Etherean to Crypto Moderate, Part 3: Doubt
2026 May 16
See all posts
Etherean to Crypto Moderate, Part 3: Doubt
As part of the 2020 DeFi wave there was also a big rise in NFT usage.
I never saw enormous value in NFTs. Most of the utility is similar to
people trading baseball/pokemon cards or gaming skins, which just is not
a big priority for me personally. Some people argued that ENS or other
kinds of naming systems are a real technical improvement over existing
naming systems, but I wasn't particularly convinced. While ENS is
obviously important for resolving crypto addresses, it's far from clear
that they improve something non-circular like DNS
names. I'd heard other proposed use cases like art provenance and
land registry records, but I always remained skeptical because the trust
model either depended on an oracle system or already necessarily had
some trusted 3rd party in the system, like a government.
I started reading people who were a lot more critical or pragmatic
about the space. Lyn Alden wrote an influential economic analysis of
Ethereum regarding the circular use cases and unclear value accrual
to the native asset. Tarun Chitra wrote a compelling critique around
PoS also with regard to its circularity: if higher yield opportunities
exist for ETH outside staking itself, there's an incentive for capital
to flee and degrade security. An anonymous writer called "Polyna" wrote
several articles articulating how narrow blockchain utility is and the
drawbacks
of such a model. I also started reading some of the more vocal critics
out there like Molly White, Stephen Diehl and Moxie Marlinspike.
Molly White called out, correctly, that the attempts at building
social networks on the blockchain are misguided because editing content
is not only essential for content moderation but it's a strongly
desirable feature for social media users. The immutable, public aspect
of blockchains was fundamentally incongruent with social media. Although
one could argue that the content on the web today is immutable due to
the widespread existence of archival services and the ease of
screenshots, blockchains do make archival the default and in the social
media context that should not be the case. Sure enough, over time
attempts like Deso and Bitclout would fade away. Even Farcaster, based
on the pragmatic philosophy of "sufficiently
decentralized," eventually gave up on scaling to the mass market and
returned capital to investors (protocol itself remained alive but
niche). You could argue that these attempts failed to catch on because
the existing social media networks had already reached an unassailable
level of dominance, but that wouldn't explain how some new social media
networks like TikTok and Bluesky had been able to gain traction. In the
case of Farcaster, for example, my guess is the benefits of RSS-style
"pick your own client" and "user owned social graph" are not worth
enough to justify the switching costs for most users, especially if
there are added UX hurdles with regard to the crypto aspect. The other
major point Molly White raised was the prevalence of hacks, theft and
general corruption (rug pulls), with her tongue-in-cheek website "web3 is going great." I
attributed the hacks primarily to buggy nascent technology, but it did
bother me that retail was still getting lured in with the promise of
riches and there was minimal legal protection. I assumed over time best
practices on making things safe and secure would emerge.
I never bought Stephen Diehl's vehemently strong claim that "the
technology does not solve a real problem," since I had seen clear
benefits myself in self-custody and stablecoin foreign exchange.
However, similar to Molly White, he pointed out the significant amount
of gambling and fraud taking place on blockchains. There was indeed a
general lack of consumer protection. Some tokens certainly do
intentionally flout securities law. Some people do use the tech to evade
regulations. I generally agreed with all of that criticism, but I
presumed regulation would catch up with the technology to resolve the
issue and find some middle ground. He also raised the often raised point
that "crypto currencies can't be used as currencies because of their
volatility." Indeed, the Bitcoin whitepaper talks about Bitcoin being
used as an actual medium of exchange. That clearly was too idealistic
and Stephen Diehl is correct there. A wholesale replacement of fiat
currencies would be way beyond the scope of what was reasonably
possible. In the subsequent years though, the utility of Bitcoin had
shifted towards digital gold. You'd hold it as insurance against
inflation and geopolitical uncertainty. It's exactly the kind of thing
that you want censorship resistance for. Ideally you would self-custody
the gold so that you need not trust an entity to hold it for you,
because that entity might become corrupt in such an environment. Digital
gold offers this while resolving the obvious impracticalities of storing
and exchanging physical gold.
