[00:00] (0.08s)
Hold on. So, you're using AI during your
[00:02] (2.40s)
interview process. Yes. Oh, you're not
[00:04] (4.72s)
running away from it. No, we're
[00:05] (5.92s)
embracing it. