I have an idea for un-ASIC-able mining but I need to clarify something first.
My understanding is that a miner will receive txs, run all the code, package the txs and the outputs of the code into a block, and start hashing the block. The key is the separability between running the code and hashing the block. So if I were part of a mining pool, we would only need one node to actually execute all the code, and pass the results along to the miners, who can then spend their time on the proof of work hashing.
Is this correct, or does each individual miner (in a pool) actually have to execute code?