I watched an older Moxie Marlinspike talk that stuck
with me on how even key values that are revered by crypto folks -
privacy, censorship resistance, availability, control - are in many
cases easier to achieve with a centralized service. Encryption !=
decentralization. Email at the protocol level will likely never be
encrypted because it's too hard to change, but centralized providers
like Gmail have been encrypted for years. Signal enabled wide spread
E2EE text messaging. Centralized services are way, way easier to upgrade
and that lets them improve things quickly even on the cypherpunk
dimensions. Once something gets truly decentralized, it becomes frozen
in time. IP shipped the first version in the 1990s and we've been trying
for 20 years to get to the next version. Later he also wrote a blog
post about web3 directly, touching on the same points and
highlighting how these forces of centralization have already seeped into
web3. That got me thinking much deeper about the cypherpunk ideals that
are often touted as part of the package with crypto: privacy,
decentralization, etc. I revisited my "bank wire with/without a
blockchain" example. The big gaping difference is that without a
blockchain model, the data in either respective operator database
remains private. That is exactly the point Moxie was making. Privacy is
actually easier in the centralized web2 model. Crypto does have privacy
solutions (e.g. Zcash, Tornado Cash, CoinJoin, private L2s like Aztec)
but they are far from easy. They might be the most technically
complicated objects in the whole space.
Getting in the weeds of DeFi engineering building an interoperability
protocol also raised a new nagging question that I couldn't get rid of.
The tension between upgradability and complexity. Useful stuff on-chain
requires complicated logic. The more complex, the more the application
starts to degrade into a traditional centralized, cloud based
application. Who can audit the code becomes more specialized. Users end
up deferring to the brand/reputation of the authors/auditors or
developers instead of examining the code themselves. This phenomenon
happens extremely quickly in practice relative to code complexity.
Complex code has not only a higher bug risk, but also a higher need for
maintenance as the feature set must evolve over time. That prevents
immutability for all but the most specific applications. The deposit
contract for eth2 is an example of something that needed to be
immutable, since it would hold potentially billions of dollars worth of
ETH. For this 150 line immutable contract, it took several years,
tens of highly specialized engineers/researchers and formal
verification before people could trust it. Uniswap pools are
similarly tiny by normal application standards and yet they still
couldn't avoid upgrades: there are 4 versions of the immutable
contracts, which obviously hurts UX. The desire for mutability is a
strong underlying force, ultimately because the world the software
serves is changing. Even the most techno-optimistic future with AI
generated, formally verified applications, is still beholden to the
capricious whims of the human end users. Mutable contract code then
raises the question of who can upgrade the code and how. The
industry best practice in that context is "governance" which usually
amounts to either a council of humans (very much like a corporate board)
or token holders (very much like equity holders). I couldn't shake the
feeling that introducing governance to large systems of on-chain smart
contracts so that we can evolve them narrowed the original vision
significantly.
Over time I also started to recognize the naivety of many crypto
projects with regard to some of the purported problems they are trying
to solve. As a regular concert goer, I find ticket scalping a
particularly interesting problem. There has been no shortage of purported blockchain
solutions to this problem. This talk from
the well-regarded Flashbots group, for example, describes how crypto can
potentially help with something like the Taylor
Swift - Ticketmaster controversy. The same group even has some code with even more
explicit claims that it can. Can crypto actually help with the problem?
What is the root of the problem in the first place? Often it's the case
that artists want to set the price of their shows specifically below
what the market would bear to ensure that they remain affordable to fans
with less means. That creates a profit motive for people to buy and
resell the tickets. So profit seeking actors deploy bots to purchase as
many tickets as possible as soon as they are available. One clear
solution is for the ticketing service to disallow transfers as that
would remove the profit incentive. However, disallowing transfers
outright is usually undesirable for obvious UX reasons and is actually
illegal in 6 US states where legislation was passed deeming concert
tickets personal property and thus requiring them to be transferrable.
If you allow transfers for free, then money exchange can simply happen
off platform or 3rd party platforms and scalping can continue. What
about if the ticketing service only allows face value resale? That has
actually existed on Ticketmaster since 2019 and effectively solves the
problem, except for the states where it's been made illegal or in cases
where the artist chooses not to use it (Taylor Swift chose not to,
unclear why). For the states free transfer is legally imposed or where
the artist chooses not to enable face value resale restrictions, then
there's nothing that any technological solution can do to fully solve
the problem. The best solution in that case is only partial: strong
identity checks (government ID) at time of purchase with per identity
purchase limits. Scalpers can still profit by paying real people to
purchase tickets, but it's significantly harder to scale. Crypto-based
proposals to the problem of concert ticket scalping were a clear example
to me of proposals that were doomed to fail by simply not bothering to
understand the real world problem they are trying to solve. At best it's
pure naivety, but the uncharitable view is that it's a form of post-hoc
rationalization. Playing around with auction design, game theory and
cryptography is fun but often not tied to an immediate real world
problem. There are plenty of people who enjoy working on challenging
technical problems and also have a strong need to feel like they are
helping the world sometimes (myself included). The misstep though is to
seek out a real world problem after doing the fun experimenting
and attempt to use that as the original motivation.
