Some of the subjects it treats are low-level enough (and of Unix-ish heritage) that their treatment is probably not irrelevant today. The performance considerations of threads and processes are probably out of date, but their API (sections 3 and 4) is largely still the same. Same for shared memory and other IPC mechanisms described in section 5. The API re. section 7 still applies -- /proc has expanded considerably since 2001, but it's still used in the same manner. Chapter 8 is probably more out of date than others, but you can still follow it. Stuff in 9 is probably not used much anymore, an #10, while adequate, would probably be considered incomplete today.
I'd say it's still useful to someone who comes to Linux from other platforms, or to people who are already using Linux but aren't familiar with systems programming under it.
Edit: although, as someone else pointed out, APUE is probably a better choice (and was a better choice back in 2001, too).
If you're looking for a better maintained book, I'd recommend "Advanced Programming in the UNIX Environment" by W. Richard Stevens, revised by Stephen A. Rago.
It's not a free PDF however, but well worth the money.
With the exception of systemd versus init, the big ideas are pretty much unchanged. Details around hardware and distribution features should be kept in historical perspective and application context, by which I mean that a 2001 64 megabytes of RAM desktop might be today's embedded system and most people won't have to configure networking over two dial up modems.
For me, the critical issue with technical books is readability and accessibility. If it's in front of me and I enjoy reading it because I'm learning something, that's often good enough. If I were already an expert Linux programmer, then I might have different criteria. I know I ain't.
This book is dated - there was the kernel 2.4 and a lot of system calls were changed or removed - but I think it is one of the best book available on GNU/Linux system programming.
I love "UNIX Systems Programming: Communication, Concurrency and Threads", it is simply great.
This appears to be the first edition of the book, published in 2001. Is this book still useful, considering it's 15 years old?