Live data from Hacker News

What to do once you admit that decentralization never seems to work (2018)

hackernoon.com

41–50 of 148 posts

Re: What to do once you admit that decentralization never seems to work (2018)

#41
post #18
post #13

Earlier quoted context omitted.

The Web and email are both decentralized. The Internet is decentralized to the extent there's no single "Internet company" which can kick you off. Various companies try very hard to make you forget these things.

Almost no new B2C web company can work without Google.

That's a pretty bold claim, care to elaborate?

Re: What to do once you admit that decentralization never seems to work (2018)

#42
post #34

Decentralization the way the crypto crowd wants can only happen if everyone has a server at home, some consumer-grade ARM box that you plug in and forget. It should be configured to work out of the box without the customer ever having to use the shell. • your own Nextcloud server to replace Dropbox, GCal • email server (replace Gmail) • torrent seedbox (replace Netflix, Spotify, etc.) • personal Bitcoin node (replace…

> You lose out on the discoverability that centralized hubs give you.

This isn't a real trade off. Consider how the web traditionally worked. Everyone has their own website, then you have search engines like Google or Yahoo that allow you to find anything anywhere.

A third-party search engine doesn't have most of the centralization problems you get with a Facebook or a YouTube, because people only use them to the extent that they're good, which means returning the search results people want. Moreover, if YouTube goes bust, all your videos disappear with it, but if Altavista goes bust, Yahoo and Google continue to exist and index all the same information.

You can have things that index all the other things without them being a single centralized monopoly.

Re: What to do once you admit that decentralization never seems to work (2018)

#43
> You lose out on the discoverability that centralized hubs give you.

Obviously you can have decentralized services that address this important 'presence' issue.

IMO the issue is that we have conflated the 'functional goal' with 'architecture'. The 'goal', for me at least, is public information systems that protect civil rights and resist the rise of tyranny. That goal, after thinking about this space for the past 25 years, can be served with a combination of centralized and decentralized services, and a mix of grassroots and corporate players.

To get there, we need (a) societal consensus that privacy in fact matters quite a lot to, and is a bedrock of, our civilization, and (b) [the recognition that] ideological bias is a deadly sin for a system architect.

Re: What to do once you admit that decentralization never seems to work (2018)

#44
"The apparently free, participatory open-source software communities have frequently depended on the charismatic and arbitrary authority of a “benevolent dictator for life,” from Linus Torvalds of Linux (who is not always so benevolent) to Guido van Rossum of Python"

GRAB. FORK. It's all there. Nothing stops you. Sleepless nights, stay on your toes.

Nothing stops people from complaining either. And if you fork, people who didn't will begin complaining about you.

Re: What to do once you admit that decentralization never seems to work (2018)

#45

"The apparently free, participatory open-source software communities have frequently depended on the charismatic and arbitrary authority of a “benevolent dictator for life,” from Linus Torvalds of Linux (who is not always so benevolent) to Guido van Rossum of Python" GRAB. FORK. It's all there. Nothing stops you. Sleepless nights, stay on your toes. Nothing stops people from complaining either. And if you fork, peopl…

Are you arguing about the centralized role Linus or Guido have? I think what you point out, the lack of a formal requirement for centralization just makes the point that centralization emerges from these networks anyways a bit stronger. And hence the point that we need to build accountability (git annotate for source) into the new systems is strengthened.

Re: What to do once you admit that decentralization never seems to work (2018)

#46
post #42
post #34

Decentralization the way the crypto crowd wants can only happen if everyone has a server at home, some consumer-grade ARM box that you plug in and forget. It should be configured to work out of the box without the customer ever having to use the shell. • your own Nextcloud server to replace Dropbox, GCal • email server (replace Gmail) • torrent seedbox (replace Netflix, Spotify, etc.) • personal Bitcoin node (replace…

> You lose out on the discoverability that centralized hubs give you. This isn't a real trade off. Consider how the web traditionally worked. Everyone has their own website, then you have search engines like Google or Yahoo that allow you to find anything anywhere. A third-party search engine doesn't have most of the centralization problems you get with a Facebook or a YouTube, because people only use them to the ext…

Maybe someone can come up with a decentralized search infrastructure. Something agent based perhaps. Otherwise the search is both centralized and in the position to make money from ads.

