This is your regular reminder that if you want a decentralized anything, you need to solve the spam problem again, which means censoring content.
Blockchain tech is good at one thing: making numbers expensive. That property has made them really unpopular, along with doing it by turning huge amounts of electricity into heat, but it's useful in that an identity is just a number, and if you can make identities allodial and somewhat expensive, you can solve the problem at the user level, by banning bad actors.
If a user ID costs ten bucks, spammers have to make that plus labor back before everyone shuns them. That's discouragingly difficult to do.
Bitcoin's proof-of-work system is inspired by Hashcash.
But yes, the solution to spam is to make it expensive: either monetarily or energetically.
> the solution to spam is to make it expensive: either monetarily or energetically.
This is used nowhere in practice? In practice, the solution is for a small number of dedicated central spamfighters to cooperate on lists of known spam IPs, or correlate across a large number of emails to produce statistical markers of spam.
These days the performance difference is about less than 10x on the cheapest of phones. A hashcash that takes 100ms on my PC and 1s on my phone to send an email would be prohibitively expensive for any spammer.
(The "inventor" of HashCash is also responsible for the corporate take over of the GitHub repository of Bitcoin Core).
If not Chia coin works that way, maybe that algorithm could be used? In either case it would make ASIC useless.
Here, I've googled it for you: https://en.m.wikipedia.org/wiki/Hashcash
I also don't ever remember having received an email with Hashcash.
The biggest mistake Satoshi made was including a reference to HashCash in his white paper
Fortunately I don't really rely on my gmail email for anything important, but it's filling up with that spam and marking it spam doesn't make any difference.
Interestingly I don't get any of those on my primary email which is self-hosted (and the address is a lot more public and older than my gmail address).
I'm currently trying to get rid of them by filtering the messages. I do
Settings --> Full settings --> Filters and Blocked Addresses
Create a new Filter
Add the display name or similar to the 'Includes the Words' field
Create Filter
Select 'Delete it'
> Spam needs censorship?
Spam needs to be deleted or hidden by someone or something or your inbox/feed will be full of spam. Trying to manage this individually with keyword filters or even the more sophisticated Bayesian anti-spam is exhausting.
> But isn’t spam only profitable at scale?
Yes. So what? Spam scales. It's like cryptocurrency, you burn trillions of FLOPS of computing power to make a few cents.
> Can’t you just charge a nominal fee or something?
Fee? For Facebook? This absolutely kills customer acquisition. It has often been suggested and never worked; the only site I think that does it is Metafilter?
Facebook went in the other direction with subsidised data plans that were free to access Facebook. The economics does not permit them to charge the eyeballs to look at Facebook. On the other hand, they can't afford to let people advertise for free on Facebook, since that's the service they do charge for.
In practice this is why accounts get tied to phone numbers so often, because it's slightly more difficult to acquire another few thousand phone numbers after the first thousand are banned for spamming, than it is with IP4 addresses.
My own personal gimmick ideas for twitter, that will never be tried, would be "premium" which gives you abuse reporting that actually works (and maybe some other stuff), and a system to enable people to pay for premium (or part thereof) for other users.
There is also the fact that it can be hard to see the overall gain when each interaction is very marginal. For example, updating your timeline with a photo and a status by itself doesn't seem worth paying for even if the cummulative effect over an entire year could be very beneficial for your networks.
But there’s also nothing inherently wrong with censorship. It just tastes extra bad when the things we like get censored.
There’s censorship right here on HN for example and that’s OK.
Presumably you would hardly see messages that haven't been approved by someone who has been transitively approved by you. Decentralized systems are usually networks, and if you got spam or uninteresting messages you would reduce the weight of that part of the network.
You can't trust a centralized party to censor for you, because then that entity has power over you through what you see. It has to be your secretary, i.e. it has to actually work for you and be a reliable servant. What better to use as a secretary than network of people connected by who is interested in the output stream from whom?
Sure you could install filters, but decentralised networks in practice has spam (the only ones that don't have any spam and don't have any filters are those too niche that spammers won't bother).
In a distributed system of the kind I propsed above, if you got a spam message you wpuld reduce your trust in those who transitively approved the source of the message, immediately.
Someone spamming would be cut off quick, as would anyone with high trust in a spammer.
You've fallen into the naive assumption that spammers use the same identity for each piece of spam being sent out, which is simply not true. I get plenty of spam phone calls with fake phone numbers despite the fact that STIR/SHAKEN has supposedly been implemented. Forged digital identities are cheap. And if forged identities are cheap, there's no trust mechanism that can work.
Presumably you'll counter by saying that you can simply assign a negative trust to unknown identities. In the real world there are plenty of us that have to communicate with people we haven't communicated with before (my business can't afford to turn away customers I don't already know), so we can't simply drop unknown identities.
The spam problem is hard because identities are hard, and digital identities are doubly so. Time and time again has proven that simple hand wavey solutions do not work.
A message from an entity which has not yet sent a spam message does indeed have the save priorization characteristics as one from any other entity withour connections to anything, but presumably a new user is invited by someone.
Furthermore, a new user could presumably say things on the spam-filled rubbish subforums, until he gets an upvote from someone who is trusted by somebody, and then after a while he becomes a participant.
The readon you see problems is, I think, that you imagine a system of private, individual, communication, while I imagine an enormous public forum.
Writing code is dead easy if you remove all the requirements from users.
Furthermore, among high-trust users PMs can be permitted even on such systems; and I csn actually give a superb procedure for dealing with spam in PMs among moderately trusted users: by letting a receiving user move a PM into the public flow of the sender.
Plus you still have the problem of how to treat an individual that is a first time poster. On many forums I never bothered creating an account until I felt I had something useful to contribute. In your system, how do you distinguish between the first time poster making a beneficial addition to the conversation versus a first time spammer? I feel you're hand waving this really hard problem away.
The Donald was genuinely popular. Some people didn't want it to be, but it was. I believe that this contributed more to its banning than any incidents relating to it.
The problem you're bringing up isn't really a solvable problem unless you have a human secretary as intelligent as yourself, and then it would be as appropriate for you to be the secretary for the person you'd have as a secretary.
Curation by some entity is going to involve being manipulated, which is very dangerous.
I won't think too deeply if your solution works theoretically, but in reality in our imperfect world most people would just choose to outsource the job to the point that we're back to the email problem. Majority of people liked Gmail - majority chosed Gmail and spammers galore we have. Even Mastodon is not resistant to centralization despite their efforts to be decentralized - most wouldn't bother to set up their server and so a lot ends up signing up to mastodon.social.
I quip, but I think there are solutions in that kind of thinking. Not everything needs to connect to everything.
Furyhermore, it's not like people can't post illegal content even if they can't spam. In a way, your readoning regarding it is like responding to a threat as if the person in question can't realize it even if his conditions are met.
Gee this sounds like telephone billing...
There are many other problems with this (not least of which being the amount of energy required now just to send a message).