Respekt vor der Leistung. Andere Ausschnitte aus der Sendung prallen leider komplett an den Rhetorikfähigkeiten ab, da merkt man 'wie leicht' es ist im 1:1 einfach alles abzubügeln oder abzulenken und dabei nichtmal schlecht für die eigene Zielgruppe auszusehen :(

Yes definitly. For small coding challenges I'd always encourage use of LLMs and wait for the canditate that surprises us by being twice as fast with the same high quality final result while saving 50% of keystrokes.
And thus eventually even inspire the interviewing team.

I'm sure this will happen.

This is an important thing. However 'understanding code' cannot be tested with a code challenge that forces you to code without LLM?

So if we consider this to be important, I'd rather then create a 'make improvement suggestions for this (big) code snippet and refactor with an LLM' hiring task?

I understand people that don't see benefits in Copilot yet, however using Cursor since two weeks and not writing code anymore but only intentions is yet again a complete new level of assistance.

So hiring without allowing LLMs looks for the wrong skills for the next years imho.

Is 'can code without LLM' really a skill for Juniors in the future?

I rather think that we now can get much faster to 'how good are you in breaking down problems and how fast do you iterate towards a solution (with an LLM)' . Even for Juniors we can now use more complex (+ interesting) challenges.