You’ve invested $x in a certain number of hard drives and have merrily filled them with plots
You are actively farming these plots, hoping for a lucky chia coin drop
That’s all well and good, then you decide you want out of the Chia biz and wish to sell your farm. Is this possible?
Because as others have pointed out:
the original owner knows the private key to those plots, so you must trust that the seller is not secretly maintaining a copy of the plots and farming them behind your back, thus double-farming
You can ask that the seller destroy the private key…
… but you can’t prove that this has happened. I guess there needs to be a way to force a plot to belong to someone else, a command you can invoke that changes the private key? I’m not sure. I thought we should have a topic about whether it’s possible to sell a farm or not.
a command you can invoke that changes the private key?
I asked a related question, in relation to the language people use when talking about hpool, a while back:
The consensus is that, at least right now, there’s no “re-keying” of plots. I suspect that it may be be technically possible, based on my (limited!) understanding of what plots are, and how public-key crypto works. Though I’m not sure it could be done without sharing secrets.
But even if it is possible, I’d assume the data are the same (or maybe more precisely “equivalent”) in both the original and (theoretical) re-keyed plot – meaning it doesn’t get you any protection against double-farming.
So this is a really interesting topic. First off, full disclosure. I have a vested interest in this subject, given I am currently offering a farm for sale (and believe to be the first to do so, at least on this forum).
If we ignore the technical aspects for a moment, the answer to the question “is it possible to sell your Chia farm” is, of course, yes, providing someone is willing to buy it. Whether or not an individual chooses to buy is entirely a function of their desire to invest after taking into account their risk appetite. By way of demonstration, would anyone here be interested in a 1000TB farm for $5? My point is, at a certain price point (which will be different for every person), trust ceases to become an issue.
From a technical perspective, as far as I understand (and it would very much welcome anyone more enlightened than myself to correct me on this), the plots are inextricably linked to a private key, and that private key cannot be changed, nor can the plots be reassigned.
So where does that leave us? I think we can reasonably assume that there will be no technical solution in the short-medium term. Dev team priorities will (naturally) be focused elsewhere, and there is little literature on the topic for others to be able to move the topic forward. Are there some clever things you could do with chialisp to enforce contracts? Probably. But how much working chialisp code is out there in the wild?
All in all, I think right now we have more questions than answers. Thanks for kicking off the discussion. Will definitely be following with interest.
In these early days, if the transaction was large enough (say, $25K USD or more), as a buyer I think I’d need the counter-party to live in the same country as me so that we would create some kind of legally enforceable contract - especially around ensuring the seller deletes their private keys.
Of course getting a lawyer to draft the contract is the cheap part ($500 tops).
Using courts for actually enforcing the contract if it all goes sideways? Hopelessly expensive.
But sometimes just having the signed piece of paper feels better.
The fiction of Chia is that this investment is the time and drive-space holding the actual plots… but that’s just not reality.
The power requirement to maintain this drive space over time is negligible.
The REAL investment is the destruction of expensive NVME drives as these plots are created…
However, even the most expensive drives are still much cheaper than the electricity used in PoW…
A firecuda520 is $415 and can be expected to last for 3.5 Petabytes written… so well over a year in most Chia plotters…
I’m burning that kinda of money on PoW electricity in weeks…
As @codinghorror rightly states above, the value of prior SSD wear is indeed a part of the value of a farm, but arguably a small component.
The overall value of a cloud farm is based on three key factors:
Expertise to set up integrated VMs, plotting, storage and monitoring
Time and money saved from not using your own equipment
Foresight to start plotting X weeks ago (anyone have a time machine?)
In short, the value of a cloud farm, depending on your personal valuation of the above, and future projections of returns, is somewhere between zero and infinite. In much the same way as the value of Chia. We are all here because we are taking a bit of a punt.