Finding anomalies that develop into a mean reversion trading algorithm is actually one of the more fulfilling things I've done. Only a few other things in my life have matched the amount of grit and brainpower I've had to use to get something like this accomplished. A team or company stretching the limits of statistics and computer science to do the same seems like something worthwhile to me.
Now we can get into some areas that may be borderline unethical, like certain types of front running, but a blanketed statement isn't a fair thing to cast on the industry as a whole.
I’d argue this cross exchange arbitrage does still provide some value by keeping prices of securities across exchanges/the world in sync, despite being quite unfair and taking value from those putting in large orders.
Liquidity provided by algo market makers is also a service to market participants because they take risk to ensure there is always someone to buy or sell - this reduces volatility and risk for everyone.
Algo trading is also required for keeping ETFs in line their benchmarks, which is an entirely separate subject you could fill a book with.
So no, all algo trading is not the same thing, there are valid and productive uses of code rather than people shouting across a pit or running slips up and down roads to keep capital flowing through markets efficiently.
