Nick Szabo  🔑

why is hash power supposed to decrease at long run?

It may if tx fees don't pick up slack from decreasing block rewards. Has been said to present a security risk, but I haven't actually read a good argument showing it would degrade security. Likely reflects oversimplified mental models of how hashrate contributes to security.

is the tx/s rate increasing being considered in this analysis? Millions of txs, millions in fees being collected. I been always certain hash power will grow as far as adoption can.

I wish more people knew about the "just wait a few more blocks" response to the hash power attack scenario.

Covered that later in the thread!https://twitter.com/mhluongo/status/1092581486392475648?s=19 …

Mentioning != covering. Several ways for recipients of high value to protect themselves, including (i) break tx into multiple txs across multiple blocks (which BTW will also increase miner's fee revenue), (ii) wait more blocks before relying on.

Hash power is very unlikely to decrease because the cost of a single hash falls by ~roughly 50% every year. Relative hash power is the metric you're looking for.

Does it still? I thought Moore's law has been slowing down lately.

It has, and this is going to be closer to 0% than 50% in the future, although still somewhat unpredictable.

Unless there's a massive archecture breakthrough it should decrease over time.

If I was being more precise in the original tweet and had enough room I would say "value of capital investment in hash power." But it's fine if absolute hash power decreases too, as long as it's a gradual and at least roughly industry-wide trend.

Moore's law is about the density of transistors. I'm talking about the price/performance ratio of generic information technology, which has fallen by 50% every year for the last 75 years.

In other words, how many calculations can you get for a dollar in constant dollars. You can get twice as many every year in inflation adjusted dollars.

It's become twice as many every year because chips have reliably doubled in density every year until recently. Are you saying that price per calculations has gone down, not as a factor of increasing density and consistent pushing of the upper limit, but for some other factor ?

It depends on how low the hashrate can go. Bitcoin hashpower at a low single digit % of total SHA256 hashing would be an utter catastrophe. Trades would require days worth of confirms to go through with safely.

Competitors can't fund that much hash power from inflation unless the demand for them is much greater than the demand for Bitcoin, in which case Bitcoin is hosed anyway for independent reasons;.