The core knowledge in STEM would be easily recoverable. There’s a lot of people with great memories. Even tricky stuff about the right way to mix things in sequence in a chemical reaction is in somebody’s head somewhere. (But see below.)
Other fields might not be so lucky. E.g., literature. How many people have memorized a few well known books as in Fahrenheit 451? And the “next tier” of books is going to be almost completely gone.
Wiping all the computer software off the planet is going to be a problem. A basic form of Linux could be rebuilt in a year. Windows would have to be redone from scratch and would take a good long while. (OTOH, it’d probably be less buggy.)
As noted though, there’s a chicken-and-the-egg problem here. To get all the Linux developers out there to communicate, allocate work, check others’ work, etc. requires basically the Internet. So you have to figure out a way to get all that up and running to a small extent using old tech (e.g., “sneakernet”).
One big issue is where software meets process. I.e., a computer program was used to work out how to control a chemical reaction. The data for all this was on the computer. All that’s wiped. Right now it wouldn’t be a long term issue to start over and redevelop that. But as this gets more common and becomes more complicated there could be data on some complex processes that would never be recovered as it used to be. A similar process might be found but there’d be no way of ensuring it was the same as before.
While all this is happening of course there’s no problem with food, energy, clothing, etc. distribution, right?