There's a (sort of) solution to this tail emission discussion, not mentioned in the article.
We will have to hard fork Bitcoin in the future (the year 2038 problem), don't panic, we have time until 2100-ish. With this hard fork, we might do more changes at once, like increase the block size, something we still don't need. I'm sure there will be other issues waiting to be sorted.
Anyway, what we can also do, is to go smaller units than satoshis. We already do use millisats (three decimal spaces) on the LN, we might go much further.
Now, the last halving in 2140-ish will dictate the block reward at 1 satoshi. What if it won't be the last halving but we will have a following cycles, rewarding miners 0.5 sahoshi, the next 0.25 sat, etc?
This way, we won't need to implement the tail emission and the 21 million coin will still be intact.
Note: there's this limiting factor, the max value of a 64 bit integer (18,446,744,073,709,551,615) but I'm sure we'll find a way around it too.
reply
When the block subsidy is 1 sat, the block reward will be 99.99999% fees which hopefully be orders of magnitude higher than sats.
Allowing sub-sat block subsidies does not make sense when they will be eclipsed by fees.
Either the fees will be sufficient or they will not but allowing endless halvings is definitely not the solution.
reply
The havings can't be endless and I agree, the fees will be the main source of miners's income before we have such a fork.
reply
Yes, that's a good point you raise as technology improves the average node could probably handle a lot more and we can safely increase the block size without cutting people off from validating the network and those bigger blocks may result in more fees
Hmmm, interesting so halving forever into units we might not care about now, but care about in the future, or even just having block subsidies held until it hits 1 satoshi before payout
reply
I'm pretty sure the 2038 problem has already been corrected in bitcoin, and didnt require a hard fork because the blockchain format had already taken it into account. The issue was around the difficulty adjustment calculation and the unix timestamp is only used as part of a difference calculation therefore count-overflow is canceled out in the two-weeks of seconds measurement.
reply
Ill make a prediction.
If dollars are still around in a hundred years, bitcoin will be in the billions making the 1 sat block reward + fees in the millions.
If bitcoin is not in the millions or billions of USD, I believe most BTC maxis will mine at a loss to protect their "stake"
This is the right wah to do proof of stake by the way.
Too bad none of us will be around to see what happens though
reply
I wonder if in 100 years we would still have fiat or would we have let bitcoins deflationary forces take over in repricing everything and mining would be a circular economy, and how would that effect security, my brain can't get around like if energy is now cheap in bitcoin terms what does that do for security
reply
there are already coins with tail emissions, and i dont think it belongs in proof of work. block reward should only be used to distribute coins as fairly as possible, and not to be used to pay for security.
but this is maybe a narrow point of view.
reply
I am with you on this "narrow" point of view, I would not be for tail emissions, because I believe in the hard cap thesis that with supply inelasticity, things around it will reprice, what I do wonder is how expensive do on-chain transactions become, maybe it does price out most people, but that's why we need second layer solutions, that help them but don't take too much fees away from blocks mined
reply