Hacker News new | past | comments | ask | show | jobs | submit login

The problem is that this gospel has been said for almost a decade now and despite, literally, billion dollars of assets that Vitalik has, and others flowing this is not happening.

Not saying that this could not happen as an hypotheses but cryptocurrency foundations are far far from business execution basic practices.

As an insider I can say that most money flows to a very small group of people and the governance is not really decentralized. For example, very few people can decide on Bitcoin and Ethereum protocol changes, and these people cannot be changed...




It does seem slow, but they did manage the proof-of-stake transition pretty smoothly despite delays and widespread skepticism, so I give them some credit for that.

I have no idea what Vitalik is funding. Do you?


He was talking about proof-of-stake since the beginning and every year was talking about the next year. Not blaming Vitalik himself but the whole thing, it is bad to give false expectations.

Another thing, no pun intended, is that the proof-of-stake upgrade maintains prohibitive the network fees for transactions while other technologies have low fees.

> I have no idea what Vitalik is funding. Do you?

The funding of projects is through the foundation but if I remember well the original people and contributors received the ~50% of the total ethers until now.


> proof-of-stake upgrade maintains prohibitive

Consensus algorithms have nothing to do with transaction fees.

> while other technologies have low fees.

Any "Ethereum killer" that showed up turned out to have the same if not worse problems as Ethereum in the moment they started dealing with minimal real-world traction.

> contributors received the ~50% of the total ethers until now.

First: source?

Second: "50% of total ETH until now" is doing a lot of work here. How much was during the pre-mine and how much was due to the sale? The pre-mine sale raised < 20 million USD. Are you counting the people who bought ETH in the pre-mine as "original contributors"?


> Consensus algorithms have nothing to do with transaction fees.

Please don't tell bullshit. Look at Algorand and other protocols, consensus has a relationship with fees because it is linked with the cost of reaching consensus!

You can even read that in the Ethereum subreddit [0].

> Source?

It is repeated ad nauseam in Internet [1] and you can analyze the blockchain genesis to check it.

> Ethereum killer?

It is not about the protocol but the community you create. Algorand has solved the PoS before Cardano and Ethereum but they are #58 now and the creator is one of the parents of modern cryptography, Turing Prize, etc. Solana is #5. Beyond comparing the Solana protocol with Algorand it is a matter of "business" execution, technology is a smaller part. Probably if Livra from Meta was accepted it would be in the top 10.

Even when you think about Solidity as a programming language, it was not well designed (e.g. security) but that doesn't matter.

[0] https://www.reddit.com/r/ethereum/comments/ru9dsq/the_proof_...

[1] https://www.google.com/search?q=how+much+the+original+contri...


I asked for "source" because I know that this is "repeated ad nauseam in the Internet" while being provably false.

The very first result on your google query is a bitcoin.com page that is 404, but archive.org has this:

  The Ethereum network started off with a supply of 72 million Ether (ETH).   
  Eighty-three percent of that (60 million) was distributed to people who had 
  purchased ETH in a crowd sale that was conducted in July and August of 2014.


  (...)

  Of the remaining 12 million ETH distributed at the launch of the network in 2015, 
  half was split amongst 83 early contributors to the protocol based mostly on time
  contributed. The other half were set aside for the Ethereum Foundation.

So, the "50% to contributors" is actually 8.33%.

> consensus has a relationship with fees because it is linked with the cost of reaching consensus!

Wrong. Fees are determined by network activity and the amount of transactions competing to get into the block being "mined". The cost to validate a full block is not really different than the cost to validate a block that is not completely full.

If Algorand or Cardano ever got close to the transaction volume from Ethereum, you can bet that their average transaction fees would go up accordingly.


> If Algorand or Cardano ever got close to the transaction volume from Ethereum, you can bet that their average transaction fees would go up accordingly.

I'm not sure what you mean by this. I don't know about Algorand or Cardano transaction volume, but many EVM-based blockchains process a similar number of transactions to ethereum (or more), with lower fees. They do all have different (proof of stake still) consensus models though

For comparison: https://etherscan.io/chart/tx

Polygon: https://polygonscan.com/chart/tx

Polygon is an L2, so arguably not as decentralized.

But then there's Avalanche: https://avascan.info/stats/network-activity

Or Fantom: https://ftmscan.com/chart/tx


Let's get Polygon out, because they are not a base-layer blockchain.

> a similar number of transactions to ethereum (or more), with lower fees.

Are we talking about the base currency (Wei) or the dollar-equivalent amount? If Wei, the only way that the transaction fees can be lower is if the chain has a different set of costs for the operations.

If you are talking about the dollar-equivalent amount, then yes, transactions are going to be "more expensive". But even then, it is not related to the consensus algorithm and just the "price of the base token".


Transaction volume meaning.. the number of transactions processed by the network?

If you meant monetary volume, you should have used a different term than one which is well-recognized to refer to the number of transactions (both in and out of blockchain applications of that term)

edit: I see, you're suggesting the fees are cheaper because the token is cheaper, and somehow seem to think EVM networks will have a straightforward relationship between the number of transactions and the cost denominated in their gas token.

I don't see how this follows. The fees are entirely a function of network constants and usage, which have more to do with what people are willing to pay to get their transaction into a block.

Ethereum has a limited amount of block-space, and a fixed number of blocks per year. The gas price isn't entirely a bidding system, because there's basically a floating multiplier which adjusts automatically based on the "fullness" of the most recent however many blocks, but the principle is that you need some form of congestion control

In blockchains which have larger blocks, or more numerous blocks, or a number of blocks/block-size which adjust based on usage, it is not as costly to get a transaction included.

So I don't know about Cardano or Algorand, but many networks can handle as many transactions as ethereum while having much cheaper transaction fees, which seemed to be the point you were arguing against


You know that Algorand and Solana supports a bigger number of TPSs that Ethereum and with lower fees and different consensus mechanisms, if you don't know that I am talking about someone that tries to show expertise but don't have any real one. It is a fact.

Initial investors are also contributors. The number allocated initially is really huge.


I also know that Solana had frequent outages and issues where they could resolve their transaction sequence, because their hardware and connectivity requirements make it prohibitive for "casual enthusiasts" to run a node.

And if "initial investors are also contributors", then you are just parroting the "Ethereum is pre-mined" from Bitcoiners, and we can safely end the discussion here.


Again you are not following the argument, you cherry picked Solana, Algorand didn't have any issue. Follow logical argumentation please...


I didn't mention Algorand because its overall network is a blip compared with Ethereum and Solana in any metric, and it can barely be considered as Battle-Tester in a "real world" scenario.


Do you have a paper for that?


TVL for Algorand, 111M USD: https://defillama.com/chain/Algorand

TVL for Ethereum, 53B USD: https://defillama.com/chain/Ethereum

If doing DeFI on Algorand is so much better/cheaper than on Ethereum, then why is it only 1/500th of Ethereum's size?

Mind you, it seems that this calculation is not attributing "bridged" TVL as related to the chain size and activity. If it were, there are 330 Billion USD that depend on Ethereum's base layer security.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: