Lossy vs non-lossy compression is orthogonal to the sampling rate and bit-depth (which is what this article is about). While the MP3 standard effectively means sampling more than 48kHz is useless, there's no reason you can't have a lossy comprssion scheme that attempts to capture higher frequencies.
But the article makes a compelling case for why > 48kHz is completely pointless.
A lot of listening today occurs through streaming services, and huge uncompressed songs will eat into data plans quickly.