AN ESTIMATE ON COMPANY'S FINAL PAYOUT FIGURE, page-232

Note that he wouldn't be able to mention anything that hasn't been publically announced and so even if they had observed higher endurance, a podcast wouldn't be the appropriate place to announce it given that previous announcements have been market sensitive.

As for your other question, when testing endurance we typically use an Arrhenious Equivalency or similar to accelerate the testing and get accurate results. This is done at different temperature points which will be what some of the "tuning" here is all about. Different write/read speeds impact temperature which in turn impacts endurance (faster = hotter). Hotter will mean less endurance as we'll start seeing single bit read errors on a static data pattern sooner (though this isn't necessarily linear and instead tends to have a shallow decline into a rapid increase once a particular temperature is hit). So we can go faster, but at the cost of endurance.

That said, the above is a bit simplified and this is also new tech so I don't know if traditional memory logic holds true for it either. Hope that helps!