypto. So much is yet to be created. Let’s focus on the parts of the Web3 vision that aren’t about easy riches, on solving hard problems in trust, identity, and decentralized finance. And above all, let’s focus on the interface between crypto and the real world that people live in, where, as  Matthew Yglesias put it when talking about housing inequality, “a society becomes wealthy over time by accumulating a stock of long-lasting capital goods.” If, as Sal Delle Palme argues, Web3 heralds the birth of a new economic system, let’s make it one that increases true wealth—not just paper wealth for those lucky enough to get in early but actual life-changing goods and services that make life better for everyone.

Post topics: Web
Post tags: Deep Dive