The top 15 schools according to the report are UVa, Yale, CMU, MIT, UCLA, UPenn, Rice, Caltech, UCSD, Swarthmore, UC Boulder, Columbia, Stanford, USC, and UIUC.

Depends on the source. Here's from US News 2022 Computer Science Best Undergrad List:

https://www.usnews.com/best-colleges/rankings/computer-scien...

Carnegie Mellon

MIT

Stanford

UC Berkeley

Cornell

Georgia Tech

UI Urbana-Champaign

Cal Tech

Princeton

UCLA

UT Austin

UW Seattle

This is voluntary data, the average CS grad from Berkeley doesn’t need a code signal test to find a job. The cream of the crop at Swarthmore likely does.

In my experience, coding tests for fresh graduates is a pretty terrible metric for how well that person would perform as a member of a software development team for business applications.

Data/science/research is a different story with a different skillset.

whats the point of even caring about what school a person went to if you can rank candidates based on some test score?

I felt pretty proud of myself when I scored 99 percentile on a codesignal exam. Especially since I consider many of their tests to be more legitimate than leetcode.

Some of their tests though are odd. I don't think the bugfix style is that good. But that also might be because I would just rewrite a buggy function

I've never heard of the General Coding Assessment. I don't know anyone who ever put his or her score on a resume and have never heard of it from anyone in the valley.

What's the point for a student to take it?

I've been out of software for about two years...is this test really now a thing or is this a submarine article to promote its use?

easy. github. no contest.