I have several points to criticise the "rare sats" concept. I don't repeat pooya87's and Ambatman's points but some are somewhat related.
My main point is that in an open system like Bitcoin, it's likely that the value of rare sats will decrease continuously because the sats will become increasingly less rare over time, due to certain incentives.
First because there will be a proliferation of "historical" satoshis. There are a lot of transactions, dates and events you could consider "historic" in the Bitcoin world.
I give a non-obvious example: What about if we consider the first satoshi of the person who bought at the bottom (i.e. to the lowest price) of a historic bear market "rare"? We had (depending on counting) between 4 and 6 deep bear markets, and those who bought at the bottom were the "heroes" who bought into a falling knife, creating the base for a new bull market.
Second, the method of considering a sat "rare" due to its position is also open for new approaches. We already have the example of "black" sats, the last ones in a difficulty or halving cycle. In another thread I have brought up the following examples:
1) A sat which is situated exactly in the center of all coins mined during a cycle.
2) As an extension to 1), a sat which is situated in a mined transaction in the exact center of a cycle, not in the mined coins. The interesting property of this approach is that there may be cycles without such a "center sat", which means that they are more rare than the traditional uncommon sats.
3) A sat which is exactly positioned at the "golden ratio" between the first and last block of halving cycles.
Third, what if some whale decides to create an own ordinals protocol using franky1's approach to count the fee first when sats are transacted, because he has bought (for almost no money) a lot of these sats? We could get two rival versions of the protocol, and while one will perhaps "win" it's possible that both are used.
All of these approaches allow to create a very high amount of rare sats over time, and that will make the average rare sat less rare. Incentives are always there: you identify an "interesting" event, mathematical property or whatever in secret, buy the sats in secret (e.g. on P2P exchanges) and then promote and establish them in the same way NFTs are promoted ...
That is a fair point, and some of those categories might even make sense. However, that doesn't dilute the value of sats from the first bitcoin transaction for example... in the same way that altcoins don't dilute the value of bitcoin.
About franky's point: he might be onto something, but there are arguments for the fees being the first to flow out or the last as well. I believe that's the
single arbitrary convention about rare sats.
Your point can apply in practice with the current indexing method tho: someone searching for some category of sats, then creating that category. In that case, it's up to the people to decide if there's really a value proposition with new rare sats categories or not. On
magisat.io we list only what makes sense. Also, one shouldn't just jump into any category of rare sats. Some make more sense and have real historical meaning, while some other categories that don't can hit the market in the future, but that doesn't make the ones that make sense weaker imo.