yaocheng.eth retweetledi
I respect you a great deal, so when I read your praise of Kyle's talk, I immediately watched the video.
I tried to keep my feedback short but Kyle is an absolute fountain of misinformation and marketing spin, and I just kept writing, so here we go:
Kyle's sensationalist rewriting of Ethereum history is shameful and unfactual. As just one example, Kyle suggested that the rollup-centric roadmap was adopted in Oct 2020 two months after defi got started - he implies it as if the rollup roadmap was a knee-jerk response to defi-induced congestion - but the actual founding of defi might likely be traced to Dec 2017 when DAI first launched, three years before the community aligned around rollups.
Kyle says a lot of other garbage about Eth's history and roadmap, but by far the most nonsense thing he said in this segment is that "the Eth community is now questioning the rollup-centric roadmap". The fact that one guy named Max had a couple good ideas followed by terrible theses on the L1 economics and roadmap does not constitute the community questioning the L2 model. The L2 model is working. We're tripling down on it.
The second most nonsense thing Kyle said during this "eth history lesson" segment is that "ETH is not accruing value because it surrendered fees to L2s". What a convenient time for Kyle to take a temporary fee situation and project it as something surely to continue in perpetuity. In reality, Dencun's massive L2 capacity increase coincided with a bear selloff to produce specularly low L1 fees, a situation that is virtually impossible to persist. I have argued that L1 total execution and blob fees will remain very high on average over the medium to long term. L1 fees are already steadily recovering. Blobspace is on track to saturate in ~Q1.
Kyle's suggestion that L2 fragmentation is bad for defi and thus the L2 model is a bad strategy misses the broader point that fragmentation is unavoidable because humans want to control and specialize their chains and no one chain can hope to serve even a fraction of the coming global demand. Solana is now learning this with their "network extension" L2s.
He says "the idea that ETH is money is nonsense and non falsifiable. The litmus test is you can go to the coffee shop and prices are in ETH". For one thing, ETH is often the quote currency for many DEX pairs, NFT sales, gas fees, etc. For another, I'd argue that in this era when any token can possess many of the classically defining characteristics of money, money has instead primarily become about monetary premium aka confidence premium for people to prefer holding the money. Confidence can come from many sources. The main source of confidence in ETH is that Ethereum is the hub of the next-gen global economy.
Kyle talks about how L2s are "anti aligned". This concept of alignment has gotten way out of hand. The only entity in the entire world that isn't aligned with Ethereum is Lido and that's because they threaten to destabilize the base layer. The very concept of "eth alignment" was created by Lido proponents to negatively emotionally conjugate our publicity campaign against Lido dominance. L2s are aligned with Eth in the sense that they help build Ethereum's network effects as a global hub and a distribution mechanism for ETH as money, and, in the future, L2s will pay high fees in aggregate to the L1.
Kyle argues that Ethereum has an identity crisis. It's true that the Ethereum community itself has struggled to settle on a crisp one-sentence value prop for Ethereum or ETH. But a chunk of this identity "crisis" comes from Ethereum evolving over the years, including changing its mind when the facts changed, and from having a broad community base of diverse supporters with opinions that conflict in a healthy way. Solana is experiencing a similar identity crisis now - Anatoly often says they are focused on being the best trading venue (NASDAQ vision), and yet Sol has a diverse app layer that speaks to many stakeholders' needs, not just trading. Sol will soon feel this divergent stakeholder tension more strongly, such as with some of their teams starting to prefer L2s.
Kyle says "Eth can't tell us what they want to optimize for because they don't know" which is an outright fabrication. Eth optimizes for credible neutrality and being the global economic hub. Nobody in ethereum, except like Max, thinks credible neutrality is a bad overarching mandate.
Kyle tries to show that DA is worthless. But really all he can show is what we can all see, which is that today's L1 DA fees are de minimis and the L1 has a lot of would-be DA competitors. DA competitors are not identical products to L1 DA and the actual equilibrium pricing premium of blobs remains to be seen. I think blob fees will be high and argued that this week.
