Please no. Spammers have large amounts of more CPU available to do PoW. Consider a remote controlled botnet army.

Spammers have a massive advantage over ordinary honest users to do PoW. So ultimately you gain not much at the expense of draining the user's battery.

IMO there is opportunity in expanding social web of trust spam filtering. By following someone that means you trust their opinion. More of your friends reporting the same thing means you probably don't want to see it. Perhaps extend this social filtering model beyond per-post to affect entire accounts. I don't have a full plan here. This requires design thinking.

#[1]

Reply to this note

Please Login to reply.

Discussion

But I didn't mention spam control :)

Relays can enable PoW for any reason they see fit. If an Amethyst user wants to post into those relays, they will need this feature.

Why require PoW if not for rate limiting or spam control?

In any case this doesn't improve anything because attackers have lots of stolen CPU botnets at their disposal. This will only waste user battery and time without gaining anything.

Users aren’t meant to do proof-of-work… indeed, that is not a viable defense strategy like Adam Back originally intended due to botnets et al.

The ingenious innovation was when Satoshi applied Adam’s proof-of-work spam deflection to miners (servers)… rather than to users sending emails as Adam Back intended.

Satoshi created the first long-lasting user-server model that’s decentralized — problem is it’s very expensive to store data in blocks.

The whole point of nostr that makes it distinct from IPFS is users no longer need to spin up a server to participate — nor do they need to expend tons of computationally power or bandwidth… User costs on nostr should remain low like they are for a bitcoin user’s lightwallet via SPV.

Not for us to decide. We are just a client. If somebody creates a relay for the rich only, they can enable proof of work to make it happen.

They are free to make bad choices. I'm only pointing out the underlying logic is flawed. It punishes ordinary users without gaining anything.

#[5] ser, non-zaps or zaps can help fix the perceived problem of zero cost replies.

That doesn't work as a rate limiter because zaps can be costless if you pay yourself.

🤔

Pay per write event relay?

pow is basically a desktop users only filter

Legacy tech is already by and for the rich. They don't need nostr.

Pointless LARPing… no need to create unnecessary costs.

We want high security with a low barrier to entry.

#[7]

That means people will be locked in their echo chambers.

What sort of censorship-resistant medium is it if I can only see what “people I trust” want me to?

If spammers have more resources for attacking the network than legitimate users for maintaining it you have serious issues a web of trust isn’t going to solve.

Your argument could be used to defend changing Bitcoin to proof-of-stake, do you realize that?

Quite the contrary. Botnets are only a threat to CPU mined shitcoins.

ASIC's being single purpose hardware millions of times faster than a CPU means botnets using stolen electricity can't compete.

Why would PoW need to be CPU-mined in Nostr?

And BTW that means Bitcoin was a bad idea before the development of ASICs?

You can’t both have someone that can resist censorship and posting on it for free.

You either require some proof of payment or you delegate to someone else the decision of what you should see.

If I have to delegate it to someone else why not just choose a company that has a certain level of public scrutiny rather than some random people on the internet I find amusing?

If we go with some sort of payment, zaps to other people can be faked.

So it’s either zaps to each recipient or PoW.

PoW is unbalanced and ultimately an unworkable solution due to attackers having more resources than ordinary battery limited mobile users.

People will waste time implementing this. It will seem to work for a while. Then a malicious attacker will bypass it with ease at no cost to themselves.

PoW doesn’t need to be done in your phone.

You can pay someone else to do that, or have your own PoW rig at home do it for your phone.

I’m not talking about doing a billion hashes on a phone battery.

If that’s the proposal I agree it makes no sense.

That doesn’t mean abandoning PoW

OK you're right. If you're talking about renting time from the fastest ASIC that's actually an unfakeable cost. That would be a meaningful rate limiter!

Botnets are also an opportunity cost. A user playing on a social media app is not considering what else they'd do with their CPU work, at that moment. A botnet, of which there are really only a few, do have other uses and potential monetization. We don't need users to do more work than an botnet, just make it cost enough more that it's more profitable to do something else with it.

Setting to only allow posts and comments to be shown from those you follow and those they follow. I think Nostrgram has this option as filter for global feed.

Amethyst is already almost that. I like it.

devs: please consider that follow is not trust. There has to be a way to follow a list of people, yet still curate a separate list of those you recommend or trust. Make this easy for the average user and the platform will be unstoppable.