Disagree. The journals are now acting like a paid certification. If they admit any old slop, who would pay to submit their papers?

The service they are providing is peer review and applying a reputable quality bar to submissions.

Think of it this way, if you have a good paper why would you publish on Arxiv instead of Nature? And then if you are Nature, why would you throw away this edge to become a free-to-publish (non-revenue-accruing) publication?

Small correction to your point: they perhaps provide a reason for peer review to happen, but it's scientists themselves who coordinate and provide the actual peer review.

That is, unless ACM and Nature have a different approach to organizing peer review, in which case my correction is wrong. But I believe my point stands for many conferences and journals.

Reputable quality bar isn't the right metric. Quality is a better metric. To the extent it can be estimated, impact is another. Neither of these require journals specifically.

A different way to look at this is to question what "old slop" actually means.

The reason not to publish in Nature is that it might take a long time to get everything right in the paper to publish, to the point it takes years to get it read. Publishing fewer results faster spreads the results faster.