Comment by uncomputation | Hacker News Reader

Comment by uncomputation | Hacker News Reader

An interesting problem that I haven’t seen addressed is that, despite the marketing, web3/defi/crypto projects will be antithetical to “builder”/hacker types. Decentralized protocols are slow to develop by nature due to many people depending on them. If you’re developing Yearn finance, for example, no one wants you to do anything but make sure that stays running and delivering optimal returns. Once you have a project with supporters, it becomes impossible to leave or modify. Not so with a centralized project which can adapt quickly and “pivot.”

Moxie Marlinspike talked a bit about this in his talk “The ecosystem is moving.” Where Signal could implement video chats any time they wanted, distributed systems and protocols are hoist with their own petard: consensus. You must get community agreement and development on the new direction for anything meaningful to happen. This is both the strength and weakness of decentralized systems and one that is less suited to hacking and more to code review.

I can't help but think about all the non-web3 protocols this applies to, like email, html, etc. It's slower, but it's more powerful. While Signal could implement video chat any time, true...we don't need another closed video chat.

Decentralized systems are more powerful if governed properly, absolutely. But they won’t attract hacker types because they require ongoing maintenance and community consensus. When was the last email upgrade? HTTP and HTML have remained essentially the same their entire lives. Most of the people on the W3C are veterans, not “builders.”

Also Signal is open source, unless you have a different meaning of “closed video chat.”

they don't necessarily require ongoing maintenance, it depends how they are coded. some might even say the "correct" way is a non-upgradeable back-end (on-chain solidity contracts) and hash-addressed, ipfs hosted front-end. anyone can then pin the front-end themselves to ensure access or just run the code locally. dapps should rely on centralised, server based components as that mostly defeats the purpose.

uniswap is a good example. when they release an upgrade users must explicitly adopt the new version. in fact uniswap v1 and v2 still see usage despite the release of uniswap v3.

The Internet standards were built by "hacker types".

This can be done through a hybrid though, go off, build video chat, get some adoption and then leverage that to get consensus. “I’ve already done it this way I’m not changing it” is a dick move in an average PR request, but effective.

I was going to include this in my original comment and then deleted it, but you just described Netscape before Internet Explorer. You may enjoy the history of the img tag: https://thehistoryoftheweb.com/the-origin-of-the-img-tag/

I don't think its related really and just happenstance. Its just a sign of market maturity that communities haven't accepted a survivorship or succession path that all organizations have to content with, instead of putting all their confidence on an unsociable do-it-all savant.

all of these founders should just leave en-masse and let people pick up the scraps. I love buying the dip on things that are misunderstood as codependant crypto traders freak out, has been my best trades.

(The current state of market maturity, market immaturity)

That's basically the BTC approach and it shows. Stable, but significantly more slower dev ecosystem. ETH, and by extension web3/defi are a result of builder/hacker types leaving BTC to get something more dev/app friendly.

What’s hilarious is consensus is effectively the path to “centralization” and build by committee.

When only a few coders are allowed to decide on design consensus, how a system works is out of the hands of users.

No different than politics and banks. Unfortunately for computer science minds, there is still a real world to crash into, not the endless void of infinitely big little numbers.

Reply on news.ycombinator.com