@sunmoonbuffalo @ForestNanoo @mira_hurley@DigiEconomist Then the PoW alg would need to be changed to something like RandomX, to prevent optimization with specialized hardware. Otherwise like he said people will always find a more efficient way to make more money.
@takitus@pkulak@kjr_r@nano So there really isn't any system of "spam detection", it only ties right of usage with the stake of the account in the network, regardless of it being a real user
@takitus@pkulak@kjr_r@nano Currently the spam can only affect wallets in the same tier they're in, and even then, the least used accounts have priority in the same tier. So there can be legit people that transact frequently, but without a high investment the spammer can only harm those with low balance
@takitus@pkulak@kjr_r@nano I agree that the spammer could make another attack with $millions spread in a meaningful number of wallets to get aligned with regular users behavior, same as filling btc blocks with high transaction fees - there is no way around against someone willing to spend enough resources
@takitus @pixles_nano @pkulak@kjr_r@nano Now with v22 it pretty much can. The only transactions being delayed right now are those with very extremely low value AND highly used accounts in a short time frame.
@TransactionFees Not sure which node it's getting the stats from, but the finality time hasn't been more than 0.3s. Maybe should take the median value from multiple nodes.
@PlasticRaptor @nano This essentially moves the staking aspect that we have in consensus, to spam prevention, that's why it just works.
The "rich being prioritized" exists today when someone can buy orders faster PoW and keep the network saturated, and not be invested in Nano as for 51% attacks.
@PlasticRaptor @nano Notice how Nano is resistant against traditional Sybil attacks (because nodes with more weight is what matters for voting, more nodes can't control the network), but not against "sybil transactions", because we're giving the same weight to every account in the network.
@PlasticRaptor @nano Can you give 1 example? For years I've read tons of proposals on this matter, mostly modifications to PoW (memory-hard, work expiration against pre-calculated attacks, etc).
But none that are resistant to attackers with theoretically *infinite* PoW.