Much of Kyle's "SOL will flip ETH" thesis in this talk turns around his metrics. He suggests his metrics show that SOL is already starting to flip ETH and so it'll become a self-fulfilling prophecy. But will it?
On DEX volume, Kyle excludes L2s from his Eth comparison chart which fits nicely with his supposition that L2s are separate chains and don't accrue any value to ETH. There has also been reporting (eg by Wazz) on the high level of wash volume included in some Sol metrics.
Personally, I question the value of volume as the top metric because volumes are naturally higher on lower-fee chains with lower blocktimes (for sol, this is achieved via centralization) due to increased opportunity for bots and small arbs.
Kyle says "in our view onchain trading DEX volume is the most important metric for a blockchain". I prefer the metric of capital that people trust on your chain.
Solana has 12x less app capital than Ethereum and 3.2x less after normalizing by native token market cap. Sol also has 4x more memecoins per dollar of app capital and 25x less stablecoins per dollar of app capital. Data from Aug 31st.
Kyle shows a nice chart of MEV tips on Sol vs Eth. This chart is misleading for two reasons. One, Kyle probably picked just MEV tips instead of total fees because total fees makes Eth look better. Last 30d gas fees (excluding mev payments) on eth L1 are 6x sol... this is during a time of L1 fee collapse. The other and more important reason that this chart is misleading is because the Eth community is deep into a multi-year effort to protect users from being fleeced by unnecessary MEV, whereas Sol is only starting this journey.
In mid 2021, when Eth was terrible at protecting users from MEV, our "best" fee day ever was $118M in one day.
So, Sol's MEV numbers do not help show at all that Sol is going to flip Eth, it shows that Sol is fleecing the F out of retail memecoin traders.
Kyle's next chart shows that Solana stablecoin transfer volume is increasing on a relative basis vs Eth. However, keep in mind that Sol stablecoin transfer stats have been subject to huge wash volume, see Wazz's reporting on this. And, Kyle shows only the stablecoin transfer comparison vs the L1. He's happy to ignore all L2s, again based on his (bad) view that the L2 model is destructive to ETH. Kyle also picked transfer volume for his chart and not stablecoin market cap because, as I already mentioned, Sol has 25x less stablecoin market cap than Eth L1+L2s (drops slightly to ~22x if one insists on omitting L2s).
Kyle ends on three "structural advantages" of Solana vs Ethereum: token extensions, firedancer, and scaling with hardware.
Token extensions are effectively already being done (pioneered) on Ethereum via many custom tokens and L2s.
Firedancer is not close to delivering on its promised tps. First it must be fully implemented (vs frankendancer that is part rust client, part C codebase), and then firedancer must be trusted to run the entire network because the rust client won't be able to keep up. Plus, firedancer's "handcoded C modules" open up a ton of new security risk, as well as key person risk due to the relatively small number of individuals qualified to maintain such a low-level codebase.
Scaling with hardware is something Eth is doing, too. See ongoing gas limit increases on Base and Linea, MegaETH taking L2 hardware specialization to the limit, and parallel EVM implementations (not just monad).
Kyle's "unique" structural advantages of sol reminds me of the "Only Possible On Solana" campaign where pretty much every example they gave is actually possible on many L2s.
imo, Solana's true structural advantages are leveraging centralization and vertical integration to accelerate growth, and, for now at least, a higher level of curiosity and aggression in the app layer than is usually seen in Ethereum, such as with depin and consumer crypto.
In short, Kyle's stories about Solana's advantages and progress are over-embellished or total fiction. He omits any detail on Solana's structural disadvantages, such as that they sell the same 50k tps to every customer and that 50k tps is really more like 1.1k tps tops today.
SOL will never flip ETH. Solana is having its heyday this cycle. In the future, the fastest L2s will be faster than Solana. Solana's speed+cost advantages will narrow to zero. It will become obvious to everyone that the L2 model was a genius move and scales the way the world actually works, while providing excellent value accrual to ETH.
English
