I have a knack for “finding” non-free epubs, so sometimes I read those.
But for CC stuff, what I often do is find interesting documents from Media Archive. One problem with this is that a lot of books - or, at least the ones I happen to be looking for - are often badly OCR’d from scanned paper, automatically generated with no corrections applied. So every so often what I do is patiently OCR the PDF version and correct everything in LibreOffice Writer. So I am basically transcribing the book as I read it. Then I use a plugin called eLAIX to export the Writer document to an epub file I can load onto my reader for next time. And/or release on the interwebs for others who want a decent copy which is actually readable.
It’s a bit of bother, but I often think worth doing.