CO₂ Coin: Decentralized Carbon Capture Blockchains (2021)
gwern.net
gwern.net
Ignoring the perverse incentive problem...
... enhanced weathering only works under certain conditions. Randomly spreading some olivine around at a specific GPS location does not effectively sequester CO2 on a useful timescale, for example.
Olivine also needs water, and the reaction is very slow, so almost all olivine proposals end up dumping the material in the shallow ocean (or grinding it small enough, say <2um, so that it stays in the mix layer before it fully reacts with CO2). Grinding energy is not linear, so a bunch of 100um rock (easy to produce) requires a lot less energy than a bunch of 2um rock. Vespa gets around this (they claim) by sticking the rock in the wave/tidal zone to break it apart naturally, and keep it from sinking out of the mix layer.
There are also issues with this from a biological standpoint (e.g. olivine contains some amount of nickel and cadmium), so doing this a gigaton scale would require a lot of research before incentivizing it.
TLDR, there are things one could do on land that might work, but not what is proposed.
===
Long story short, I can't see a carbon coin being super successful if it is overly prescriptive on the method of CO2 capture, unless it gets very very lucky and happens to pick the right one.
Since the proposal already requires an auditing methodology (which is basically all the proposal really is, if I am reading it right), it may as well drop the exact technology being audited. At that point, it mimics a normal carbon market, but with printed money, so not clear how it's going to get traction over a fiat currency.
I'm curious what these might be.
Everything else in your post was great, but there's certainly a precedent for even obviously unworkable ideas attracting much investment in the crypto space. Having something that could be beneficial is already way better than many crypto launches.
Unlike mining traditional crypto currencies, a carbon coin would require a large amount of capital equipment up front before mining. Bootstrapping without fiat capital would be an extra wrinkle.
BECCS isn't really scalable past a certain point, either, which leaves direct air capture and enhanced weathering.
Enhanced weathering is basically moving around alkalinity, which the CO2+water (carbonates) will neutralize. There are other ways of doing this besides dumping olivine rock into the sea (which gives you a very large surface area to absorb co2), though dumping rock is very inexpensive (probably less than $20/ton, modulo grind size and bio toxicity issues).
If humans disappeared today, the CO2 would revert back to pre-industrial levels predominantly through the ocean. The ocean is a large sponge for CO2, and it will stay in equilibrium with the atmosphere. As it turns over (~5000 years), it brings the extra CO2 down to depth, where it reacts with sediment. Extra CO2 will initially raise the carbonate compensation depth, reacting slopes of marine snow with the extra CO2 (kind of like snow melting) fairly quickly (on a geologic timescale). If/when that is exhausted, the ocean floor (which in places is covered with basalt and other material that will neutralize CO2) will (very very) slowly eat the rest of the CO2. As the carbonates drop in the ocean, the ocean will start to absorb more CO2 from the atmosphere. The ppm will drop slowly, until it hits the biosphere equilibrium (marine cycle, etc) of ~280ppm.
One scalable way (personal disclaimer, one of my research areas) to do carbon capture is to just speed up this process. There is no reason CO2 has to react with basalt, it can just as easily react with NaOH or some other source of alkalinity, which can be minutes instead of millennia. In fact, some direct air capture proposals (long ago) were just waterfalls of fine droplets with NaOH embedded as a sorbent.
So, long story short, if you can create alkalinity, you can easily turn that into sequestered CO2. Literally throwing an antacid tablet into the ocean sequesters CO2. Depending on the method of creation (e.g. chloralkali process), you may also have to dispose of an equal amount of acid, but that can be injected underground (e.g. poured out onto a bunch of olivine). Doing it all on land is a little trickier, as the ocean has some nice surface area and mixing ... but a large pond with a vertical firehose of high PH solution may be effective, though it starts to bleed over into the definition of DAC at some point and eats more power.
Say I have 1 ton of carbon captured, I am paid 1 coin and I "prove" that I captured 1 ton. Secretly, the mechanism of my capture is temporary and I realize not only the 1 ton, but an additional ton as it costed me energy to make this scheme work. What are you gonna do?
Another, more likely example is that I say I have 1 ton captured, EnergyCorp has Bob audit me, I tell Bob, you approve of my fraudulent capture notice and I'll give you 10% of the profits. Bob agrees, I am paid my 1 coin. I give Bob 0.1 of the coin. We both win. Later Bob is found to have participated in the conspiracy. What are you gonna do?
You, the smart reader are reading this and think - OK fine, we'll build a way to confirm capture without anyone else being involved. Well, I point to scenario 1. What're ya gonna do? The post basically assumes trusted auditors, which is fine, but why even do blockchain?
Blockchain cannot beat reality.
For what it's worth, both scenarios apply to traditional fiat, but at least in that scenario punitive damages apply and can be applied. I'm not sure how a tax or fine can or would be applied with any blockchain. It's also not even clear to me why blockchains are even necessary after skimming the OP.
How is accelerating exothermic mineralization reactions 'temporary', exactly?
For a concrete example - let's say planting a 50ft tree to maturity constitutes capture. In other words, if I point to a lat and long and show that there wasn't a tree here before, and now there is, I receive the coin.
Well, if you spent energy moving a mature tree from another place, you've not only killed the existing tree, but eventually the tree you've replanted will die too.
That's the sort of scenario I'm referring to.
In any case my concern is more around the second scenario I posed in my first post than the first one.
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.
