The thing that’s always missing in this kind of reporting is: with what budget and what level of uncertainty?

You could take any existing product, which was originally developed in a fog of uncertainty but now serves as a robust spec, and throw millions at a dev team to rewrite it using… whatever, Dvorak-layout keyboards or something… and get the correlation you wanted.

And especially if you’re after performance as your trophy, the initial release of the rewrite is the ideal time for that, before all of those pesky bugfixes, security gaps, compatibility concerns, feature changes, and observability requests come in to bog things down.

Edit:

Same thing happened with the security reporting around Mythos/Fable. Mainstream media said “look how much better AI is at finding issues!”, but if you looked at the neckbeards who were discussing the kind of budget that issue-hunting stunt received, the real story was “look how underfunded security has been for decades! …and will continue to be, once the press cycle is over”.