Well I agree with you about churn, but IMO it has largely happened because the recruiting is all wrong. When you recruit people to come and make money (from a giveaway), when the money is small and declines (largely due to the market price), then people quit.

When the market price was higher or increasing, churn was much lower. Users stayed and the overall user base would grow.

When the money increases, even if it is small, people see potential. But you can't guarantee a rising price.

The idea of the whole value proposition resting on people getting free money is broken. The promise is powerful as an attractant, but the reality is always a disappointment, leading to churn. It needs to be drastically deemphasized.

BTW: Everything you say are assertions without prove. It's something you believe in and you present it as it's axiomatic truth. But it's not. Hitchens's razor applies.

You can say it is without proof, but it is entirely consistent with observation. That can't be said to be literal proof (as little about the real world ever can be), but it is strong evidence.

When we introduced linear reward with minimal penalty at the low end (only the dust threshold) we saw exactly this phenomenon, with huge growth of scammers creating massive numbers of comments with small votes to extract a large amount of rewards, a little bit at a time. It didn't exist under n^2 because these sorts of small payouts weren't possible.

When we introduced the convergent linear, the swarm of small milking comments went away as a strategy, and indeed much of the decline in overall count of comments that happened rapidly was due to these milking comments going away. It is hardly an extraordinary claim that not only moving back in the direction of linear but even farther to sublinear (which would be even more vulnerable), would result in the same exploitative behavior. The math also makes it clear this would have a larger payoff than with a linear curve, which both increases the incentive to engage in it, and increases the portion of the rewards extracted for the same effort at exploiting.

The premise that there will be people who exploit vulnerabilities in the system seems well-supported, so when you point out an vulnerability, an appropriate response hardly seems to be to claim it doesn't matter or can be swept under the rug because it isn't literally "proven" that it will occur. The belief that it would is reasonable and well-supported.