I agree with your assessment here. The crux this entire theory is built on is the auditing. People will find a way to defraud it, especially if it relies on physical evidence of carbonization. It would have to be highly regulated to deal with your conspiracy scenario. But the fact that it's decentralized makes that incredibly difficult. It's definitely something that has to be considered as a possibility before being practically applied.

> my question is why is blockchain necessary if an auditor is involved? Isn't blockchain supposed to be zero-trust. Why even use a blockchain at all?

A common misconception. Blockchains are zero trust within the state/data space of the blockchain itself. We can refer to these space domains as "on-chain" and "off-chain".

Moving data or state information about the real world, say about the verification of successful carbon capture, to a blockchain is a core part of the oracle problem - which is the problem of migrating off-chain data, to the zero trust on-chain environment.

You use blockchains for the properties of immutability and transparency. In the case of gwern's proposal I can see a market environment where you can claim to capture carbon, provide data, make money (i.e. earn tokens) and then open yourself up to being "attacked" or "slashed". If someone else can make money/tokens by proving you are a liar you create an environment encouraging oracles (data providers) to be honest and accurate.

Final note: gwern pretty much states all of the above in his post, so I would suggest you not skim it.