Simpson's paradox requires that you be looking at averages. An example would be if you compared average salaries of SF vs. Austin, and both tech and non-tech salaries were rising in both places, but tech workers were moving from SF to Austin so the average includes relatively fewer tech workers in SF and relatively more in Austin. The article is comparing specifically tech salaries, so it's free from this sort of bias.
It could have a similar bias by virtue of all being collected from Hired, though. Top tech companies do not use Hired. Hot startups generally do not use Hired. Mid- and low-tier tech companies that want to outsource their interview processes use Hired. If these are also primarily the companies that are going remote-first or moving their engineers to Austin or SD, you would expect to see tech salaries rise in those locations and fall in HCOL regions like the Bay Area. It's a real effect, but not necessarily reflective of the whole industry, which includes a number of large and very high-paying employees that aren't in the dataset because they don't use Hired.