Open this lesson in your favourite AI. It'll walk you through the why, explain the demo, and quiz you on the try-it list.
A block has to be proposed, propagated, and accepted. The identity of 'who gets to propose' and 'who decides if it's valid' is the single most politically charged question in blockchain design, because it determines who earns block rewards and who can be bribed or coerced. PoW answers it with hashrate, PoS with stake, and PoA with a whitelist.
On Ethereum right now there are ~1M validators, each staking 32 ETH. Every 12-second slot, one validator is pseudo-randomly selected (via RANDAO) to propose a block. A committee of ~128 other validators attests whether the block looks valid. If >2/3 of the committee attests within a slot, the block is accepted; after ~13 minutes (2 epochs) it becomes finalized — meaning reversing it would require slashing ~$10B of staked ETH. On Bitcoin, the equivalent actors are mining pools — four of them (Foundry USA, AntPool, F2Pool, Binance Pool) currently produce ~75% of all blocks.
Use these three in order. Each builds on the one before.
Who proposes the next Ethereum block, and who proposes the next Bitcoin block? Name the actors and the selection mechanism in each case.
On Ethereum, walk me through exactly what happens in one 12-second slot: who is selected, what they publish, who attests, and how finality accumulates over epochs.
If one mining pool (say Foundry) controlled 51% of Bitcoin hashrate, what attacks could they realistically pull off and which couldn't they? Compare to an Ethereum scenario where Lido (staking pool with ~30% share) went rogue. Which chain is more vulnerable today, and to what?