Mostly a class thing but also a reflection of the fact that the exams are crammable without understanding. MCQs work that way unless you make the question back very large. A computer adaptive test could bypass this but it doesn’t feel like it assesses your knowledge the way even a test composed of short answer questions does. It’s not like anyone in the US who’s passed the bar exam holds it in particularly high esteem and for practical purposes doing that presupposes you have a JD.
There's a growing sentiment that degrees are the same; it's certainly not the prevailing view.
One part, in my view, is the exceptionally temporary nature of a certificate. A degree typically teaches things that are fairly static.
If you're a biologist, I would expect 90% of the things you learned in college to still be true, or mostly true, when you retire. A certificate, on the other hand, validates that you know this particular thing well, until the next version comes out. Depending on what "thing" is, and when the last version was released, it could be less than a year. So even if I accept that the certificate does mean you know this system really really well, it doesn't mean that's true for very long.
They're also an external validation of your skills, so people who aren't confident in their skills will often mention their certificate when they feel their skills are being challenged. I think this creates a bias in how we view certificate holders. I've noticed the same thing with people who have degrees from prestigious institutions. I've got a bad taste in my mouth from MIT because most of the people who went to MIT and I'm aware that they went to MIT like to bring it up when they don't feel confident. That's not to say everyone who went to MIT is like that; the rest just don't feel the need to constantly mention it, so I don't even know.
They're typically multiple choice as well, which are much easier to fake understanding on. "Which network architecture accommodates X best?" is probably going to be easier to answer than "You have a network with X servers that need Y bandwidth, etc, etc, explain what network architecture you would choose, and why."
And lastly, I think there's a certain sense of derision for certificates for things created in our lifetime. Things that came before us have a sense of tradition that gives them a sense of legitimacy. Of course your accountant needs a degree, they've always had degrees. New things don't have that, so you have to justify why a certification is necessary, and most don't meet the bar. Everyone managed to use GitHub just fine without a certification.
previous experience/internships at good companies that I know have solid engineering and/or CS/Engineering degree from a school I trust.
In my mind, bootcamps, certs and micro-degrees often signal someone was unable to enter the field due to a lack of qualifications. Most of them are pretty much rote memorization of commercial products, so fairly low value, or 'one trick ponies' where they know how to create a web app with one and only one tech stack (by copy-pasting templates!).
Really, I expect anyone with a good degree or previous experience to just be able to read git and GitHub's manuals and get it.
I think part of the reason is that when you choose to get a degree, you don't know any better, and it probably seems as good a way to get into the field as any (in lots of other fields, it's the only way in - e.g. bio-tech, so it's an easy assumption to make). By the time you get far enough in your career to get a certification, you're also far enough along to know that it's worthless, so it's more of a red flag to others if you choose to do so anyway.