10 crypto predictions for 2023

10 predictions for what will be an incredible year for technology and probably a terrible year for finance.

Writing my crypto predictions for 2022 a year ago, I vowed to make them as “absolutely wild as possible.” If by “wild” what I meant was “actually not going to come true in 2022,” then wild indeed they were. On the other hand, if by “wild” I meant “anything up to but not including the wholesale immiseration of an entire industry at the hands of scam artists, which I have completely failed to predict,” well, then, 2022 was about to teach me what “wild” meant.

What I will say for my 2022 predictions is that I still believe in most of them—my timing was just really, really off. Even if our beloved industry hadn’t decided to self-immolate this year, most would have taken 5-10 years to play out.

So for this year’s predictions, I tried to be contrarian-pragmatic. What I mean by that is that many of the following 10 theses for 2023 are deeply unpopular within crypto—I’m not even sure I’ve fully convinced myself of one or two of them. But they also almost all entail projects willing to test these theses over the next year. In some cases, those are projects of friends that I work with and have invested in. In a number of cases, the project is simply my own.

To put that another way, I’ve tried to take something of an investor’s stance for each: the point is not to be right 100% of the time, or even half the time, but to be really right 20 or 30% of the time on the shit nobody else could believe. Ultimately that just means that I’m still a student of this space, and 2023 is likely to be a very harsh teacher.

Without further delay, 10 crypto predictions for 2023.

1. You integrate or you die

If smart contracts are internet-native laws, integrations are internet-native treaties. The lifeblood of an on-chain future, integrations represent blockchains’ unique value proposition—composability, ie the ability to permissionlessly interoperate with any on-chain service—and, I’d argue, blockchains’ moat. But this is a fairly counterintuitive idea these fearful days. Brave is the investor who dares to challenge the prevailing wisdom that composability is crypto’s biggest threat to crypto products’ defensibility: after all, we hear, if anyone can permissionlessly build on top of another service, they can also permissionlessly fork it, drive down the fees, and steal its users.

But listen to that argument more closely and you’ll intuit that in fact, the real threat is not integrating. Don’t want to collaborate with potential friends and foes? You become a nation without allies, stranded from a referral network of services that could keep each other alive—and enable on-chain executability across services. In precis, you’re fucked.

Let’s break that down. First, a protocol that captures users on the interface level is in a strong position to drive them to other services—which means that those other services are incentivized to build on top of the protocol rather than compete with it in order to get users. Second, the protocol and partner services can all share revenue from driving value to one another, and even do so permissionlessly.

And it’s why, as we approach the first era of cheap gas fees, we’ll finally see that composability is the moat.

2. A major fund collapses

Will it happen in 2023? Maybe not quite so soon. But funds that deployed at last year’s valuations will struggle to return a profit, and many are already facing contagion from portfolio companies that held funds on FTX. Constant down rounds and dwindling appetite among LPs for losing funds are likely to cause a domono-ing confidence crisis; a major fund collapsing in venture capital’s first ‘08 moment is not out of the question. 

Bears are simply the reapers coming to collect what the bulls have sown.

3. The long laughed-at social token thesis reemerges—in adult content

User-generated content in the 2010s meant that any creator could become a star; user-generated content in the 2020s means that any *fan* can become a star (see: TikTok fans remixing content). And here we see the whole problem with social tokens to date: they've been designed to give creators opportunities to monetize, rather than giving *fans* opportunities to monetize, and we’ve quickly learned that there can be no opportunity for the former without the latter. After all, there’s simply been nothing to do with social tokens except to join a Discord or collect an NFT.

Now, however, imagine social tokens as ways for fans to express themselves: paying to participate in contests to remix songs, telling writers what topics to write on, or voting on the content they’d like to see a creator make. There is no better vertical here than adult content, which always lives years ahead technologically and has particular incentives for fans to want to buy up tokens to get a say over the content produced—offering a whole host of new revenue opportunities, including paying-to-submit, paying-to-vote, and simply buying up supply to get extra fan power.

We can imagine this model solving the commonplace issue of DAOs depleting their treasuries while missing out on the upside of their own tokens that are sold in secondaries; here, creators could have fans pay to participate through their own token in order to constantly replenish their treasury and limit supply. Just as significantly, fans with winning ideas could win rewards—not only tokens, but the opportunity to cocreate content and start building their own brand too. (This is one of many, many use cases for what we’re building at jokedao.)

