TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Ethereum Services Are Centralized

216 点作者 iamnotarobotman超过 3 年前

31 条评论

carlosdp超过 3 年前
Echoing the other comments, this is a pretty naive understanding of how a general purpose smart contract blockchain works and the incentive models involved that keep the kind of censorship described here from happening. They also don&#x27;t seem to fully grasp how smart contracts that are in-fact decentralized work, and are cherry-picking the ones that have an authoritarian controller because of the nascency of the technology.<p>The article isn&#x27;t really detailed at all, so it&#x27;s hard to know where to start with why these &quot;services&quot; aren&#x27;t in fact &quot;centralized.&quot;<p>It&#x27;s effectively the equivalent of someone saying &quot;you can&#x27;t keep your password secure from hackers if you store it in a database, if someone hacks it then they can steal it and use every site you used that password on&quot;: it&#x27;s just obviously false to people who know how this works, but now I have to explain how hashing works and such.<p>I&#x27;d just encourage reading up on how Ethereum transaction mining works before just taking the article&#x27;s claims at face value.
评论 #28844954 未加载
评论 #28844667 未加载
评论 #28845537 未加载
评论 #28845381 未加载
评论 #28844873 未加载
评论 #28844966 未加载
ojr超过 3 年前
Ethereum enthusiast are quick to disagree, the author of the article might not have a deep understanding of how everything works but headline of the article holds somewhat true.<p>Metamask is a centralized service with centralized team, if Google removed this extension from Chrome as they temporarily did before, it would make adoption for new users harder.<p>Open Sea has centralized operators, the orderbook is not on the blockchain I believe.<p>Uniswap the &quot;decentralized exchange&quot; has acentralized team with vc funding and registration with the SEC.<p>Infura is another centralized service, that so many Ethereum projects depend on.<p>It is hard to call these services decentralized. Although the small mainstream success of these services and markets, show that fully decentralization is probably not needed for most mainstream users.
评论 #28848375 未加载
评论 #28846242 未加载
评论 #28846880 未加载
woah超过 3 年前
This blog post is pure nonsense. Nobody can alter an Ethereum smart contract unless they control all full nodes, or can convince all full node operators to install a fork of the software that reaches into the database and modifies that contract.<p>It is possible for smart contracts to contain code that allows them to update themselves when receiving a message signed by a certain key, but if this is the case, then it is obvious to anyone inspecting the contract.<p>Most heavily used Ethereum contracts such as Uniswap and Compound for example do not contain any updating code because users do not trust it.
评论 #28846229 未加载
评论 #28845771 未加载
评论 #28848541 未加载
评论 #28845739 未加载
spicybright超过 3 年前
This will likely be controversial, but tech like bitcoin is a big step backwards from payment systems we currently have.<p>All the rules and regulations were created because someone was bad enough for people to agree on a rule to prevent something from happening again.<p>Block chain money loses all of those protections to stop a bad actor from, say, a hacker draining your wallet with no recourse.<p>So we find ourselves back to step one. I do think it&#x27;s inevitable layers of regulations and rules will crop up, either from governments, or power holders requiring layers of smart contracts to respect transactions.<p>It will just be a matter of time till the modern banking systems &quot;port&quot; themselves onto the chain.
评论 #28844690 未加载
评论 #28844646 未加载
评论 #28845288 未加载
评论 #28845058 未加载
评论 #28845146 未加载
评论 #28844812 未加载
评论 #28844692 未加载
评论 #28845599 未加载
评论 #28846928 未加载
评论 #28844664 未加载
factorialboy超过 3 年前
The moment a small group of people bailed out their friends&#x27; DAO, Ethereum lost all credibility. I am referring to: <a href="https:&#x2F;&#x2F;www.gemini.com&#x2F;cryptopedia&#x2F;the-dao-hack-makerdao#section-the-response-to-the-dao-hack" rel="nofollow">https:&#x2F;&#x2F;www.gemini.com&#x2F;cryptopedia&#x2F;the-dao-hack-makerdao#sec...</a><p>Since then Eth has proved to be a successful investment, so this is not a commentary on the performance of this digital asset. But rather a comment on the false immutable, decentralized claims of this blockchain.
