Thomas Jay Rush

See you in Cannes 🫡

@SCBuergel @VitalikButerin That’s very funny. It’s you and me, brother.

@tjayrush @VitalikButerin oh no, the other node runner also showed up :p

We should be open to revisiting whole beacon/execution client separation thing. Running two daemons and getting them to talk to each other is far more difficult than running one daemon. Our goal is to make the self-sovereign way of using ethereum have good UX. In many cases that means running your own node. The current approach to running your own node adds needless complexity. Short-term, maybe we want some more standardized basic wrapper that lets you install dockers of any client and make them talk to each other easily? Also good that @ethnimbus unified node github.com/status-im/nimb… exists. Longer term, we should be open to revisiting the whole architecture once @leanethereum lean consensus is more mature.

@SCBuergel @VitalikButerin Here here.

I think this is entirely missing what Ethereum developers *honestly* think. I guarantee you, nobody building a serious app is thinking "oh, if only I would have to run one container less, I'd run my own Ethereum node". The entire discussion of running your own Ethereum node isn't even happening! Every developer knows that "running an Ethereum node is so broken and painful that OF COURSE we're not doing that, lmao!". I've actually had these exact discussions because I did attempt to make people who build apps (including those at @gnosis_ ) run their own nodes. It's not happening, there's not even a discussion about it. Running an Ethereum node is absolutely not in the Overton window for [d]app devs. So are Ethereum [d]app devs just stupid or unwilling to listen to Vitalik? No, I think Vitalik is just unaware of how bad the reality really is. I challenge you: take your agents on a one day journey to build any contemporary Ethereum app. I don't mean some sexy smart contracts, I mean the mundane web frontend that of course needs an indexer, make it AA because you like it sexy and then of course you need a bundler and then of course you need a price feed provider and then of course you need cross chain RPC providers to render a total balance. Build that on top of one single Ethereum client that you run yourself. Then (because you think that we have current diversity, right...?) swap that one client for another that you also run yourself. I know you're talking about individuals and not devs, but that's another level up from [d]app devs, especially because of all the additional services which a contemporary Ethereum app needs and that has to FOSS alternatives that anyone could run even if they wanted to. My point is: Ethereum nodes don't cater to the needs of actual [d]app developers, they cater to the imaginary battle against Solana and Base. That is my biggest ivory tower criticism of Ethereum.

Happy to see Vitalik backing up this idea, but also, I would like to remind people that Erigon had this for the past 2 years? 🤷‍♀️ I think this was one of my most controversial ideas at the time and even got high-ish degree of backlash at the time for even daring to try this out. I usually don't do this but I think some credit should be due here.

@VitalikButerin Been making this argument since 2016. It’s not going to be enough to just get people to run nodes. RPC also needs significant improvements. Particularly indexing. See everything by TrueBlocks since 2016.

I feel like at every level we've implicitly made this decision that running a node is this oh so scary devops task that it is ok to leave to professionals. IT IS NOT. We need to reverse this. Running your own Ethereum infrastructure should be the basic right of every individual and household. "The hardware requirement is high, therefore it's okay for the devops skill and time requirements to also be high" is not an excuse. Even people who can afford high-end hardware, dedicated staking boxes, etc often do not have a lot of free time. Nodes should be easy.

@VitalikButerin we do this with client.buidlguidl.com ser

I'm not trying to be argumentative, but that's of no interest to an end user. Easier for dev. Costly for user. Non-browser apps can read binary directly from disk. 2 (or more) orders of magnitude faster than JSON stringification. That means 100 times less heavy or 100 times more capable apps.

@tjayrush @tkstanczak @androolloyd With a browser worker you can debug the front end of any app. It’s actually really nice.

soon when you buy a new phone it will be some empty OpenClaw-like OS that will just wait for you to tell it what applications it should create for you hello phone, I need an alarm clock, a calendar, and a browser

@creepercharge56 @tkstanczak @androolloyd Local software never has to rate limit so the entire end user machine is devoted to one user. We’ve found it to be extremely fast in comparison to accessing data from a shared resource like an API. Plus private (no login). Plus free.

@tjayrush @tkstanczak @androolloyd Wouldn't that be extremely slow

🪚 Fixed Fee Swap: How Splitting Yield From Principal can Reduce Risk for Uniswap v4 LPs Thanks to grant funding from @UniswapFND, we built & open sourced proof-of-concept LP yield tokenization on Uniswap v4, based on academic work by Zachary Feinstein & Fayçal Drissi 👇

In or near Philadelphia? Love OpenClaw? Join us this Thursday for the first OpenClaw meetup in Philly! luma.com/96e1770h

@aditiitwt I pre-date angle brackets in include statements. The only thing I learned you could use would be quotes.

You may be old But are you this old

@marilyn100x @DeFi_Dad @ethereum Search for an article I wrote about two years ago called “Revenge of the Nerds” at tjayrush.medium.com. I tried really hard to explain my answer to exactly that question.

I understand your POV. What according to you should have been the path? I'm trying to understand how would someone actually be interested in running a node just to use the wallet/network? How feasible is it? I mean not everyone could do that or wait till the node syncs and then use the wallet and everytime a tx needs to be executed, you have to sync the wallet to the current block and then transact?

Setting up an @ethereum wallet in 2015 took three days. Three full days, before you could do anything. > 10-20GB of storage. > A computer powerful enough to handle it. > An internet connection that couldn't drop. > Days where you didn't need your laptop for anything else. The official wallet was called Mist Browser, built by the @ethereumfndn itself. Here's what using it looked like: + Downloading blocks. 12%, 3 days remaining. + Close your laptop -> sync pauses. + Internet drops -> sync corrupts. + Sync corrupts -> start over, back to zero and you needed three more days. Smart contracts were live, developers were building. The protocols existed but none of it was reachable. You couldn't just try Ethereum. You had to earn the right to try it first. The technology was ready. The world wasn't allowed in yet. Then in 2016, Aaron Davis and Dan Finlay asked one question. What if Ethereum lived in your browser instead of your hard drive? A Chrome extension, two minutes to install. Twelve words to write down? That's it. The exact same Ethereum, without any barrier? Thats how @MetaMask came into existence. It made life so easier that, a developer could build something and send a link. Anyone in the world could click it, connect in seconds, and use it. The full node requirement didn't get easier. It got deleted. @MetaMask did a lot of good for the space and hardly gets the recognition it deserves.

Everyone accepts that AI agents will run 24/7 on your devices. But suggest running a Lightning node and suddenly it's "unrealistic." It's the same thesis.

1/ We put together a list of projects doing GPU-accelerated crypto & ZK proving on client-side devices. i.e. browsers, phones, laptops No server. No trusted backend. Just your hardware github.com/zkmopro/awesom…

Saying the quite part out loud. All these SaaS crypto accounting firms are terrifying data honeypots. Opt out, be free and use @rotkiapp. A local tracking & accounting tool that keeps your data with you, not at someone else's server.

It turns out some AI models can step through EVM bytecode in their head. We built EthIQ, a new benchmark to test how well models actually understand Ethereum protocol internals. 325 questions across two evaluation modes, and one alive canary 🐦 Read more 👇

@URozmej @LefterisJP @base @Nethermind I’ve been saying this since 2016. What’s the disc space requirement to run a Nethermind Archive node? Is it erigon sized or Geth sized? Someone should point an AI at true blocks and nethermind. I’ve run out of money and time.

@LefterisJP @base With EVM chains scaling indexers should be built directly in the nodes itself. Skip JSON RPC, go directly to memory. Get live events of new blocks, fork choice rules or reorgs. Have direct access to state modified by new blocks. That is why @Nethermind has a robust plugin system.

I hate @base with a passion. It's the chain that does not care about its ecosystem, its community and the tools used to interact and process it. That means that tools like rotki end up getting bad data, or data ends up being unavailable. Fix your shit.