Regarding the first year of Bitcoin I sorta agree, these "extremely old" sats may be really relatively rare. But recently one of the old miners from 2009 has moved coins. That's 5 billion new potential rare sats, "common" ones but old ones, which can dilute other old common sats, e.g. the Pizza sats. From time to time this will occur, although not very frequently (we simply don't know what Satoshi is doing with his coins).

But for all coins after 2010 the number of "lost" Bitcoins and sats is much, much lower than the lost units of "analog" collectibles like stamps or coins. Even if the 30% some sources claim that are lost forever (these are those not moved for several years) this means 70% of the coins, and thus also of the rare sats are accessible. Probably the number is much higher, as other estimations about "lost" coins range between 10 and 20%. In addition, since the "rare sats" concept is known, most of those who could hold a rare sat are probably knowing it and "protecting" their sats from getting lost.

And as rare sats don't deteriorate, there's no indication that the structural problem that "rare sats" do not become more rare over time if demand is static, could change. The only chance would be an extreme crash and crypto winter where many people would "give up" their coins.

In reality I'm not against rare sats, I think it's a funny concept. Perhaps I'd even try to get[1] a sat e.g. from the Pizza transaction, which are cheap. Just for fun. But the concept has too many problems to be really a meaningful "investment".

---

[1] "getting" does not necessarily mean "buy", it could imply to search for them via an alternative method, like franky's Would be a fun exercise of blockchain analytics.