"web3 social" will never be successful because censorship is a desirable feature of a social-media platform. If there is no way to censor a group of people, there is no way to censor spammers either. You will be able to filter at the client level of course, or set up maintained filter-lists (which will be trivial to game with a "sybil" approach), but this is going to be "running an email server" times a million in terms of how abusable that "censorship-resistance" is going to end up being. You will need "spam assassin for social media" immediately and it will be a continuous battle.
Spammers are perfectly happy to pay-per-impression if that's what it takes to get ads (or botnet droppers, etc) in front of victims. They are more willing to pay than an actual user is, really.
Require a stake to setup an account, say $10, and if a spammer is too spammy and too many people complain they lose the stake. Or maybe it's just paid out on a long delay. Now spamming costs quite a bit. Add to that a blacklist feature and a shared consensus on spamming addresses and it starts to work a little better.
> Require a stake to setup an account, say $10, and if a spammer is too spammy and too many people complain they lose the stake.
Oh, so, if I have $1m I can register a bunch of accounts and kick people I don't like off the platform? You said this thing was supposed to be censorship-resistant, right?
Hell people will do it for free if they disagree with someone, you don't need $1m of accounts, you just need to say something that angers 100k people who report you and get you banned... isn't that something that people specifically complain about with regards to "censorship" with current platforms? Like I'm not saying they have to read it, but if others have the ability to stop me from publishing than that's not really censorship-resistant, it's just mob-controlled censorship. Censorship-resistance is something like IPFS - you have no control over what I put on IPFS. You don't have to access it, but you can't stop me either.
Also, if I register enough accounts so that 100k people never complain about any one account do they gradually go back to being healthy accounts? Or if not, then when I spam a bunch of people then I can get them kicked off they will be banned permanently (or have to pay another $10 to restore access), right? How long will this service last when I'm kicking every single user off every single day?
"Consensus" will always take into account the consensus of attackers too... sure you can always have a trusted party who oversees censorship and you can depend on to make unbiased and fair decisions that 100% of individuals will agree with (lol) - but in that case, why do we need crypto at all? Why not just let that guy run the server/why not let him build blocklists for the mastodon server?
it's the SMTP copypasta except the thing you have chosen to ignore is sybil attacks, and the massively complex thing you are handwaving as being trivial is building a reputation system, and that would solve all the problems for the non-crypto space too.
(really, it's the exact same problem, it doesn't matter whether the format is facebook posts or emails. Email is actually an extremely decentralized and censorship-resistant protocol - there is no central authority who can stop me from putting my message into your mail server, or even see what I'm sending you (if we connect directly). We can have completely private conversations that no central authority can stop! And what's what makes it terrible, because it turns into utter white noise with vague "reputational" signals like "dirty IPs" - those are your "blacklisted commenters" on facebook, and spam emails are still a problem. Or at least they were, until a centralized authority (Google) effectively came to take over the market and solved the problem for us... but then there's censorship again. And if you are winding up about "PGP message signing" or anything else, you need to go back to the "spam solutions" copypasta and find the bits that apply to you, because people thought of that 40 years ago before crypto was a gleam in satoshi's eye. Is spam solved? No, because your idea doesn't work.)
Spammers have been willing to "pay per account" forever, and that's the problem, they are more willing to pay than the actual users are. So any model needs to survive in the face of an asymmetric foe that has more resources than you... and almost all of the "asymmetric" strategies will require either a huge amount of cooperation between "honest" participants, or a few "super participants" who have actual power (ie censorship) over the system.
And remember - the original idea behind bitcoin literally came from HashCash - which was an attempt to reduce spamming by requiring a unique proof-of-work for each message.
Edit: Just to clarify, in my response I think blacklist != censorship, if a blacklist can be altered by anyone on their own system.
You seem to have interpreted what I said slightly deeper than what I meant. I didn't mean to suggest there could be some widespread censorship. I imagine you could stake to create an account and then you could vote someone as a spammer - voting causes you and the person you voted to have to wait N more time before withdrawing (harms both accounts). If many people vote V for someone as a spammer that account would have V*N delay. I'm not saying they're censored, I'm saying they just have to wait longer before withdrawing to make a new account.
Add to this a simple blacklist system like how some antivirus tools collaborate, sharing hashes. You can pick your own third-party list of spam hashes based on different rules, or modify/build your own. Using your example of an account getting 100k spam reports on-chain; auto-add them to such a blacklist. Realise your favourite person has been spam-attacked in the way you describe? Whitelist that account. Like how you allow some apps you know are safe to run on your computer.
This is super rudimentary and pretty quickly thought of, but I feel like you could get halfway and that's all I'd really care for. No-one here is censored btw, as in like how Trump was kicked off Twitter, they're simply blocked by default for you based on an algorithm you choose and can be whitelisted again if you want.
You could add smarter methods like checking strings for common spammy terms, such as all those fake Elon accounts on Twitter saying the same things, and auto add them to a list. You could have a quadratic spam voting system where if someone is voted repeatedly the wait they have to withdraw gets longer and longer. You could have a reputation system based on the number of people who say someone is spamming where others have too, but who has never been called spam themselves, who then gets more weight for future votes to blacklist someone.
You could even create psudo-private groups on the system by adding your own accounts to a primary blacklist and then whitelisting each other. Obviously not private because someone could just not use a list, but you'd only see each other to make things easier, like a subset of Twitter.
I don't know, there's loads of small tweaks you can make here, but it feels like you could get a P2P application working better than Twitter in terms of spam if you have the ability to pick your spam prevention model and accounts cost a deposit to setup. Maybe you've worked on this way longer and have thought about these all before and they're provably wrong! FWIW I quit working in crypto earlier this year and am far less invested than others, I'm just also less dismissive of it off the bat.
I think the main reason it won't be successful is because people simply don't care about decentralization/privacy. Historically, there have only been a handful of successful social media companies. Starting a new one with the key differentiator being decentralization, while also adding the onboarding friction of crypto wallet management, is essentially a guarantee of failure. Mastodon tried this, without the unnecessary wallet friction, and is still an extremely niche platform.
Same reason crypto won’t take over for payments (even if it could scale).
We already have payment networks that work well. People aren’t going to switch to a less convenient solution to what they already have for <insert political reason about decentralisation/privacy/anti-statism>.
> We already have payment networks that work well.
I assume you don't live in a country built around M-Pesa or EVC Plus, but it's very hard to build banking records around mobile credit. For example, in Somalia which uses the USD, has US, UK, and many other countries financial sanctions. So you trade airtime in the form of USD, but without a real banking infrastructure behind it. There is an increase in moving from this to a crypto-based system over SMS that allows for unique accounts, credit histories, international payments for remittance and micropayments (these SMS tools don't work out of the country), etc... So here Crypto might be the better payment network. Here's a guide for EVC Plus to understand how they trade there:
Spammers are perfectly happy to pay-per-impression if that's what it takes to get ads (or botnet droppers, etc) in front of victims. They are more willing to pay than an actual user is, really.