Google coding round. Went so bad, that I am considering cancelling the rest of interviews. I talked through a general apporach to solve this, but it did not address one of the edge constraint. Of course did not have time to do a dry run, i did not even have a complete algorithmic pseudocode.

the problem: Count the number of squares

x----y-------z----a
|    |            |
|    |            |
b----c            d
|                 |
|                 |
q----w-------f----u
|            |    |
|            |    |
v------------j----k

constraints: cannot use string, char, image or binary for inputs
Squares must have edges to be counted. In above example xycb and xauq are squares, but bcwq and xzfq are not.

This is NOT the detect squares problem. The points must be connected by a sqaure and must also account for squares formed by multiple points.

is this a realistic question to solve in under 40 min, if you have not seen it before?

This is for L5 in US.
Question for those at Google with insight into Hiring committee decisions - is it even worth talking the other interviews? Will failing one coding round be an automatic rejection during HC? does the HC have insight into 'tough' question, or just the interviewer recommendation?

1 coding round went well, others are still pending (1 more coding, design, behavior). Interviewing for general backend (not ML or specialized roles)