Thank you for this!
It is deeply disturbing thatit always needs a catastrophe to get people to act. People have warned about the fickleness of databases for decades, see, e.g., slides 19-22 from 2013:
https://www.slideshare.net/slideshow/the-three-infrastructure-crises-in-science/15079538#19
or with explicit PubMed mention on slide 6 in 2014:
https://www.slideshare.net/slideshow/making-science-open-by-default/31885784#6
We have known that this was a problem for many, many years, I was not the only one pointing this out, of course.
And yet, nothing happened.
Even now, when people start #SafeguardingResearch we are making the same mistakes again! As you point out in your article, all our outputs are in danger, be it research data or papers. And GitHub is a single point of failure foir scholarly code, too!
All of our outputs are under threat. If we had to invent infrastructures for our outputs, nobody would invent different silos for each output, so linking code, data and text together would be made more difficult, rather than easier?