I'm a bit of a scholarly infrastructure purist. The paper has a DOI, it leads to a landing page that has the full text, and the content is open licensed.

Like if someone posted a link to an archive.is version of a Wikipedia page, you'd probably prefer to get the canonical link to that content.

ResearchGate is a bit of commercial enclosure of infrastructure that is, and should be, open. Who knows, maybe it has other value. I'm not an academic so I don't know.

My position is that when it’s open access, we might as well link the primary source. ResearchGate is generally legal. It’s the responsibility of the authors to upload accepted manuscripts if the final document is not open access. AFAIK it does not do anything dodgy like archive.is does.

They're a user hostile attempt to extract money from people. They make their website hard to use.

If the original is available, posting anything else is by definition less desirable.

ResearchGate isn't open access.

The researchgate link has for me a deceptive ad at the bottom: it has a PDF logo and says "download now", suggesting that this will download the paper. It then links to a download page which in the fine print says it will actually download some kind of ebook collection which costs 30 bucks per month. That's a scam.

As far as I understand this is a pre-print under CC-BY license, if this answers your question?

The clue is in the name, research "gate"