How can they take over PoW? You either do the work or you don't. There is certainly a level of work that would make replying to every note instantly untenable, without being overly costly (or even noticeable) to the average user
Discussion
ASICs lol
something that casual users are not going to have at their disposal
they can then spew out valid proofs at a far higher rate to make their garbage pass PoW filtering
It adds a cost to an activity that currently has none... PoW is the best onboarding solution to WoT that I've seen.
I don’t know how spammers would benefit from doing PoW. If the goal of these bots is to attempt to steal zap, it must be an extremely small amount of income, which would be wiped out by the cost of creating volumes of valid proofs.
can you mine bitcoin with a PC?
what is the level of difference between them and what is the cost per hash, and time per hash, with current ASIC hardware?
what makes you think that the same arms race won't play out on nostr PoW antispam?
any conservative estimate is going to be wrong because it's probably based on the current size of the userbase and not on a future situation where half the internet has been driven off the mainstream social networks, that's hundreds of millions of users, and that's definitely a target for scammers
I’m just not seeing an economic benefit to scammers doing this here if it costs them even a little bit.
you are bearish on nostr then
and i guarantee that whatever they pay will be less than the average user by orders of magnitude, and if you end up having to pay for PoW then why not just skip wasting energy on it
Any way to look at it imo, spammers drive the cost of using the network up. I suppose WoT is the only one that doesn't here?
yep, it is the lowest cost spam limiter
any spam limiting with a higher cost than calculating social graphs might as well just charge for relay access, you see?
even simpler filtering for paid relays can literally be, only accepting events signed by paid subscribers, or you can make a second level with them and their follows, or you can go further and compute those graphs and update them every time a follow event shows up that was let through the filter, and the administrator can decide on an arbitrary accept threshold to apply, according to their stomach for fluff - but it might also be preferable for the users if they like to get off the beaten path a bit more
these are all pretty cheap computationally, for the most part up to more than 4 billion users a 64 bit fingerprint of npubs is conflict resistant enough so it's really a small amount of memory being burned too
Know the Knuth: “Premature optimization is the root of all evil” 😉
scrypt solves this. ASICs can grind a lot of compute, but memory doesn't scale the same way
lol! that's what CHARLIE LEE said 13 years ago lol, see how that turned out?
one of the last standing PoW coins but only because it's got such a big miner useerbase
i still haven't heard a reasonable, cogent argument from any PoW antispammers on nostr explaining to me what happens when the amount of hashpower increases, they haven't even thought about the fact you need a consensus on the network hash rate in order to actually prevent spam!
i hope you all enjoy your arse backwards method of antispam as much as this rake-boarder:
I like ❤️