评论 #28845497 未加载
评论 #28845653 未加载
评论 #28845193 未加载
评论 #28848489 未加载
politician超过 3 年前
Is Medium playing games with browser history? When clicking the back button after clicking &quot;read more&quot; it kept refreshing the same page. I hate when sites do that.
评论 #28844993 未加载
评论 #28846018 未加载
robcohen超过 3 年前
As other commenters have said, this article largely appears naive on how contracts work. Looking past that, the main thrust of the argument is that if &quot;the market&quot; wants things like chargebacks and censorship, then there will exist large blockchain &quot;startups&quot; that fulfill that need. They also make the argument that such a progression is likely inevitable.<p>What they fail to understand is that is not important. What is important is that blockchain enables people to vote through their choices. If they prefer decentralization, they can have that. If they prefer centralized features, they can have that. Until blockchain, it wasn&#x27;t even possible to have the choice of decentralization because it simply technologically didn&#x27;t exist. Crypto-hopefulls believe this choice will better society. Maybe it will, maybe it won&#x27;t. But even a subset of society can benefit from having more choices. Just because your candidate loses doesn&#x27;t invalidate the benefits of having a democracy.
betwixthewires超过 3 年前
&gt; Ethereum nodes can in theory filter modifications to the smart contracts they execute.<p>This is the crux of the point, and as far as I&#x27;m aware is incorrect. A transaction to change a smart contract must be signed, a node cannot modify a transaction. A miner can in theory refuse a transaction to change a contract, but the only way to reliably prevent a change is a 51% attack. A node could refuse to accept a change, but it would fork the network and the node would be running a non canonical chain, which makes it pointless.
评论 #28848448 未加载
togilvie超过 3 年前
This is a really bad take, with a limited understanding of how Ethereum nodes and smart contracts operate. I clicked hoping for more...
aazaa超过 3 年前
&gt; But this [knowing that an internet service will behave exactly as advertised] fails to hold true for general purpose chains acting as a VM, including the front-runner Ethereum. Ethereum nodes can in theory filter modifications to the smart contracts they execute, but in practice, node operators have no reason to inspect or reject these upgrades, because they have relatively little stake in the success of individual contracts or their ecosystems. If an owner of a smart contract publishes a new change, nodes will mindlessly run it. Ethereum is blockchain AWS.<p>The author fails to define &quot;VM&quot; here. The only thing that makes sense is &quot;Virtual Machine,&quot; which is a thing, but only barely makes sense in this context.<p>The fact that node operators have little incentive or ability to inspect or reject changes to contracts beyond what the protocol dictates isn&#x27;t a problem in itself. That is, after all, the entire idea. Code is law. It doesn&#x27;t matter how desperately a single miner might want to change the rules. At least in principle.<p>This is more of a problem on systems like Bitcoin where miners of side chains literally have no knowledge, from the protocol itself, of what the hashes representing transactions on other chains mean.<p>The bigger problem, not addressed in this article, is that the vast majority of things you want to do to make a block chain useful outside of financial transactions requires something called an &quot;oracle.&quot;<p>What does that NFT actually secure? Go consult the oracle. Want to know the exchange rate of some non-ethereum token? Go ask the oracle. Want to know what that land title says? Go ask the oracle.<p>The problem is that an oracle is just a server. Maybe it&#x27;s a group of servers - doesn&#x27;t matter. Servers are corruptible in ways that Ethereum is not (or at least should not be) through hash chains and proof-of-work. Servers use logins, admins, and undocumented security procedures. And they offer little protection against Sybil attack.
russellbeattie超过 3 年前
The post is mostly nonsense, but that said, I&#x27;m always astounded by how few nodes there are out there. Checking now and there&#x27;s just 2,713 of them - though this number is low because of a bug in the latest software which removed thousands in September for some reason. It&#x27;s never been higher than 7,000 or so.<p>How long would it take you to write a script which DDOS all of them? How many are run by organizations which can protect against attacks? My guess is not enough. Considering the amount of money relying on those measly ~3k servers out there? It&#x27;s truly frightening.
