Sticking to 8192 signatures per slot post-SSF: how and why | Hacker News Reader

Sticking to 8192 signatures per slot post-SSF: how and why

ethresear.ch

44 pointsanaoum27 comments

Really interesting to see where the scaling and complexity limits are for blockchains. In this case, it's signature aggregation for the huge numbers of machines (~hundreds of thousands now in Ethereum) that attest to the validity of state.

There is a lot of interesting programming + mechanism design + cryptography happening in crypto.

I don't envy the client teams that have to keep up with all these changes. Ethereum really is the most impressive open source project in the world, IMO, because it's a live system with half a dozen different clients written in different languages that all have to interoperate with each other and perform complex cryptography in short timeframes. Not a lot of space for bugs and any breaking change has to have an elaborate migration coordinated among at least 500K different validators globally.

> Ethereum really is the most impressive open source project in the world, IMO, because it's a live system with half a dozen different clients written in different languages that all have to interoperate with each other and perform complex cryptography in short timeframes.

I am curious, are you aware of linux?

Seriously... Too many people just buy the excuses of these systems even though there are massively distributed systems that work on existing technology quite well. Crypto is just an excuse to scam others with the promise of $$ rewards using words they don't understand and seeming to be high tech.

Linux is of course very impressive and most Ethereum nodes run on Linux. But you can peg a machine to a specific version of Linux and never have to upgrade if you desire. Your machine is ultimately independent of other Linux machines. The key challenge with a blockchain is that you have to keep what amounts to a globally distributed and permissionless data center in sync with protocol upgrades and the cost of making a critical mistake can be in the billions of dollars. You have to account for not just technical vulnerabilities but game theory vulnerabilities.

The challenges are of course, entirely different in OS kernel land. All of them are impressive.

I'm fully aware of Linux... It's also a very impressive open-source project. To elaborate, reason I consider Ethereum to be more impressive is the collaboration between maintainers and releases of different clients and client versions in different languages. This happens with different packages in Linux to some degree but on Ethereum it's a different ballgame because it's a live system, it's not like you can just let some users upgrade to beta versions and fix bugs as you go, everything has to work together perfectly with a system-wide change happening in under 12 seconds globally with billions of dollars on the line. There are testnets but no true analogue for mainnet. (Edit: chrisco255 below said it better)

My question is when and how we can reduce the gas fee. It’s the big problem now.

Is it? Layer 2 solutions are here and working well.

OP and Polygon L2s. Fees are like fractions of a penny or max 1-2 cents.

and/or sharding

What people normally refer to as "Polygon" is their Matic sidechain, which is not a "layer 2" system. It's its own consensus system with its own stake who can steal funds transferred to it from the main Ethereum network. To be considered a L2 system it has to not require additional economic majority-trust assumptions like that.

See also extended explanation: https://ethereum.org/en/layer-2/ (which doesn't mention systems like payment channels since those are limited to just payments)

Polygon the company has developed a rollup based on zero knowledge proofs but afaik that hasn't been deployed with real money yet.

Polygon zkEVM is still in alpha, but it is live and it has more than $100m in assets live on chain. Other rollups with close to $10B in assets, like Arbitrum, do exist though:

https://l2beat.com/scaling/summary?#active

I feel like crypto in general was designed by speculation, and not real use-case.

If you read the original bitcoin whitepaper we have veered into a completely different direction.

I was fortunate enough to build and operate a payment system, during covid 2019 - 2022. It was able to do everything required to process payment maintain an accurate ledger. Our little team of 6 built it in 9 months.

It was centralized yes, and that made things a lot simpler than crypto so comparing it to crypto is a bit unfair admittedly. But you know the difference is it was designed based on real use-case. It could process billions of $ in transactions. Could scale up / down etc...

More importantly we had real people using the thing. People who had no idea how payment systems worked. People were sending money to their mothers for her birthday. Sending money to pay for school fees, hospital bills, etc...

What the crypto community including Vitalik Buterin has failed to realize is the nodes on the network (nodes accepting any given currency) is everything. The regulators have done a very good job of keeping crypto's contained within just 'speculation'. I feel the window of opportunity for it to spread has been shut (for now), mostly by the crypto community themselves with all the fraud and lack of any improvements in UX this long into the game.

While the crypto developers will happily hack away and fabricate use cases based on speculation, and refuse to reduce friction / fees and improve UX. Crypto in it's current form is dead, something new will arise with all the learnings of the current-generation crypto. As much as I believe in the mission of crypto. I'm confident that this proposal like all the other proposals refuse to acknowledge the problem with ethereum and any other crypto, which is I still can't buy bread with it. This proposal is another evidence to show that they're simply focused on the wrong set of problems.

Edit: Update team size mentioned above from 4 - 6 for accuracy.

> As much as I believe in the mission of crypto.

What is the mission of crypto and what is the most impactful application to date?

If the latter is underwhelming, perhaps that should make one question whether the mission is worthy.

> A purely peer-to-peer version of electronic cash would allow online payments to be sent directly from one party to another without going through a financial institution.

Depends on how you read this, but to me, this is a replacement for fiat, I should be able to pay my butcher / baker, that hasn't happened yet. If i walk to a merchant near my house they will all reject crypto in any form.

For one no cryptocurrency is really designed to be used for actual payments. They all are deflationary by design so the future value of cyber bucks will be higher than today's value. If you spend cyber bucks today you're losing out on those future gains. It's irrational then to ever spend cyber bucks.

Unless cyber bucks are accepted to pay taxes in your jurisdiction (they're not) they need to be converted to a currency that is accepted. Thanks to the deflationary nature of cyber bucks, their conversion rates fluctuate wildly and constantly. The conversion rate can even change while waiting for a transaction to complete.

To boot every transaction is caveat emptor, there's no protections against fraud or even simple errors. Access to a wallet's private keys (even if they're not your keys) confers access to the wallet and all of its cyber bucks.

If the goal of cryptocurrency is to replace fiat currency transactions it's like the worst possible design for the task.

Seriously, check out Monero.

Valid critique. Thanks.

Out of curiosity, in what country do you live?

Thailand

The company I work for uses USDC for quite a lot of B2B transactions and settlements. On the tech side it's miles easier to use and build tooling for than things like wire transfers through our bank. Seriously, I've gone back and forth with our big bank for months trying to get some simple API integrations and they're pathetic. For USDC payments, we quickly built a fantastic initiator/approval system and live monitoring on incoming and outgoing payments. I wish all payments were on crypto rails, it would make a lot of jobs easier.

> I wish all payments were on crypto rails, it would make a lot of jobs easier.

Not having any way of reverting transactions definitely makes the jobs of all sorts of shady characters much simpler, that's for sure.

Centre (the consortium behind USDC, including Coinbase and Circle) does have a mechanism to freeze shady accounts and accounts used to steal funds. Which has been used to freeze the TornadoCash funds when they were ordered to.

Yeah, cause we want to stick with the crummiest blockchain that repeatedly rewards the select few and never manages gas fees right....

breaking: ethereum ceo advocates to change consensus rules again and further incentivize centralization.

Unfortunately I don’t think that there is no amount of changes to blockchain that would make it useful in any way in the real world.

There are way too many negative externalities of crypto.

And after nearly 15 years, we should just stop this.

Yep. Crypto is a solution searching for a problem and it still hasn’t found it. Imo the best use case remains and likely will be drug/black market purchases. Otherwise, it doesn’t have enough practical advantages to beat traditional digital currency, philosophy aside.

Reply on news.ycombinator.com