pysel☕️

Check out my talk on Orderbooks and Automated Market Makers (AMMs) that wrapped the quarter for CS 190N Foundations for Blockchains and Cryptocurrencies at @UCSB. youtube.com/watch?v=1Ewh5Q… TL;DR below 👇

Got a decent amount of pushback for this one. Ethereum is an incredible social phenomenon and technology, and I have mad respect for everything the community has built. However, today's thesis of Ethereum is questionable, in my opinion. That was the whole point of the post. I am really surprised by how close to heart some eth maxis take any criticism towards Ethereum. This makes the community look like a big echo chamber where you either agree with everyone else in it or are simply wrong. Thanks to those who engaged in rational discussions -- that is a great way to respond to something like this. To those who took it so personally to engage in insults, I recommend this book.

@sui414 @buffalu__ They also quote pretty wide. Wide spread, along with a huge volume of non-toxic flows makes it very lucrative for MMs to provide liquidity in those markets.

@buffalu__ there are sources u can buy data feed with p good latency advantage (much better than audience’s phone-tower-server latency) but anecdotally i heard it starts at $100k access some times

how do prediction market MMs make markets on sports? presumably the live feed has too much delay. intern in the audience?

@jchaskin22 How so? I was under the impression that 7702 makes EOA's outgoing logic configurable, not incoming. If some shitcoin calls transfer() with my address as a recipient, how would 7702 help me?

@pysellll Any wallet can build this today using EIP-7702

Dear Ethereum, I would happily sacrifice some decentralization to see this rate of shipping and solving real problems that users have. Best, Pysel

@TokenBrice I wish you all the best in getting there then.

@pysellll Good, Ethereum is not for everyone. If you are satisfied with the current state of the political and financial systems, you will likely find your happiness somewhere else. Ethereum is for the punks. Overcompliants need not apply.

@TokenBrice And I am happy for Ethereum, but I don’t feel aligned with it

@pysellll Dear Pysel, You are free to sacrifice decentralization as much as you want on your corpochain for enhanced compliance, however Ethereum has other ambitions. Best, Brice

@wiggyhop I understand the thesis, I just don’t feel aligned with it anymore

@pysellll You don't understand ethereum then

What if I wish to focus on different parts of the stack? I had the opportunity to do exactly what you said at EF, but I decided not to, because I wish to do other things. If somebody criticizes Rust for bad async code, would you tell them to start contributing to Rust to solve it? I think people can have opinions.

@pysellll i mean if you want to see faster changes, you can participate in forum discussions and champion eip that you want to be included

let’s make endless weekly changes to a global decentralized protocol securing hundreds of billions of assets like a corpo chain with 4 validators securing over $15m worth of assets make it make sense

I am not sure I am still following our conversation. We started with consensus mechanisms vs sybil resistance mechanisms. PBS (or ePBS) doesn’t have to do with neither of those, it is just a mechanism to “democratize” MEV revenue (which hasn’t played out well, but that is a separate line of discussion). Consensus mechanisms are BFT-style + longest chain. Sybil-resistance mechanisms are PoW or PoS. I attached a source from one of the most reputable researchers in the blockchain space. I worry we just have different definitions.

@pysellll @Tim_Roughgarden also, unless you're talking about another chain, ePBS isn't live (or close to deployment?)

I’ll be the annoying guy and say that PoW and PoS are not consensus mechanisms. They are sybil-resistance mechanisms. Here is an awesome lecture on this by @Tim_Roughgarden timroughgarden.github.io/fob21/l/l9.pdf

@c477bfef6df4311 @Tim_Roughgarden What do you mean by expected next block? Like, you have locally txs T1, T2, but you get a block with txs T3, T4? If so, that is not what I was talking about, I was referring to a literally invalid block from formatting perspective.

@pysellll @Tim_Roughgarden 1. wrong for now, ePBS is not online; you can have an expected next block and still get a different proposed—yet valid; 2. because your node isn't voting on anything unless it is mining, which is the whole point for ordering tied to reward (+ cost)—cont.