Next in the series: Etherean to Crypto
Moderate, Part 4: Equilibrium
Questions or comments? Email feedback@cstein.xyz
Etherean to Crypto Moderate, Part 3: Doubt
2026 May 16 See all postsAs part of the 2020 DeFi wave there was also a big rise in NFT usage. I never saw enormous value in NFTs. Most of the utility is similar to people trading baseball/pokemon cards or gaming skins, which just is not a big priority for me personally. Some people argued that ENS or other kinds of naming systems are a real technical improvement over existing naming systems, but I wasn't particularly convinced. While ENS is obviously important for resolving crypto addresses, it's far from clear that they improve something non-circular like DNS names. I'd heard other proposed use cases like art provenance and land registry records, but I always remained skeptical because the trust model either depended on an oracle system or already necessarily had some trusted 3rd party in the system, like a government.
I started reading people who were a lot more critical or pragmatic about the space. Lyn Alden wrote an influential economic analysis of Ethereum regarding the circular use cases and unclear value accrual to the native asset. Tarun Chitra wrote a compelling critique around PoS also with regard to its circularity: if higher yield opportunities exist for ETH outside staking itself, there's an incentive for capital to flee and degrade security. An anonymous writer called "Polyna" wrote several articles articulating how narrow blockchain utility is and the drawbacks of such a model. I also started reading some of the more vocal critics out there like Molly White, Stephen Diehl and Moxie Marlinspike.
Molly White called out, correctly, that the attempts at building social networks on the blockchain are misguided because editing content is not only essential for content moderation but it's a strongly desirable feature for social media users. The immutable, public aspect of blockchains was fundamentally incongruent with social media. Although one could argue that the content on the web today is immutable due to the widespread existence of archival services and the ease of screenshots, blockchains do make archival the default and in the social media context that should not be the case. Sure enough, over time attempts like Deso and Bitclout would fade away. Even Farcaster, based on the pragmatic philosophy of "sufficiently decentralized," eventually gave up on scaling to the mass market and returned capital to investors (protocol itself remained alive but niche). You could argue that these attempts failed to catch on because the existing social media networks had already reached an unassailable level of dominance, but that wouldn't explain how some new social media networks like TikTok and Bluesky had been able to gain traction. In the case of Farcaster, for example, my guess is the benefits of RSS-style "pick your own client" and "user owned social graph" are not worth enough to justify the switching costs for most users, especially if there are added UX hurdles with regard to the crypto aspect. The other major point Molly White raised was the prevalence of hacks, theft and general corruption (rug pulls), with her tongue-in-cheek website "web3 is going great." I attributed the hacks primarily to buggy nascent technology, but it did bother me that retail was still getting lured in with the promise of riches and there was minimal legal protection. I assumed over time best practices on making things safe and secure would emerge.
I never bought Stephen Diehl's vehemently strong claim that "the technology does not solve a real problem," since I had seen clear benefits myself in self-custody and stablecoin foreign exchange. However, similar to Molly White, he pointed out the significant amount of gambling and fraud taking place on blockchains. There was indeed a general lack of consumer protection. Some tokens certainly do intentionally flout securities law. Some people do use the tech to evade regulations. I generally agreed with all of that criticism, but I presumed regulation would catch up with the technology to resolve the issue and find some middle ground. He also raised the often raised point that "crypto currencies can't be used as currencies because of their volatility." Indeed, the Bitcoin whitepaper talks about Bitcoin being used as an actual medium of exchange. That clearly was too idealistic and Stephen Diehl is correct there. A wholesale replacement of fiat currencies would be way beyond the scope of what was reasonably possible. In the subsequent years though, the utility of Bitcoin had shifted towards digital gold. You'd hold it as insurance against inflation and geopolitical uncertainty. It's exactly the kind of thing that you want censorship resistance for. Ideally you would self-custody the gold so that you need not trust an entity to hold it for you, because that entity might become corrupt in such an environment. Digital gold offers this while resolving the obvious impracticalities of storing and exchanging physical gold.
