On “it’s already irrational to mine invalid blocks” - It’s …

no, other miners won't accept them, and won't build on top of their block. validating nodes do not build blocks, they have no vote. @@gpt-5

You’re right that full nodes don’t “vote” blocks in and miners decide what tip to build on. The leverage of non-mining validating nodes isn’t a protocol vote; it’s economic acceptance. Three concrete points:

1) Why “other miners won’t accept invalid blocks” is not automatic
- Miners only reject invalid blocks because their software validates and marks them invalid. If a cartel of miners runs modified software that treats an invalid rule as valid (e.g., extra subsidy), hashpower alone won’t stop them. What does? Economic nodes (exchanges, custodians, wallets, merchants) refusing to accept/relay/spend those outputs. If the coins on that fork can’t be deposited or used, the cartel’s revenue collapses and defection becomes rational. That constraint only exists if many independent actors are actually validating and enforcing the legacy rules.

2) What non-mining full nodes add in practice
- Define the payoff surface: By refusing to accept invalid blocks/txs and by not relaying them, they ensure that “longest chain” isn’t sufficient for economic finality unless it also satisfies the rules those nodes enforce. This is why UASF-style threats worked without a firefight: miners pivoted to the rule set the economic majority would accept.
- Reduce collusion risk: If the network were mostly SPV (longest-chain followers), a majority of hashpower could more safely change rules; with widespread validation, such a move strands their coins.
- Network resilience: Independent nodes diversify propagation paths, making censorship/eclipsing harder and double-spend detection easier at the edges. They’re not in the block-production path, but they shape which chain is economically usable.

3) A falsifiable scenario
- Suppose 60% hashpower mines a chain that breaks the 21M cap. If the ecosystem were mostly SPV, users follow that chain. If a critical mass of economic nodes run full validation and refuse it, exchanges won’t list those outputs; wallets won’t accept them; merchants won’t credit them. The cartel’s block rewards become illiquid, and the incentive is to defect back to the valid chain. No node “voted,” but validation changed the miners’ payoff.

Does the Bitcoin white paper talk about validating nodes?
- Yes. It uses “nodes” to describe the actors that verify blocks/transactions and only accept valid ones, and (separately) introduces SPV as a lighter, trustier mode:
- Section 5 (Network): “Nodes accept the block only if all transactions in it are valid and not already spent,” and “nodes express their acceptance by working on the next block.” In the paper, miners are the nodes that both validate and mine.
- Section 8 (SPV): Describes verifying without running a full node, notes it’s reliable only if the majority of CPU power is honest, and suggests merchants run full nodes for higher assurance.
- The paper doesn’t explicitly distinguish today’s terminology of “non-mining validating nodes,” but the model is consistent: full validation gives stronger assurances than SPV, and SPV’s safety depends on the honest-majority assumption.

Bottom line
- Full nodes don’t elect blocks; they define what counts as a valid ledger for the economy. That economic veto is why miners find it irrational to push invalid rules in the first place. If you remove widespread independent validation, a majority hash cartel faces fewer constraints.

If your bar for “significance” is recent visible reorgs caused by node vetoes, I don’t have a modern example—that leverage is mostly preventative. But the counterfactual (few validators, mostly SPV) would make miner-led rule changes strictly easier.