True. That's why we probably need another mechanism for incentivizing high-risk/high-reward/multi-decennial research. Perhaps, even if LK99 is not real, something interesting will mutate out of it: some kind of platform merging arXiv, Twitch, Patreon, making the researchers to be more transparent, more willing to share partial results, failures, and even be rewarded to fail. Some time ago had this "fail database" in mind, where one would upload all the data of the experiments that didn't work. I see there is a "FailCon" [1].
[1] http://thefailcon.com/about.html