I still do it for fun, but not methodically, and not regularly. It's a great way to look at code, to learn, and sometimes it pays off.
e.g. Reporting a bunch of trivial predictable filename issues in GNU Emacs, including something referring to the (ancient) Mosiac support:
https://bugs.debian.org/747100
Fuzzing is definitely useful, and I've reported issues in awk, etc, but fuzzing tends to be used when you have a specific target in mind. I'd rarely make the effort to recompile a completely random/unknown binary with instrumentation for that.
I still do it for fun, but not methodically, and not regularly. It's a great way to look at code, to learn, and sometimes it pays off.
e.g. Reporting a bunch of trivial predictable filename issues in GNU Emacs, including something referring to the (ancient) Mosiac support:
https://bugs.debian.org/747100
Fuzzing is definitely useful, and I've reported issues in awk, etc, but fuzzing tends to be used when you have a specific target in mind. I'd rarely make the effort to recompile a completely random/unknown binary with instrumentation for that.