Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Limits to Blockchain Scalability

vitalik.ca

Post reply on HN
Print view View original

91–100 of 465 posts

Re: The Limits to Blockchain Scalability

#91

Post by jollybean »

The notion of 'decentraliztion' as presented by the blockchain community is basically extremist/absolutist, and it's ruining their own projects.

Our entire civilization and everything in it depends on networks of trust. Without it, we'd fall down instantly.

To require a system of 'absolute decentralization' when 'partial decentralization' would work just as well, doesn't make sense.

We technical people have an odd way of falling into these kind of traps - we're trapped in theory, unable to map to the real world pragmatically.

DNS on the internet works on the basis of 'partial decentralization' and it works very well. There wouldn't be that much point in making it 'fully decentralized'. We could push it maybe a few inches more that way, but there's zero requirement to go 'fully decentralized' for it.

You get vastly diminishing marginal returns to effort when you start to go absolutist.

Much like Rust, which has some very nice outcomes but trades absolutely everything for those outcomes, some small tradeoffs in decentralization for Blockchain would yield probably some big benefits.

Re: The Limits to Blockchain Scalability

#93

Post by rawtxapp »

542458 wrote: post #11
javert wrote: post #6

You could also just not scale the blockchain, and instead use layer 2 solutions, which is the bitcoin approach.

BTC’s layer 2 method also doesn’t scale because (among other reasons) you still need on-chain transactions to handle opening connections and creating wallets, which the main chain does not have the throughput to do at scale. That’s also not to mention all the practical difficulties of using lightning (have to constantly monitor for fraudulent closing of connections, locks up liquidity, etc) and technical issues (rout…

Lightning channel factories will reduce the on-chain transaction requirements very significantly by batching many channels open/closes into a single transaction.

Re: The Limits to Blockchain Scalability

#94

Post by diegocg »

seibelj wrote: post #39
jeremycarter wrote: post #16

Blockchain, Decentralization and Smart Contracts have had nearly a decade to prove their value and disrupt the marketplace. So far nothing but whitepapers, pipe dreams and exit scams - nothing my mother can use. What's the point? Nobody is using crypto as an alternative to fiat. Prove me wrong.

I have responded to a thousand versions of your comment on HN. No matter what I write the haters never change their minds. HN is not an enthusiastic place for blockchain devs.

That's because on HN we know that 99% of the blockchain hype is about people discovering asymmetric cryptography and pretending that the blockchain is the only way to use it.

The only real difference that blockchain solutions bring is decentralisation. I have to yet hear one single argument about why decentralisation is good. So far I have only heard the typical libertarianesque arguments about states, banks and inflation. As as I'm concerned, decentralisation is unnecessary and so are blockchain.

Re: The Limits to Blockchain Scalability

#95

Post by wmf »

dmitriid wrote: post #82
gruez wrote: post #67

Earlier quoted context omitted.

This feels like a misunderstanding of how the syncing process works. You don't only download the last n blocks and assume the current state is accurate. You download the entire chain, validate it, and optionally discard the redundant information. As far as your client is concerned it still verified the entire chain end-to-end. The only difference with pruning is that you can't revalidate the chain end-to-end without…

what happens when everyone on the chain discards old transactions?

That would be bad. Archive nodes are needed.

Re: The Limits to Blockchain Scalability

#96

Post by tromp »

john_alan wrote: post #66
deadalus wrote: post #61

Earlier quoted context omitted.

Are there any other significant untraceable cryptos other than Monero and Zcash(using privacy fetures)?

I work in applied Cryptography so I’m well placed to judge. Other than Monero and ZEC (the latter of which has some issues for me which I posted about before) Grin is the only other project that isn’t a clone.

Grin is currently not untraceable (it only hides amounts and addresses, but not input-output links, which are mostly visible in the mempool), but could be if the coinswap proposal [1] is implemented and widely deployed.

[1] https://forum.grin.mw/t/mimblewimble-coinswap-proposal

Re: The Limits to Blockchain Scalability

#97

Post by Closi »

s17n wrote: post #65

If Vitalik is correct that the only viable defense against the attack described in this article is users validating nodes en masse, then blockchain is doomed, because you're never going to convince more than a small percentage of users to validate blocks. Personally, I think that custodial entities such as Coinbase (or in the future, banks) can probably be relied on to not do anything too controversial.

So back to centralisation of power?

Re: The Limits to Blockchain Scalability

#98

Post by noxer »

mvzvm wrote: post #26

Scam after scam, that's all blockchain is. Just another way to fleece the average consumer. There has not been a single valuable use, a single product, that actually improves anyone's day / process / life / anything. I am very open to changing my stance if someone presents evidence to the contrary.

https://ripple.com/ripplenet/on-demand-liquidity/

This is a real product, real people use it (without knowing that they do).

Re: The Limits to Blockchain Scalability

#99

Post by mrkramer »

"It's crucial for blockchain decentralization for regular users to be able to run a node"

Yea just like Larry Page is running Google server in his bedroom to help scale Google's infrastructure.

Nodes need to be in specialized server farms and compete with each other for profit. Capitalism is all about specialization and competition in order to improve quality not everybody doing the same thing and amounting to nothing at the end.

Re: The Limits to Blockchain Scalability

#100

Post by greg7mdp »

s17n wrote: post #65

If Vitalik is correct that the only viable defense against the attack described in this article is users validating nodes en masse, then blockchain is doomed, because you're never going to convince more than a small percentage of users to validate blocks. Personally, I think that custodial entities such as Coinbase (or in the future, banks) can probably be relied on to not do anything too controversial.

Looking forward to the bright future where we rely on centralized exchanges and banks!
Post reply on HN