chat

He also doesn't know how to program (in the computer science department) so he has no concept of the actual quality of my code.

This is the biggest of the red flags, but the other communication issues seem serious as well.

Finally understands" my place in on the team, as the one who "Understands things"

This shit is equally enraging and terrifying by proxy. I'm sure this could be charitably interpreted many ways, but in context of this guy's vibecoding, this reeks of someone who would rather press the "theorem is obvious/intheliterature/leftasanexerciseforthereader" button than really engage with the work they're supervising. It's not required (and it would be weird) for your advisor to understand every aspect of your work that they supervise at the depth and facility you do, but they should feel like they broadly understand it and nearly every supporting concept involved unless your work is absurdly multidisciplinary to the extent that your advisor has never really studied computer science directly.

I was being yelled at while doing another thing that was asked for, explicitly due in 2 hours.

Yeah, this and bringing all of this up in a group meeting with other researchers where presumably the task was not to berate a student for being late is more what I was referring to about direct communication.

In my experience, to communicate issues meeting expectations (outside of maybe "lab manager" type duties that might affect everybody's work when whoever is on duty is late or fucks up), professionals only bring this stuff up in direct 1-1 meetings.