I don't disagree with some of the points, but I unless you can point me to specific examples I can't see where that hashrate could go and still be profitable. Is there not a large amount of hardware that's no longer able to mine Eth because of memory restrictions already on those coins?
Profitability depends on the current price of the coin and the cost of energy. There used to be plenty of older Turing, Polaris, etc. cards on the used market, but even if they're not ideal mining cards, eventually the price increases made miners scoop up all of those cards. I don't mine, but I understand that ETH requires slightly more than 4 GB which leaves a lot of cards out, but at some point even older cards with that much memory might not be efficient to use due to being built on an older process. I don't know if other currencies face the same memory restrictions as ETH does.
Eth is the 900 lbs gorilla of the alt coin space, and even now it's the most profitable going. It's market cap is also orders of magnitude more than other minable coins.
I checked one website (
https://coinmarketcap.com/) that seems to have information related to currencies. ETH is #2 and is collectively worth as much as #'s 3 - 10 combined, but it's by no means the only cryptocurrency. The gap between the next currency and ETH is approximately the same ratio as the gap between ETH and Bitcoin. All of those top ten currencies have a market cap as large as or larger than many companies in the Fortune 500. There are plenty of companies that aren't at the level of Apple or Amazon, but it would be foolhardy to dismiss them as unsuccessful.
Yeah someone could develop RT coin, but who's going pay fiat money to purchase it? Hobbyists will mine and HODL random useless coins to speculate and for funsies, but unless people are willing to inject millions into your ecosystem you can't pay off your cards or power bill.
Even Dogecoin, which to my understanding was started as a bit of a joke, has a valuation of $7 billion. The Brave web browser created a cryptocurrency (Basic Authentication Token) as a means of handling advertising in a less intrusive way. It currently has a market cap of almost $1.2 billion (edit: I had an incorrect number here originally, but it wasn't too far off the correct value) according to the website I previously referenced. All you need is someone to accept the coins as payment for something and they're valuable to anyone who wishes to purchase that thing. From there they can be exchanged for other currencies.
I've thought that most of these cryptocurrencies would be far more useful if they did something other than pointless hash calculations. Obviously a PoW-based scheme requires a problem that's hard to compute, but relatively easy to verify the solution to, but there's no end to real-world problems that require a lot of heavy compute that are probably better candidates.
I merely used ray tracing as an example that's easy to grasp because it's quite clearly a gaming related function, but I'm sure there are other workloads that would be similar enough that could cause false positives. Hell, you could even create some kind of digital currency centered around providing a distributed Stadia-like service where someone is effectively paid to use their GPU to render games for someone else as opposed to using it to find solutions to a hash. Google hasn't killed Stadia (yet) so presumably there are people who would buy some digital currency that pays for the service.
Facebook
Twitter