评论 #28848209 未加载
xrd超过 3 年前
This is pure garbage:<p>&quot;Ethereum nodes can in theory filter modifications to the smart contracts they execute, but in practice, node operators have no reason to inspect or reject these upgrades, because they have relatively little stake in the success of individual contracts or their ecosystems.&quot;
评论 #28847661 未加载
thesausageking超过 3 年前
&gt;If trust on the blockchain boils down to trust in a mere social system, what has been gained?<p>This has already happened once when the original DAO project got hacked and the community decided to undo all of those transactions with a fork. The changes on chain were completely valid from a code point of view, but the community around the project decided to undo it.
nootropicat超过 3 年前
&gt;If an owner of a smart contract publishes a new change, nodes will mindlessly run it.<p>Utter nonsense, contracts can be both immutable and mutable. There&#x27;s inbuilt owner capability, code has to explicitly implement upgrade capability. All nodes do is run code.<p>Uniswap for example has no upgrade capability.
评论 #28847618 未加载
crazydoggers超过 3 年前
This lacks an understanding of how Ethereum contracts can function. He’s assuming that a single person or entity must own a contract, and that’s not true.<p>That’s the whole point of DeFi. Uniswap is a great example, which is a fully decentralized exchange.
评论 #28848238 未加载
X6S1x6Okd1st超过 3 年前
Find people that are skeptical about the benefits of blockchains&#x2F;cryptocurrencies AND have a strong technical understand is so rare. This is not one of them.<p>So far the best person that has that intersection I have found is Angela Walsh.
Geee超过 3 年前
True decentralization means that every end user can choose which version of the software they want to run and how the software runs. This is only possible if they <i>can</i> run the software on their own hardware, in the event that they don&#x27;t agree with the remote service provider.<p>That&#x27;s why general purpose blockchains like Ethereum are doomed to fail. They can never be both competitive with centralized trust-based services and lightweight enough to be run by end users.
评论 #28845643 未加载
dabedee超过 3 年前
Not to remove from some of the valid concerns made in the article, but I find it ironic that a post decrying centralization of Ethereum services is written on Medium.
评论 #28848558 未加载
trident5000超过 3 年前
Price is going up. Looks like its time for the daily anti-crypto articles on HN now.
chmsky00超过 3 年前
It’s centralized in that the developers are now bankers; we have to trust and empower them with a Byzantine abstraction to do daily routine?<p>We’ve already seen how that worked with bankers. Nope.<p>I’m really looking forward to the day computer science folks finally accept if it involves human participation it will come along with human bias.<p>Biology has not gone anywhere, folks. We cannot decouple ourselves from this existence and be information; yet? Maybe. For sure not now.<p>And networks like Ethereum are not going to enable that.<p>It may be a nice and clean abstract in their head, but everyone has that privilege. Bringing people aboard Ethereum is putting a new cognitive burden on people. But the goal is to not do that?<p>Environment issues aside, blockchain is like nuclear powered rocket car hype and not anywhere near as cool.<p>I’ll stick to a Wireguard tunnel between people I trust literally before running this stuff for a bunch of randos profiting big off my effort.
dwmbt超过 3 年前
after reading the article and some comments, i thought i would comment my own grievance with the article. i think there is a disconnect between the quantifiable trust a blockchain offers and the authors interpretation. specifically, he mentions that &quot;the code these nodes are running is available for anyone to inspect (trust)&quot; and that &quot;smart contracts can be trusted to a large extent because they operate in the open, are kept in check by the ecosystem, and even often have voting protocols in place to govern changes.&quot;<p>yeah, that&#x27;s actually a good enough explanation of what&#x27;s going on there. yet, this doesn&#x27;t seem to be enough for the author. i guess my biggest annoyance is that they seem to think a smart contract is just a series of if&#x2F;else conditionals that can be ported onto an AWS instance.
cryptica超过 3 年前
The basic premise is true. Ethereum nodes do not provide even basic search&#x2F;querying features out of the box so you have to rely on centralized services to index the blockchain data for querying. So if you want to build anything useful which integrates with Ethereum in any way, you need to trust a centralized service provider to accurately report the data... I was shocked when I tried to integrate with Ethereum for the first time. It almost seems as though they have made it difficult to integrate with on purpose.<p>It&#x27;s been over a decade. Couldn&#x27;t they have written some kind of open source service which exposes the Ethereum node&#x27;s data via a simple REST HTTP API? This would not be a difficult project to implement.
