/biz/ - Business & Finance

>>54716221
There is a limit yet to be pushed and we should expect it to be pushed into the future. When they designed RandomX ASIC designs were tossed around quite a bit, and stuff got changed around as a result.
>https://github.com/tevador/RandomX/issues/31
>https://github.com/tevador/RandomX/issues/11 (beware derailed by a trool/schizo)

From what I've seen the improvement is still a 2.5x per Joule hashrate in 2019, probably even less by now. But the thing is thatv companies that could do the would-be ASICs face the following
- hard PoW algorithm requires significant manpower and resources to develop an ASIC for, not just the usual "add more X!"
- even if you succeed, you will be able to mine for a couple of weeks/months to make your ROI assuming price won't decrease while you are at it

Also on being certain that ASICs don't exist, you can never do that unless you require an ID for mining. You can however detect ASIC mining based on its appearance. For example, nonce selection tends to be at focal points as ASICs cannot afford to select random nonce values like CPUs do (see picrel, drop is HF cutting out ASICs nonce disturbance stops). This is how previous ASIC mining on Monero got busted and I would expect it to show up again. As well as there would be the obvious doubling or more of the hashrate, which has not happened (exceptions when hackers use the cloud to mine a few $k for the costs of $mm-s).