The risk is low for desktop applications that might be using log4j (e.g. MPLABX). An attacker would need to cause something to be logged by the application, which if it isn't interacting with an attacker-controlled service on the Internet, would be unlikely for a desktop app (unless it's a browser or some such). So I wouldn't be too worried as an end user.
If you host applications or use hosted applications written in Java, it's fairly likely that you are vulnerable, and there are a wide variety of potential vectors for the exploit, since all it requires of the attacker is to be able to cause some chosen text to be emitted as a log entry. That could be via a URI, request header, filename, etc. etc. etc. it doesn't matter, and would be hard to track down and verify that all cases are not vulnerable. So if you are hosting any such applications, either upgrade or apply a mitigation ASAP.
It allows remote code execution and is pretty easy to exploit, so is very very serious, as such exploits go, which is why everyone is (hopefully was, since they have mitigated by now) panicking.
CloudFlare has a good writeup on how the vulnerability works, as well as how to mitigate:
https://blog.cloudflare.com/inside-the-log4j2-vulnerability-cve-2021-44228/