@jarxiao My guess is both. Even if you assume a world of really capable models, good products must come with some form of differentiation. That differentiation is very likely good taste and a high skill of launching a product.

The lack of mainstream breakout vibecoded apps reveals a truth about building: the first 95% has always been easy. Building apps pre-AI was never bottlenecked by cognition, it was just tedious. The next few years will reveal whether that last 5% is a skill gap or taste gap.

@c477bfef6df4311 @Tim_Roughgarden You are talking about censorship-resistance, not sybil-resistance. Also, you don’t need many nodes to stop propagation, you need a lot of hashrate (which again, is different from node count, precisely due to sybil resistance)

@pysellll @Tim_Roughgarden + mempool propagation sources i.e. if your node is not broadcasting OFAC flagged txs, you will need a gigantic blob of nodes to stop propagation, orders of magnitude more

1. The block template is not proposed by a miner or staker, it is pre-determined by a protocol. If I propose a block that is not properly formatted, even if I solved a hash, it will be rejected. 2. This is not the definition of sybil resistance though. You are correct that if I have, for example, one very slow peer, and I only get state updates from it, I might lag behind the rest of the network. How is that related to sybil resistance though?

@pysellll @Tim_Roughgarden 1. your node will reject any ordering that doesn't comply with your rules; the block template is (currently) proposed by the miner/staker 2. your node seeks peers to find current state/block height/ mempool/ complete IBD—the sybil resistance comes from diverse IBD sources +

By definition of sybil resistance (left it in a diff response above), it also doesn’t “come” from node count, it is just a mechanism that prevents influencing a network by spinning up an unbounded number of sybils. Also, P2P has nothing to do with sybil resistance. I think you might have a different definition in mind, and we are just talking with different definitions in mind.

@pysellll @Tim_Roughgarden the Sybil resistance of btc and eth do not come from node count, but from how many peers you can reach through P2P networking to get a correct state/current mempool

1. Are you talking about a proposer or generally about any node? I assume the former, as you are talking about a miner. A proposer does not choose the order of the block in the chain, they just propose a block according to their internal state of the network. If they propose a block, it might be rejected by the network even if it is correct according to their internal state. This is precisely my point about a distinction between sybil resistance and consensus mechanism. 2. Sybil resistance mechanisms make sure that if I have an X amount of resources that determine my power in the network (hashrate for PoW, stake for PoS), it does not matter if I spin 10 nodes and allocate X resources between them, or if I spin a single node and allocate the entire X to it. This was my point about 10 nodes.

@pysellll @Tim_Roughgarden 1. the miner/staker orders the block along its chain of blocks... that's the whole function of mining 2. if you spin 10 nodes it does nothing until you have altered them—you can even game which transactions you broadcast or don't, which have been a current debate on bitcoin

Neither PoW nor PoS are used to order anything. They are mechanisms to select a leader that proposes a block (and fundamentally prevent situations where a single entity can spawn an unbounded number of sybils giving them a higher chance at becoming said leader, or increase control over the network). In the Bitcoin example, PoW is a mechanism to choose which node is eligible to propose a block. If I got a good hash, I can propose, and the network will accept my block. However, the fact that I propose a block does not give any guarantees about this block’s finality. The longest chain mechanism is what guarantees it. Imho, the reason why everyone calls PoW a consensus mechanism is because it is indeed very coupled with the consensus mechanism of Bitcoin (since Bitcoin is, speaking in protocol terms, very simple). However, the distinction becomes more evident if you look at PoS chains. There, the distinction between sybil-resistance mechanisms and consensus mechanisms is more evident. On your networking point, I will be honest, I don’t know much about 0.8 upgrade. However, you cannot guarantee sybil-resistance on the networking layer, because the networking layer itself is not sybil-resistant. If I spin up 10 nodes (be it, with a version higher than your mentioned 0.8) and connect them to the system, it doesn’t automatically mean that I have more “power” over the network.

@pysellll @Tim_Roughgarden PoW and PoS are not at all part of Sybil resistance; that is done through P2P networking, which was one of the few hardforks bitcoin has performed (0.8 is the furthest a modern node will peer to); pow/pos are merely for ordering