4. It becomes *easier* to do things on-chain than off-chain for the first time

Yeah, on-chain services offer users global access, composable executability, and long-term reputation building… but who cares? Until switching costs are net *negative*, retail users will not be bothered to use services that they believe subject themselves to immutable robbery. On-chain services have exactly one trick up their sleeve: wallets as replacements for usernames and passwords for one-tap login. But otherwise, signing and funding transactions with gas tokens that have to be acquired and sent cross-chain—this is the equivalent of asking earlier internet users to bring their computers to LAN parties to get access.

Sure, we delusional sci-fi geeks fancying ourselves as 24 1/2th century ethernauts—we’ll happily see challenging UI as an advanced puzzle, all the more pleasurable to solve. But for everyone else, 2023 will be the year that gasless relayers enable protocols and communities to subsidize gas fees for users, and for APIs—even slash commands in services like Twitter and Discord—to sign on their behalf. It won’t just be as easy for retail users to do things on-chain as off-chain; they may even be able to perform on-chain actions in familiar, web2 platforms. Major players here include Biconomy, Gelatto, OpenZeppelin Defender (and for signing, Lens’ Dispatcher). But special mention goes to Lit Protocol, which lets you run on-chain transactions through web2 services by minting a programmable key pair as an NFT. 

5. Account Abstraction delivers crypto’s missing piece: automation

“Account Abstraction,” as terrible a term as crypto’s custom demands, might better be described as Programmable Accounts—or simply, if you prefer, automating conditional logic. The idea of the thingamajig—I’ll call it Programmable Accounts—is simple enough. To date, a signature authenticated by a private key has always been required to execute code from a wallet. To put that simply, executing transactions has meant signing for them in the moment, so doing shit on-chain has been manual, cumbersome, and most important, unconditional: the only condition for executing code has been signing for it to happen.

But what if a transaction could execute if certain conditions were met? As long as MEV executors were incentivized to include these transactions in their block, we could effectively automate the execution of on-chain code. Fuel is enabling this in-protocol through predicates; Flashbots’ Suave could enable a simpler version through direct incentivization of block builders; and Anoma is building its own protocol along similar lines with what it calls “intents.” What we mean, simply, is that we can set preconditions for certain outcomes, sit back, and see those outcomes come to life on-chain through once the preconditions are met.

Why does this matter? For one, we solve the UX issue of making people sign on at set times to buy and sell shit by hand—now this can all be set in advance. For example, take a small limitation we run into at jokedao. We let anyone deploy and fund rewards pools that are distributed to the winners of contests: the precondition is that the person needs to win, and the outcome is that they get rewards. But someone still needs to actually execute the transaction to pay them out. Programmable Accounts should incentivize block builders just to include the transaction automatically as soon as the contest ends since the platform can set conditions for the transaction to execute—no signature required.

But more importantly, Programmable Accounts let users be more expressive about the kind of inputs and outputs they want to see on-chain. This becomes particularly useful for gaming—imagine that knife NFT being burned and minting a sword NFT if you defeat a giant on-chain within a certain time. Preconditions are set; outcomes are achieved when they’re met. Likewise, it solves the fundamental problem of recurring revenue in crypto: currently, you have to *manually opt-in* to renew subscriptions on services like ENS, whereas these intents and predicates could debit your wallet if it contains x tokens by y date, so that payment becomes automated and opt-out.

For the first time, we can have automated subscription services like we’ve long enjoyed in web2.

6. Unsexy wins.

One thesis for why so many top projects blew up this year: they only seemed like top projects because their risk profile meant 1000x returns on the way up and jail-time on the way down. The *real* top funds and projects, we could argue, flew entirely under the radar all along due to prudent risk management. And indeed, it’s Coinbase that’s trounced FTX, overcollateralized DeFi that’s trounced undercollateralized DeFi (much as we desperately need the latter long-term for this space to succeed), and the EVM that’s trounced, well, just about every other virtual machine. These are the boring, arguably subpar options, and much of what makes them seemingly boring and subpar is that they are relatively low-risk—but low-risk is of *massive* value in a hyper-volatile space that regularly scorches its top projects. Exciting might win bulls, and unsexy might win bears, but it’s ultimately what wins bears that matters most, because losing in a bull can still mean success while losing in a bear means total collapse. There is no greater value right now than being able to survive, and unsexy has the best chance to survive.

But then let’s note that this goes for unsexy theses that have seemingly been trounced as well: I’m placing a personal bet (the next years of my life) on governance tokens, social tokens, DAOs, and DAO tools. Their past failures have provided the necessary coursework for their future success.

