Blockchains Considered Harmful: Is Brute-Force Processing Replacing Good Design?

Hi. I&#x27;m Vinay Gupta, the release coordinator for the Ethereum launch.<p>Firstly, nobody believes that Proof of Work is here to stay - it&#x27;s bleeding money from the currencies that use it at an astonishing rate, and as soon as Proof of Stake (or other algorithms) can replace it, they will. PoW is a direct financial drag on these economies, and it will not last long. Bitcoin will probably take longer to clean up its act than Ethereum, but that&#x27;s largely for political reasons, not technical ones. Technically, it should be a lot easier to do than Ethereum, in fact. (It helps not being nearly Turing complete.)<p>Secondly, brute force is how things begin. The Unix philosophy has always suggested using brute force first: premature optimization is the root of all evil, as they say. We are at the very earliest stages of designing global public heterogeneous parallel supercomputers, and we should not be surprised that the early approaches are brute force.<p>It won&#x27;t be that way for long.

I agree, there are much better consensus methods than Proof of Work. It&#x27;s just that PoW is by far the most simple to implement. If you just take a peak at the ecosystem, especially around 2013-2015, it seems like every &quot;altcoin&quot; project had created their own consensus system. Of course, many of these were terrible.. but it&#x27;s a good sign that this space has been trying to get away from regular ol&#x27; PoW since people started really finding out about Bitcoin.<p>Peercoin invented the core basis of the PoS algorithm in 2012 that was used throughout the 2013 and 2014 altcoin craze. This is still being used today in various projects, in augmented forms. Some projects tried to incentivize not just holding coins, but also making transactions (PoSV&#x2F;proof of stake velocity, IIRC). Some projects tried making it so that you must burn coins to create blocks (proof of burn). And some projects tried doing genuinely useful computations instead of plain proof of work, such as doing protein unfolding (I think this was Curecoin?)... And then Dash invented the dBPF algorithm which is used by numerous coins today beyond just Dash and is similar to PoS but with a set number of stakers. (usually elected through an election on the blockchain)<p>If you look beyond Bitcoin there is a ton of innovation taking place in this space, and there has been for some time now. A lot of it is mired in scams, trolls, shills, and FUD.. but there is some genuinely awesome innovation happening in this space to drive forward consensus algorithms.<p>(source: co-founder of Qtum, a project in this space)

&gt; “Proof-of-Work” exists because money is being created. It is, then, impossible to “create a new form of money” without invoking Proof-of-Work.<p>&gt; Satoshi’s design insight was to channel that inevitable work into a cumulative process, to optimally stablize a peer-to-peer clock in a cartel-resistant way.<p>– Paul Sztorc <a href="http:&#x2F;&#x2F;www.truthcoin.info&#x2F;blog&#x2F;pow-cheapest&#x2F;" rel="nofollow">http:&#x2F;&#x2F;www.truthcoin.info&#x2F;blog&#x2F;pow-cheapest&#x2F;</a>

It always bothered me how computer engineers who try to conserve energy and optimize all the time were excited by the wasteful Bitcoin implementation only because they see a potential in it to increase their wealth. It&#x27;s like glorifying random sort because we haven&#x27;t invented the bubble sort yet!

PoW sticks because as a consensus backbone because it&#x27;s 1.) easy to understand and implement and 2.) it&#x27;s guarantees distribution of updates to the distributed state. IMO there are better ways to trustlessly secure distributed state but like all things there&#x27;s a trade-off. The dirty secret of blockchain tech is that it can&#x27;t scale - not without supplementary protocols. Even when you get it to scale you still have to deal with the storage of the blockchain. Bandwidth constrains the size of the blocks, which in turn constrains the amount of transactions. Turn up the block creation frequency (a la ETH) and you solved one problemt (Tx throughput) but created another (chain size).<p>I am fundamentally opposed to PoW because it wastes so many resources, but, for now, it&#x27;s the most viable way to secure a chain. PoS is doomed to centralization and relies too much on game theory and incentives for my tastes. just my 2 cents

This person is effectively arguing that all of Bitcoins problems would be solved by just using a centralized system.<p>This is correct, but it misses the point.<p>The whole point of crytocurrencies is to order transactions WITHOUT a centralized system.<p>There are many centralized systems already. If you want the benefits and tradeoffs of those, just go use one of them.<p>The people who do NOT like those tradeoffs will stay with crytocurrencies.

I agree with this and also fundamentally believe that trust is a feature in systems and not a bug.<p>Building systems that work without trust (supposedly) is of very limited social utility since increasing trust is what makes for a healthy society and healthy economy.

Nice article. It is unfortunately woefully light, if not completely absent of alternatives. Yes, it is a massive waste of energy, but the ability to both prove and simultaneously detract counterfeits is brilliant as far as the creation of a store of value is considered. In the long run, I believe, that the optimization of ASICS + slowly decreasing power in the hands of a much larger set of miners will be the long-term winning factor defining legitimacy for POW algorithms. Until then the one who can burn energy faster and cheaper (a baseline commodity to begin with) will win.

One thing governments could do is this:<p>- Create a digital token system that can be exchanged and accounted freely via a centralized server.<p>- Give it a name. Don&#x27;t even call it a currency, let people use it as money <i>if they want to</i>.<p>- Give a limited amount to any citizen who wants some. Or maybe sell it, whatever. Just figure out a way people can get some. It can&#x27;t be so hard.<p>- Promise there will never be more than some fixed amount in total.<p>Only governments could do something like that because despite everything, most people trust them about money.<p>It&#x27;d be kind of a mixed between the different money systems.

Vinay Gupta thinks &quot;nobody believes PoW is here to stay&quot;. Other commenters think &quot;PoS might be impossible &amp; no prominent cryptographer has some out in support of it&quot;.<p>It&#x27;s worth reminding people of the null hypothesis: Perhaps blockchains don&#x27;t work.

I don&#x27;t like proof of stake because a large stake holder could be threatened privately by a third party into corrupting the whole system. With proof of work you need to spend a lot of money very publicly for a long time to break it.

Proof-of-Work gives the token value because something was lost in order to create it

It always puts me off when an article has &quot;Considered Harmful&quot; in the title.

&quot;considered harmful&quot; meme is tired and cliched, please retire.

&gt; The underlying technology of blockchains to achieve distributed consensus has been touted as a solution to many challenges in decentralized systems.<p>I hate blockchain hype and I hope the whole bubble around it dies soon, but even I can tell, judging from the title of the article and just the very first paragraph, that the author doesn&#x27;t know a shit about how the thing works or why it is designed the way it is.<p><i>Blockchain does not solve consensus problem</i>. It&#x27;s a timestamping service. Not a consensus solver.<p>And then the author draws dumb statements about what blockchain is from all over the place. Why the hell would anybody think &quot;proof of work&quot; is a &quot;solution to denial-of-service attacks&quot;?