I failed the third phase of the Toptal interview process (where you have to solve two coding challenges, each within 20 minutes, while screen-sharing with a hum

I failed the third phase of the Toptal interview process (where you have to solve two coding challenges, each within 20 minutes, while screen-sharing with a human observer). The event was nerve wracking. On the first challenge, I only completing coding and did a first run of my solution with 90 seconds to spare. It didn't give the correct answer, and I had only about one minute to debug and try again. In that case, I felt immense anxiety and nearly gave up. But somehow I managed to keep it together long enough to spot a misplaced array index [ ] reference and produce a correct result. On the second challenge, on my first run at about 17 minutes in I discovered that my entire approach was not general enough to meet all the possible cases. As I recall, while the rules of the problem are stated, you are not able to see test cases until you run. With three minutes left and the understanding that my approach was fundamentally flawed, I just gave up. The experience was demoralizing. On the plus side, when I did live coding challenges for job interviews after that, I found them all comparatively easy and relaxed. What would be better, in my opinion, is an overnight problem to solve, followed by a detailed walk-through of the solution the next day. That presented walk-through can prove to the interviewer that the candidate did not simply copy a solution (without understanding it) from elsewhere on the internet. Of course, if the solution was copied by understood, that should be positively accepted in many cases. There's a reason we use trusted libraries instead of writing our own code where possible.