7. Fee compression turns out to be overrated in the face of good UI

At a time where valuations are collapsing, capital is drying, and revenue is mattering more than ever, it’s astonishing to hear the latest idée reçue, that fee compression will drive platforms to 0% fees as they simply fork each other and eliminate commissions to draw users with airdrops. What? Do we think the victors of a bear market will be a bunch of crypto MoviePasses? Let’s start from basics: fees are sustainable and even *worthwhile* to users as long as they don’t surpass switching costs. Switching costs are relatively minor in crypto since it’s as easy to log into one service with a wallet as another, but they’re not 0 either. Oh also: despite what the bull market might have whispered in our ear, it turns out that revenue is how you sustain and build a business.

Two-sided marketplace effects, user inertia, and liquidity moats all justify ample lock-in, but for that matter, so does that rarest of crypto beasts: good UI/UX. This is the lesson of Uniswap’s ultra-convenient four-tap trading that few in crypto seem to have internalized. And frankly, it’s the value of forthcoming multi-chain services, which reduce dependencies on multiple platforms (Uniswap + Hop just to get $MATIC!) to single, one or two-tap shops.

Let’s note the formidable counterpoint: that bear markets wipe out retail and leave only the pros who are drawn to lower fees with little concern for UI. These pros then build the crucial liquidity for projects to draw retail in the bull market.

Will this actually happen? Well, yes, with pro-facing products (which is arguably why GMX’s better incentives let it trounce DyDx right now). But it remains unlikely with retail-facing products (which is arguably why Sushi’s better incentives are not letting it trounce Uniswap right now). The barometer of the battle—between fee compression on one side, and switching costs and liquidity on the other—remains, of course, OpenSea.

8. Arweave and AI (plus a dash of ZK) start solving the provenance problem

Forget everything else, and blockchains are revolutionary for a simple reason: in an era where we no longer trust anyone, whether they be experts, viral posts, deepfakes, or randos on the internet, blockchains are trust-minimized systems for validating what’s true. To date, this essential purpose of blockchains to *ascertain truth* has been limited to the easiest and most important facts to validate—financial transactions. But we are approaching an era of content blockchains, like Arweave, where validating provenance doesn’t just become a vital new task. It also becomes a fairly subjective one.

Let’s back up. The promise here is that a content creator could set a royalty on their work for anyone else to permissionlessly quote, remix, sell; if they make money doing so, the original creator will as well. In theory, this is a powerful new economic model to incentivize free creation by essentially decoupling manufacturing from distribution: a creator (the manufacturer of content) doesn’t even need to sell their items or represent them publicly or make them go viral, because a curator (the distributor of content) can do that instead. And each can share in proceeds.

But in practice, there is a problem: enforcing provenance. Plagiarizers can simply create derivatives to cut out creators, and distributors can use platforms that don’t enforce fees. Even if there were a way to identify and enforce royalties on content, remixers might distort content past original recognition, and tokens can be wrapped in different standards to evade royalties altogether.

Let’s speculate about one solution: zero-knowledge proofs (ZKPs). As Dan Boneh recently explained, ZKPs can be used to ascertain that modified media was edited directly from an original signed by the creator. If a DJ wants to remix a song, ZKPs should be able to assert the provenance of the original sample; likewise, ZKPs should give us proofs of the ways that deepfakes have edited their original material.

Imagine this being done on Arweave, a data storage protocol that lets anyone build applications on top of the content it stores. Specifically, imagine something like a crypto TikTok being built on Arweave. A creator uploads material and sets a commission, and then a fan remixes or duets with the material, monetizing their content through ads or NFTs, with both parties getting a commission. Anyone could remix or even distort any piece of content, with a ZKP attesting that the distortion is a modification of an original work.

But what if a remixer distorts content without proper attribution, tries to hide the fact they’re remixing someone else’s work, or even tries to claim it as their own by uploading it to Arweave first? Here, we can imagine a schema where AI would be able to probabalistically identify if content had appeared elsewhere or been modified, and potentially even run more deterministic ZKPs to assign appropriate attribution and royalties. Yes, Arweave can power this model.

But more importantly, we can even imagine a world where creators are incentivized to store their content on Arweave as the Chain-of-Truth because it becomes a canonical chain to establish provenance for original work. 

9. Regulation turns out to be good for decentralization

Yes anon, I know nobody in crypto likes regulation: regulators are centralized entities policing a trust-minimized system that should be able to regulate itself, and besides, securities regulation has an unfortunate tendency of criminalizing tokens’ profit-seeking abilities that would render them useful and valuable as financial tools. And yet, we’re seeing early signs that regulation may actually serve as a forcing function for decentralization.

