While constant tail emission does in fact converge to 0 inflation over time (which bitcoin's halvings do as well mind you), tail emission does *not* solve the potential problem of mining rewards, it only delays it. A tail emission of 200,000 btc/year (~1% of the current supply) would be equivalent to halvings every ~50 years rather than every 4 years. Were we to implement this kind of thing right after the last non-"
destructive" halving, it would buy us 46 years of extra time. Nothing more, nothing less.
While its mildly interesting to know that tail emission converges to a stable point, while no inflation implies monetary deflation at the rate of loss, this feels very likely to be an insignificant problem. I think 1% loss rate per year is an absurdly high estimate these days, and the loss rate is likely to decrease as methods of storing bitcoin mature. Imagine bitcoin was worth $1 trillion (not so hard, since it was not too long ago), then try imagining people losing $10 billion of bitcoin every year. Highly unlikely IMO. A rate of loss of 0.01%/year might be more realistic for a near-future mature bitcoin. That's not going to be enough to make a significant difference even over 100s of years.
If we actually wanted to solve the potential problem of not-enough-fees to upkeep mining security, there are less temporary ways to solve that. For example, if fees end up not being able to support sufficient mining, we could add emission based on a constant fraction of fees in the block. For example, every block could emit new bitcoin amounting to 10% of the fees collected in that block. This would tie coinbase rewards to the real world (since the fee market is tied to the real economy) and ensure higher block revenue indefinitely - ie not just for another 50 years.
But its also worth saying that blockchain security (which mining revenue correlates with) does *not* need to increase indefinitely. There is some amount of security (and therefore some amount of mining revenue) that is sufficient, beyond which additional security is simply unnecessary, unwarranted, and wasteful (you wouldn't buy a $1000 safe to store $1000 of valuables). Do we, as the bitcoin community, have some good idea how much security we need? Do we have some idea how costly a 51% attack must be where we can be comfortable it will never happen? I'm curious to hear what people think about that. Because without having some kind of estimates of what "enough security" is, there's absolutely no way of evaluating whether or not its likely that bitcoin fees alone will be able to sustain enough security.