Introducing Proof of Activity: A New Consensus Model

I wanna throw out this idea for a new consensus model called Proof of Activity or PoA. Basically, it tackles some issues with existing PoS models where there's tons of unnecessary network chatter and bloated chains due to all those signatures. With PoA, only one stakeholder is chosen randomly to sign the block, which makes things a lot simpler.

That’s interesting but I think you gotta consider key management. Don’t use the same keys for transactions and mining. People need to keep their digital assets safe, like having some coins offline or encrypted. There are various methods to do this, like one-time delegations, so just make sure to plan it out.

Totally agree. And about the randomness aspect, sure there's merit to it but what if someone ends up signing a weaker fork? If their block wins, that's the chain we gotta deal with. But yeah, PoW is still needed to build on that chain, right?

Seems like you’re aiming this at preventing Byzantine attacks, but that could make small reorgs easier. Plus, in such a chaotic system, doing double-spends might get tricky.

I actually think probabilistic PoS is a cool concept. The urgency around the highlighted issues might be overstated, but it could offer unique benefits.

Not sure about that. Satoshis don’t really track like you think they do. The distribution of coins in a transaction isn’t straightforward, you can’t just say which satoshi went where. But yeah, if you had a randomness seed, you could trace it along the chain.

Hey, just chiming in, but with those delegations you mention, don’t they need a connection between public keys at a certain block? Otherwise, nodes would have to keep checking past blocks to ensure everything adds up.

Of course, satoshis can be tracked that way if you map each input to outputs in a clear manner. Like if you’re looking at a split transaction, you could follow it through. Makes sense, right?

Totally! But remember, the order of transactions matters a ton. If a miner forgets about a fee on a satoshi, that could complicate things like tracking.

Don't you think there's a standard way to sort transactions in blocks? Alphabetically is fine, as long as everyone agrees on it. But the risk for a miner messing up their coins during tracking needs to be flagged.

Can you clarify what attacks PoA is meant to guard against? 51% attacks don't really show the motive behind them, so I'm curious.

Yeah, 51% attacks might not be the only issue. Small block reorgs could slip through with this proposal. Coblee’s suggestions could make it even easier.

Double spending has to stay under wraps. If it’s known, then exchanges and networks will block those attempts. Random stakeholder selection complicates who to bribe for an attack.

Okay, here's my full thought on why PoA might not work and could even hurt us. Like for-profit double spending. It can turn into a whole business and I can outline methods to make it happen.

Could you elaborate on not needing a full block for a signature? In Coblee’s plan, signatures rely on block hashes. How does that work without the block being solved first?

Also, what about the economic model? Maybe taxing stakeholders instead of rewarding them could balance things? Like if they don’t sign, they lose some coins.

The attacker could just drop block hashes but not transactions, right? This could expose vulnerabilities in Cunicula’s protocol. If there are no penalties, then bad actors will only get bolder.

Not really buying that attack scenario. If the attacker can succeed, the currency's gonna tank fast. Why would major stakeholders help undermine their own interest?

Great point! Coblee’s idea kinda feels like a jury duty thing. If the designated signer doesn’t sign, it’s game over for that block. That might keep people more engaged.

True, but if this model introduces issues like double spending risks, it could severely impact the market. Stakeholders won’t instinctively avoid harm.