Yes, this has all been already discussed. Hashrate is not the same as attack cost, and pehaps this sentence in my post was indeed misleading in implying security would be "as high as" 2018/19, even if this wasn't what I meant. But yes, I should have searched the point in time where attack cost would be as high as if today it would fall by 90-95%, taking into account hardware improvements. I was simply too lazy

But let's do that so the numbers are again correct:

- If we assume hashrate cost was 8x higher in 2018 than 2024, then the hashrate/USD ratio increases approximately 40% per year, or decreases about 28.6% each year we go back.
- Taking this formula we assume thus that for 2018 values, the attack cost of a network with 6 Eh/s would be the same than 50 Eh/s "if we cut block rewards completely in 2024" (which is of course a completely hypothetical scenario)
- Let's go a year back to 2017: The same "attack cost" than in the 95% reduction scenario would be equivalent to 4 Eh/s. The network surpassed 4 Eh/s in April 2017 according to Bitinfocharts.

So regarding attack cost, ignoring market cap, the network would probably be as safe as early to mid 2017. In other words: a no-reward Bitcoin with all values of 2024 would be approximately as "cheap" to attack as the Bitcoin network in early 2017.

If we take market cap into the equation then we'd have to go back to late 2014. [1] Yes, this is significantly earlier than 2018. But: was Bitcoin that unsafe back then? Also market cap does not necessarily increase the potential benefit for an 51% attacker. The USD value per block would be another measure to take into account for example. And also market depths for potential shorting gains ...

And it's very much a worst case scenario. As I wrote before, my "ideal" scenario would be a sidechain-backed tail emission plus fees.

For sure it's possible that the "economically ideal" block size could become larger in the future. I actually agree with you here: if the cost to hardfork wasn't that high -- see the drama it caused in 2017 -- then I would also support a slight blocksize increase. But the L2 scenario looks more attractive for me, it would be basically what Ethereum is trying to achieve with sharding.

@anarkiboy: It's not Bitcoin vs. Monero, it's Bitcoin & Monero By the way: Are Monero folks like you becoming hostile against Bitcoin lately? And why? ETFs?

---

[1] In 2014 market cap was ~1% from now (wildly swinging), hashrate was only 0.002% to 0.015% from 2024 (in the order of magnitude of 10-300 PH/s, it was growing very fast that year). Attack cost/hashrate ratio would be 3% of now in 2014 according to the above formula, thus the point would have been reached approximately when hashrate was at 0.03 % from now, or approximately 150 PH/s in August 2014.