Let us note here, anon, that many notable crypto projects are super fucking centralized. Take a DAO, any DAO, whose governance token is predominantly held by a few founders and whales. If they decide they want to use that governance token to appoint themselves profit or paid jobs, they risk violating securities laws for good reason: they’re using a community token to enrich themselves. But if they let token-holders nominate them to paid positions, nobody is using the token to earn money for themselves. Regulation forces decentralization and prevents conflict of interests. (This is my third and final shill for jokedao, which lets communities run these nomination systems—special thanks to this week’s sponsor Uncle Sam.)

Now take a chain that operates its own validator sets as if it were little more than a private database, risking hacks and rugs and general scoffing at a community of token-holders whose only real purpose is to pump its price. Here, too, regulation might force that chain onto Cosmos to properly decentralize with the full benefits of the modular stack. By outsourcing its validators to decentralized communities around the world, the chain is not only safer, but offers opportunity to anyone who wants to support it. 

Of course, long-term, regulation could turn out to be very bad for crypto, particularly as it becomes apparent that crypto is a decentralized system of financial regulation itself that threatens the slaphappy whims of chowderheaded bureaucrats.

In the meantime, though, it’s *starting* looking as though—I can’t believe I’m saying this—regulation could be a good thing for decentralization. Fingers crossed.

10. In lieu of any sustainable revenue anywhere in crypto, nearly every major new project just optimizes to capture MEV

MEV—basically, the ability of people who assemble blocks of transactions to profit by picking them, ordering them, and adding their own—is a perfectly well-paying business to extract value from simultaneously seeing and determining upcoming transactions on a chain. Imagine how much fortune-tellers would get paid if they had the power to make fortunes come true, and you’ll have some idea of the power of the MEVers.

For a time, MEVers were pooh-poohed as something like the trashmen of crypto, performing a necessary but rather unsavory job, this rather putrid business of taking bribes to include transactions and profiting off knowledge of your trades. But more recently, a few developments have made MEV more palatable to the crypto cognoscenti. Flashbots’ MEV-boost has helped democratize MEV for the masses to enjoy, while Vitalik has proposed in-protocol commitments to decentralize block building by hiding full upfront knowledge of transactions in a block.

But I also suspect that a bear market reckoning with the unfortunate fact that MEV is about the only real way to make money in crypto right now—well, this reckoning is forcing some ideological reconsideration that perhaps decentralization works by favoring highest bidders, that perhaps we should stop calling such bids “bribes” and start referring to them as “financial opportunities,” and that such bribes, I mean financial opportunities, are the best basis for new businesses.

To that end, expect wallets trying to capture MEV, app-chains trying to capture MEV], DEXes trying to capture MEV, transaction batchers trying to capture MEV, and your local-crypto-virgin-skipping-school-to-farm-crypto-much-to-mom’s-chagrin… trying to capture MEV. The starter pack for anyone looking to raise from crypto VCs in the next six months will be little more than a slogan that should magically open the wallets of any self-respecting investor: “by leveraging abc technology, we’ve built the first xyz protocol to capture MEV.”

While I’m hesitant that we’ll see fee compression for consumer services, it’s almost unavoidable that we’ll see fee compression for MEV as devs compete en masse to match each other’s arb bots, and block builders are forced to give a greater and greater share of profit to a wider and wider share of proposers. Ultimately, the major MEV opportunities will likely come from services that can build blocks for their own transactions, though this could still be years away. For example, imagine Uniswap deploying its own chain to be able to run arbitrage trades against its users in order to balance its prices, then ensuring these are executed. Corrupt? It might seem so now, but longterm, at reasonable prices, this could just become the model for how crypto companies make money.

Bonus theses:

1. Web2 wins Web3 because it understands two things Web3 itself does not: reputation and rebranding

2. We enter the age of permissionless privacy

3. Cross-chain services enable the birth of app-chains that can now easily integrate with traditional chains

4. The next great DAOs will start as jokes

5. We realize the solution to low governance participation is to lower governance participation even further

With special thanks to everyone who had conversations, feedback, and direct edits of the above, including Josh Benaron, Jim Chang, Chase Chapman, Jon Charbonneau, Joe Coll, Tina Haibodi, Yuan Han Li, Li Jin, Sean McCaffery, Peter Pan, Cami Ramos, Ryan Sproule, and Ryan Watkins.