,
11 tweets,
3 min read
Read on Twitter
I love @rusty_twit, but I'm (still?) seriously unconvinced by his arguments about long term security concerns. Thread:
1) Let's start w/ timing. It's noteworthy that block subsidy is NOT necessarily going to reduce in purchasing power terms! The halving process influences nominal values in sats, not in actual purchasing power. Block subsidy may even very well INCREASE in purchasing power terms!
2) Actually, in most (not all) realistic scenarios assuming Bitcoin adoption/success, the growth rate in demand could follow curves that actually surpass supply decreasing rate. Of course, this cannot go on forever: eventually, the adoption curve will likely reach a plateau.
3) But timing is important: the moment in which demand growth will slow down (possibly as low as global economic growth), or halving will actually reach the final cut-off point (block subsidy is not coded to be infinitely divisible) could be quite far in the future.
4) Such very long time-frames, while can't fundamentally change the terms of the problem, can actually challenge our quantitative assumptions about it. A lot. Let's consider, for example, a very conservative hypothesis about future block capacity increases.
5) Let's assume, for example, that we can hope for a 10x block capacity increase, in terms of increases of max block weight, or even of average block rate (I don't know why people only talk about the former instead of the latter, since trade-offs are actually quite similar).
6) We know that there may be even backward-compatible ways to achieve those kinds of increases. Fine. We would need that to happen before the demand for block space reaches its absolute plateau. Which is arguably a lot of time, in most realistic adoption scenarios.
7) So much time that, actually, it could even make sense right now to do the exact opposite, directing effort into short-term capacity DECREASES, in order to reduce validation costs & to help increasing the rate of validating nodes, which is scarily low.
8) Such a basic, quite conservative, assumption would cut the average L1 tx cost estimated by Rusty, with the same constraints, to just 70$. Imho, not so bad for a global, eternally immutable settlement layer. Assuming most fast/small tx would move on upper layers anyway.
9) But even said constraints are debatable, frankly. "$700 Million USD to undo 100 blocks" (assuming present-day USD purchasing power, I guess) is quite high for a typical non-state-level attacker. And is probably too low for a typical state-level attacker anyway.
10) The point here is that in basically any scenario in which attacking Bitcoin makes sense for state-level entities, Bitcoin is so successful that we can assume demand for block space going up as much or even more than subsidy goes down. That is until very far demand plateaus.
