My take on that, with some background...
When one FAANG was very young, I knew a non-computers researcher who'd heard of the company and wanted to work there. But I figured that the company was still at the stage that they thought they only needed computer people and a few business people. (So I'd expect the non-computer researcher's resume, submitted cold, to be shredded by HR after a 1-second glance.)
So I asked a kindly and famously well-connected professor whether they had a contact there. Something like, so the researcher could explain what value they could bring that the company might not yet realize it needed.
Turns out, the professor had a strong connection to one of the top Poobahs at the company, and arranged an introduction for the researcher.
But later, when I wanted to work at that same company myself, and the company asked for a list of people I knew at the company, there were some, but I didn't feel comfortable doing that. Those contacts didn't have that much familiarity with my work, it felt more like an old-boys network or nepotism rather than a justifiable use of connections, and my software engineering resume should've been strong enough to get it in front of someone who could guess that I'd bring useful things.
Going back to the people who cold-contact strangers on LinkedIn for a referral: I'm not sure that's positive signal. It shows some effort, but it could also be seen as trying to game the system, or to abuse an employee's potential conflict of interest when referral bonuses are involved. Surely there are more unambiguously positive signals to be found in that stack of 200 resumes?