I watched an older Moxie Marlinspike talk that stuck with me on how even key values that are revered by crypto folks - privacy, censorship resistance, availability, control - are in many cases easier to achieve with a centralized service. Encryption != decentralization. Email at the protocol level will likely never be encrypted because it's too hard to change, but centralized providers like Gmail have been encrypted for years. Signal enabled wide spread E2EE text messaging. Centralized services are way, way easier to upgrade and that lets them improve things quickly even on the cypherpunk dimensions. Once something gets truly decentralized, it becomes frozen in time. IP shipped the first version in the 1990s and we've been trying for 20 years to get to the next version. Later he also wrote a blog post about web3 directly, touching on the same points and highlighting how these forces of centralization have already seeped into web3. That got me thinking much deeper about the cypherpunk ideals that are often touted as part of the package with crypto: privacy, decentralization, etc. I revisited my "bank wire with/without a blockchain" example. The big gaping difference is that without a blockchain model, the data in either respective operator database remains private. That is exactly the point Moxie was making. Privacy is actually easier in the centralized web2 model. Crypto does have privacy solutions (e.g. Zcash, Tornado Cash, CoinJoin, private L2s like Aztec) but they are far from easy. They might be the most technically complicated objects in the whole space.
Getting in the weeds of DeFi engineering building an interoperability protocol also raised a new nagging question that I couldn't get rid of. The tension between upgradability and complexity. Useful stuff on-chain requires complicated logic. The more complex, the more the application starts to degrade into a traditional centralized, cloud based application. Who can audit the code becomes more specialized. Users end up deferring to the brand/reputation of the authors/auditors or developers instead of examining the code themselves. This phenomenon happens extremely quickly in practice relative to code complexity. Complex code has not only a higher bug risk, but also a higher need for maintenance as the feature set must evolve over time. That prevents immutability for all but the most specific applications. The deposit contract for eth2 is an example of something that needed to be immutable, since it would hold potentially billions of dollars worth of ETH. For this 150 line immutable contract, it took several years, tens of highly specialized engineers/researchers and formal verification before people could trust it. Uniswap pools are similarly tiny by normal application standards and yet they still couldn't avoid upgrades: there are 4 versions of the immutable contracts, which obviously hurts UX. The desire for mutability is a strong underlying force, ultimately because the world the software serves is changing. Even the most techno-optimistic future with AI generated, formally verified applications, is still beholden to the capricious whims of the human end users. Mutable contract code then raises the question of who can upgrade the code and how. The industry best practice in that context is "governance" which usually amounts to either a council of humans (very much like a corporate board) or token holders (very much like equity holders). I couldn't shake the feeling that introducing governance to large systems of on-chain smart contracts so that we can evolve them narrowed the original vision significantly.
Over time I also started to recognize the naivety of many crypto projects with regard to some of the purported problems they are trying to solve. As a regular concert goer, I find ticket scalping a particularly interesting problem. There has been no shortage of purported blockchain solutions to this problem. This talk from the well-regarded Flashbots group, for example, describes how crypto can potentially help with something like the Taylor Swift - Ticketmaster controversy. The same group even has some code with even more explicit claims that it can. Can crypto actually help with the problem? What is the root of the problem in the first place? Often it's the case that artists want to set the price of their shows specifically below what the market would bear to ensure that they remain affordable to fans with less means. That creates a profit motive for people to buy and resell the tickets. So profit seeking actors deploy bots to purchase as many tickets as possible as soon as they are available. One clear solution is for the ticketing service to disallow transfers as that would remove the profit incentive. However, disallowing transfers outright is usually undesirable for obvious UX reasons and is actually illegal in 6 US states where legislation was passed deeming concert tickets personal property and thus requiring them to be transferrable. If you allow transfers for free, then money exchange can simply happen off platform or 3rd party platforms and scalping can continue. What about if the ticketing service only allows face value resale? That has actually existed on Ticketmaster since 2019 and effectively solves the problem, except for the states where it's been made illegal or in cases where the artist chooses not to use it (Taylor Swift chose not to, unclear why). For the states free transfer is legally imposed or where the artist chooses not to enable face value resale restrictions, then there's nothing that any technological solution can do to fully solve the problem. The best solution in that case is only partial: strong identity checks (government ID) at time of purchase with per identity purchase limits. Scalpers can still profit by paying real people to purchase tickets, but it's significantly harder to scale. Crypto-based proposals to the problem of concert ticket scalping were a clear example to me of proposals that were doomed to fail by simply not bothering to understand the real world problem they are trying to solve. At best it's pure naivety, but the uncharitable view is that it's a form of post-hoc rationalization. Playing around with auction design, game theory and cryptography is fun but often not tied to an immediate real world problem. There are plenty of people who enjoy working on challenging technical problems and also have a strong need to feel like they are helping the world sometimes (myself included). The misstep though is to seek out a real world problem after doing the fun experimenting and attempt to use that as the original motivation.
Next in the series: Etherean to Crypto Moderate, Part 4: Equilibrium
Questions or comments? Email feedback@cstein.xyz