Re: What to do once you admit that decentralization never seems to work (2018)

#47
post #34

Decentralization the way the crypto crowd wants can only happen if everyone has a server at home, some consumer-grade ARM box that you plug in and forget. It should be configured to work out of the box without the customer ever having to use the shell. • your own Nextcloud server to replace Dropbox, GCal • email server (replace Gmail) • torrent seedbox (replace Netflix, Spotify, etc.) • personal Bitcoin node (replace…

Fraud is probably a harder problem than discoverability. You can always exchange (maybe through an automated pairing) IPv6 addresses if nothing else. Supporting chargebacks, eliminating spam, blocking phishing, etc. if both harder and more important to a functioning culture.

Fraud and spam are two separate problems.

One of the best solutions for spam for a new system is to impose a proof of work requirement on sending the first message between any pair of users, or any (new username, hosting service) pairing for federated services. For any normal pairing that happens only once, but for a spammer they have to pay the cost for every message or else you can trivially block every message from that sender.

Fraud protection is harder, but it's also something that lends itself well to being an independent service. If you're buying from someone trustworthy or dealing in small value amounts then you don't need it. If you're not, you require the seller to carry insurance from someone you do trust (or don't buy from them), and then it's the insurer's problem to deal with fraud.

Re: What to do once you admit that decentralization never seems to work (2018)

#48

The author of the article wrote a book ‘Everything for Everyone’ that looks interesting. It is about organizing coops as a push back against the gig economy or large corporations. My first job was at what was essentially a huge coop: the 100% employee owned defense contractor SAIC. SAIC spoiled me for all other jobs I had because there was a good vibe and efficiency that came with all employees having a strong motiva…

Thanks. Added to reading list. Also, nice tidbit about SAIC. It seems that many coops get coopted. Suggesting there's some as yet unknown magic sauce that makes some coops more durable. I'm very interested in comparative governance models. All tips, links welcome. I just want someone to point at some real world working organizations. Kinda like how QA obsessives study the Toyota Production System. -- I joined a FOSS…

Have you read anything by Pieter Hintjens, the creator of ZeroMQ?

He wrote quite a lot on organising people, particularly software development, but instead of focusing on how to make people efficient, he focused on how to organise the group to protect from opportunistic outsiders / psychopaths coming in and taking personal advantage and destroying the group in the process.

After having his early work ruined several times by other people taking unearned credit, taking money out, causing projects to fail, ZeroMQ was organised around his ideas.

“Coops that get coopted” or are less durable, might be failing to protect their structure well enough.

Re: What to do once you admit that decentralization never seems to work (2018)

#49
post #45

"The apparently free, participatory open-source software communities have frequently depended on the charismatic and arbitrary authority of a “benevolent dictator for life,” from Linus Torvalds of Linux (who is not always so benevolent) to Guido van Rossum of Python" GRAB. FORK. It's all there. Nothing stops you. Sleepless nights, stay on your toes. Nothing stops people from complaining either. And if you fork, peopl…

Are you arguing about the centralized role Linus or Guido have? I think what you point out, the lack of a formal requirement for centralization just makes the point that centralization emerges from these networks anyways a bit stronger. And hence the point that we need to build accountability (git annotate for source) into the new systems is strengthened.

Didn't Guido just recently remove himself as a central authority?

Re: What to do once you admit that decentralization never seems to work (2018)

#50
post #13

> decentralization never seems to work Not true, and there is many technologies, both new and established, which are fundamentally decentralized. Bittorrent is the first example that springs to mind. Cryptocurrencies are another example, etc.

The Web and email are both decentralized. The Internet is decentralized to the extent there's no single "Internet company" which can kick you off. Various companies try very hard to make you forget these things.

I wonder if the case isn’t that we don’t allow new technologies to be truly decentralized. Email and web came from a time before... our market/startup forces in play today may not yield a new, true decentralization because the urge to monetize is stronger now than before.
Post reply on HN