Meh, retractions are best case. Publish or perish is causing

-low quality work: anything with a .05 result wins, whatever the contortions necessary

-zero repeatability: one of the science fundamentals, but not publishable.

-related, negative results are not published

-people leaving science for better pay / conditions cause PoP sucks donkey dicks

That's true, and we should also note for further discussion that there are related major issues.

The traditional publishing process through the major journals is entirely broken, because the journals are making a ton of money, and often the research is paid for by taxpayers, and often the researchers and reviewers don't make much at all.

Also, it depends on the country, but many universities are hiring fewer tenured faculty members. When people are on one year contracts, or on several year contracts, they have strong incentive to get things published before they change employer. If they have job security, they're much more likely to do proper research before sending things off.

Nobody wants to destroy their own reputation, but they are more likely to take a risk if it's a make or break point in their career.