I agree with the broad point of the article, but the author misidentifies what's going on. The author's problems are coming from digital printing, not the print-on-demand business model specifically and Amazon isn't the only company doing it.

The older books were printed using a process called offset printing. It needs large economies of scale to be financially viable, but it produces higher quality books. The newer books are printed with digital printing, which is just a fancy version of the laser (typical) or inkjet printer you have at home. I believe Amazon POD uses inkjet, but not sure. The result is a worse quality book, but also one that doesn't have thousands of copies taking up inventory space until it's sold. Virtually all publishers are moving low volume works this way. The fact that the quality is merely "subpar" instead of unusable is a testament to how much digital printing has improved in recent years.

Separately, paper quality has gone down industry-wide. Paper mills are simply choosing to focus on higher volume papers like those used in cardboard instead of producing fine paper. That means shortages, price increases, and publishers making do. Also, POD publishers don't want to keep every type of paper under the sun. They standardize inventory to keep prices down.

To make things even more confusing, the same work might be printed using multiple methods and different papers, with different inks. It's common to do a first run with POD to gauge market demand and then offset if sales continue. Or offset for a collector's edition, or vice versa to allow more colors.

> The older books were printed using a process called offset printing. [...] it produces higher quality books. The newer books are printed with digital printing, which is just a fancy version of the laser (typical) or inkjet printer you have at home. [...] The result is a worse quality book

Offset printing doesn't necessarily give better results than an inkjet/laser printer. My cheap laser printer from Costco produces much better output than most newspapers, and slightly better output than most old paperbacks. Fancy magazines are also printed using offset lithography and they do indeed have better print quality than my cheap old printer, but if I bought a better printer, then they'd be tied on quality again.

> It needs large economies of scale to be financially viable

Yes, but it needs much fewer copies than you'd expect. I help publish a small magazine [0], and we print it using offset lithography. We print roughly 700 copies of each issue, 3 issues a year, and 100 pages per issue. When I priced it out a couple years ago, offset printing was still cheaper than print-on-demand as long as we were printing at least ~200 copies.

Now, 200 copies is still quite a lot, but it's small enough that nearly every non-vanity-published book should have no problem selling that many copies. My impression is that the move to POD is not to reduce printing costs, but to reduce warehousing costs and the risk of overproduction. This is mostly a non-issue for us, since all the subscribers pre-pay for the whole year upfront and we mail the copies as soon as they're printed, but is much more of an issue for books where some unknown number of people will buy copies over some unknown amount of time.

The other big advantage of POD is that you can print it close to where the buyer lives. For the magazine that I help with, the cost of printing is almost a rounding error compared to the cost of international shipping, so it wouldn't surprise me if this is a major motivator for the big publishers too.

> Separately, paper quality has gone down industry-wide. Paper mills are simply choosing to focus on higher volume papers like those used in cardboard instead of producing fine paper.

Ah, that is not something that I was aware of, but now that you mention it, it does seem to match my impressions.

[0]: https://tug.org/TUGboat/

That's one part of the story, but I think you're glossing over two other issues.

First, digital printing allows anyone to sloppily OCR public-domain works (or download them from Project Gutenberg), typeset the text haphazardly, and put it on Amazon. The result is terrible for reasons that have little to do with the limitations of the technology. Take the Russell book: terrible kerning ("Proble ms"), an AI-generated artwork... and I suspect the rest is about as bad.

The second problem is the technology also encourages "real" publishers to aim lower because there's no up front investment at stake? If you have an older, low-volume book, providing a shoddy version will make you more money than letting it go out of print.