If we set the multi-sig threshold at 17 then it only takes 4 accounts to hold the funds hostage.
If those 4 accounts act against the wishes of the stakeholder majority, they will be immediately voted out.
What bothers me most about the state machine you are proposing, which differs from a blockchain, is that it doesn't seem to have all the robust security of a blockchain. For sure, it uses less computer resources, because it doesn't have all the cryptographic chain that locks in each transaction and prevents attacks. With a state machine, who knows what attacks will be possible. It's a long road to create a foundational technology that does what a blockchain does - it has to be mathematically worked out, carefully reviewed by many eyes, battle-tested over years, before relying on that technology to handle large amounts of funds.
I deeply care about the SPK network and really look forward to it. The technology it is built on has paramount importance for the success of the whole project. Do we want to let our SPK network rely on a new and unproven L2 technology?
If the threshold was 17/20... nope, 16 couldn't vote out 4 ever.
I guess you don't know that this "state machine" has a consensus mechanism, is older than SushiSwap(with pieces older than HiveEngine), and has had over 100 operators. As a layer 2 it doesn't need to order transactions, but it still has other aspects of a blockchain. Forgetting all of the the solution must live on Hive because the DHF is paying us to build it to make hive a better place.
I mean that all the people who stake the governance token can vote out any rogue validators.
I didn't mean to devalue what you have created, sorry if it sounded like that. If it works with a sufficient level of robustness against the various kinds of attacks, and does what is needed for the SPK network, then that would be a terrific innovation you have created.
Agreed, and that's the general idea... have a few layers to the governance structure to mitigate the limitations of multi-signature accounts.