Most journals still publish a list of titles in the open, so you can 'follow' influential journals, and read the relevant articles on the arxiv.

Google Scholar scrapes arXiv and coalesces the results with journal updates, i.e. it will show the entry for a new journal article with a clearly visible [PDF on arXiv] link when this is available. Google Scholar also scrapes institutional repos (at universities etc.).

Google Scholar is a service to humanity. However, I do wonder: how comprehensive is it?

is there a similar site for physics papers (getting real tired of the usual new special snowflake melting in the spotlight physics aggregators)

also interested in such sites for optics, engineering (or throw whatever at me what you got!)

ArXiv does physics too. At least for particle physics, it's all there.

I thought this was universal knowledge. I'm not requesting another ArXiv, but a site providing a similar quality selection of papers

Why isn't functionality like this in Google Scholar yet?

Scholar isn't a 'sexy' thing to work on at Google. I'm surprised it still exists, frankly. Google products that aren't career-builders tend to languish.