评论 #28845074 未加载
评论 #28848279 未加载
newacct583超过 3 年前
This wasn&#x27;t making sense to me, then I got to the meat:<p>&gt; <i>The real risk is that blockchain Twitter will begin to censor dubiously defined “hate speech”, or that blockchain Mastercard will add support for chargebacks.</i><p>This is just culture warism. In fact moderated discussion forums[1] and reversible payments are <i>DESIRED FEATURES</i> of products in the real world. Of course that&#x27;s going to happen. But to the author that constitutes &quot;centralization&quot;, but what it really means is &quot;not the libertarian utopia I imagined&quot;.<p>[1] We&#x27;re posting on one!
__MatrixMan__超过 3 年前
I think we need to start being more specific than &quot;decentralized&quot;.<p>My critique of most crypto projects is that they aren&#x27;t partition tolerant. Does that mean that they aren&#x27;t decentralized? I don&#x27;t know.<p>I&#x27;d love a taxonomy of decentralization techniques so I can say: &quot;crypto project X has decentralization characteristics Y and Z, but only Y is consistent with its stated mission so it&#x27;s less viable then project W which...<p>But instead we just decentralize=good centralize=bad and continue with politics as usual.
评论 #28847363 未加载
samaman超过 3 年前
To respond to the title: yes, yes they are, if you consider how many live in infura or an AWS geth instance and&#x2F;or quorum by JPM. Doesn&#x27;t mean its not a better means for data security...just that yes, the publisher of said smart contract can just replace or change to contract if they really wanted to if it were not part of the mainnet. There are many infura nodes though that are part of the mainnet, so that&#x27;s unlikely unless some company just forced all users to their own ETH network.
vmception超过 3 年前
Odd and inaccurate take. Some or many services have admin control. But it doesn&#x27;t matter.<p>If you are exchanging time for food and shelter, you need to be building smart contracts. That&#x27;s where we are right now, it is the most lucrative use of time possible.<p>Put down the Shopify widget development book, put down the dropshipping tutorial, and start deploying smart contracts.
seibelj超过 3 年前
Decentralization is a spectrum. A truly decentralized network, one pure from every angle of analysis, does not exist.<p>Ethereum itself is more decentralized than all traditional value-transfer networks of the old centralized financial world, and the applications built on top of Ethereum are also a spectrum of decentralization.
评论 #28844666 未加载
评论 #28844477 未加载
vmception超过 3 年前
There are plenty of smart contracts where the admin destroyed their access, and where the admin never had any privileged access.<p>Are people so out of touch on this topic that they would ask &quot;source?&quot; after being presented with negative information that matched what they wanted to read
modeless超过 3 年前
&gt; If an owner of a smart contract publishes a new change, nodes will mindlessly run it.<p>This is only true if the contract specifically allows that, right? And that&#x27;s disclosed in the contract itself. So if you don&#x27;t want that, don&#x27;t use that contract.
Animats超过 3 年前
How many miners control 51% of the Etherium hash rate? That&#x27;s the key issue.<p>For Bitcoin, it was at one time 5, all in China, and there&#x27;s a group photo.
评论 #28844807 未加载
rgovostes超过 3 年前
Meta: Unsure what is happening on the front page today. This isn&#x27;t even a link to the article[1], and somehow it rocketed to #10 in under an hour. A nearly 2-year-old announcement about SHA-1, posted 20 minutes ago, is in 3rd place.<p>1: <a href="https:&#x2F;&#x2F;michaelgummelt.medium.com&#x2F;crypto-services-are-neither-decentralized-nor-trustworthy-f43c0a19a400" rel="nofollow">https:&#x2F;&#x2F;michaelgummelt.medium.com&#x2F;crypto-services-are-neithe...</a>
评论 #28844417 未加载
评论 #28845794 未加载
评论 #28844596 未加载