It’s possible to write smart contracts in a way that an auditor, or group of auditors who reach consensus, can forcibly remove tokens and freeze accounts, of course this moves the needle towards centralisation more, (and I’m not interested in arguing the merits of that) - Im merely pointing out that immutability is not absolute in the way you are describing and exceptional circumstances and admin overrides can be coded in to EVM based systems at least.
If you are trusting a group of auditors why do you even need blockchain? The method in the OP is interesting, but I don't see the relevance of blockchain at all. Basically it comes down to some magical, incorruptible or comprisable auditor.
That's fine enough, but if you have that, you don't need blockchain to begin with.
But to your point that it's possible to hard fork, well, of course, computers being turing complete you can always go one level deeper into the platform and change the underlying system, but on a practical level forcing all of the miners to upgrade their software and coordinating this is a HUGE effort, and it seems like a very bad architectural decision to force an entire platform upgrade to deal with fraudulent users when safety can be built into the smart contracts themselves.
But I'm not arguing this centralisation is a good idea either, I'm merely pointing out that immutability isn't as final as the original post would make it seem, and there are ways of dealing with the problems they outlined (someone having fraudulently acquired tokens)
Is it that if everything goes perfectly with smart contracts being written perfectly with no bugs and all contingencies being taken into account and participants of the contract agree to the terms of said smart contracts in the event of fraud that's identified and confirmed it can be reversed? OK, point understood.
Bonding. If there’s a technique which allows for fraud 1 year down the road, then I need to have that much collateral locked in a trust for 1 year. If I behave, I get the collateral back, else I lose it. This discourages me from participating in detectable fraud over that year term.
1. "as long as the auditor is trustworthy only to the extent that they will not sign lies about dust not being present"
If they get the staked eth bounty, isn't there a pretty direct incentive for auditors to falsely report dust plots?
2. What happens with disputes? How do you challenge a false report about one of your dust plots? What if both sides are convinced they're correct & being honest -- maybe the dust farmer entered the wrong GPS coords, or the auditor visited the wrong coords, or wrote it down wrong in their notebook. What mechanism is there for adjudicating that kind of conflict?
This seems highly optimistic: if there’s money involved, cheating is inevitable. Since this inherits Bitcoin’s protections for successful fraudsters, it seems highly unlikely that people wouldn’t just try to dodge audits until the challenge window closes and their transactions can’t be reversed.
Same solution in PoS slashing for stakers who don't maintain required latency: too bad, we don't care why, it was your job to do it right, your stake is slashed. (See the NashX exploding equilibrium for another example of this: the buyer doesn't give a s—t whatever excuses the seller might come up with, because it's cheap talk; the seller makes things right, or their stake gets burned. Period.)
Flood with volume so there isn’t audit capacity & the time limit expires without a challenge.
Bribe the auditor.
The auditor turns out to be you under an alias.
Setup a mix of legit captured CO2 and something cheaper so when the auditor checks the easiest part to reach it appears valid.
Find something which will trigger the cheapest tests but costs less. Pocket the difference.
Truck the sequestered material from site to site, spending the proceeds long before anyone catches on.
Rely on the fact that since the auditing appears to be completely optional, most people will choose not to pay for it.
Bet on auditors being selected for being as cheap as possible and find the ones who will pass anything plausible.
Flood the system with fake challenges until people demand something like a quorum vote which increases the odds that someone won’t scrutinize your business too closely, or you and your pals try to convince everyone to adopt a more secure fork.
The key thing to remember is that a system which tries to be money is effectively its own bug bounty and it’s much harder to fix problems. Once you have money at stake every angle of the system will be under attack, and the security will come down to the real-world institutions backing it. Blockchains prevent silently rewriting history but that’s only one of the potential things people will try if there’s money at stake.
-collecting fake "dust" that's the byproduct of another industrial process and minting tokens off it
-bribing or threatening violence against auditors
You can add as many epicycles and “auditors” as you like and this won’t change the core problem.
The Alpha Centauri example is fair, and that does require a quasi-social process (everyone deciding to ignore the alien chain) to resolve. I say quasi-social because it doesn’t require any coordination or negotiation, but also isn’t fully automatic.
You have a blockchain, that much you can be sure of. If you want the blockchain used on Earth in 2021, you won't need to trust any specific person. If you believe everyone on Earth is lying to you, you are still hopeless and helpless.
If you could setup an audit system of the type specified, how does cryptography & blockchain improve upon things? The author seems to acknowledge that question earlier in the article but at the end when it comes saying how this is different, I didn't see a complete answer.
There probably needs to be something here about overlapping rectangles.
This portion needs to be rewritten for clarity.
1) There's no need to capture CO2
2) There's no need to record that on a blockchain or use crypto - you could use Github or Gitlab or IPFS with PGP-signed notary statements. Or push commits to all three.
3) Most enterprises are already forced to capture (or pay for outsourced capture, buy CO2 credits and other nonsense), so in order to make CO2 Coin viable in terms of (voluntary) market demand, you'd have to stop the government climate racket first. Good luck getting your friendly government department to give up on the millions of dollars they collect.
4) Considering how expensive global warming racket has become, I can't believe enterprises or individuals can be incentivized to use such a coin. It's much better to just ignore the schema - once the government is out of it (necessary condition for a CO2 Coin), damage from not participating in the racket would be negligible. Nobody wants to pay for it - if they did, we wouldn't have CO2 taxes and excises and no meaningful voluntary contributions.