Hacker News new | ask | show | jobs
by vkou 1242 days ago
They didn't solve it by 90%, they lied with statistics, and compared exponential-growth forecasted counts with point-in time counts.

The reduction was closer to 60%, and in the past few years, they've stopped investing in the program. It's been a good thing, but without ongoing effort and investment it will regress back to the mean.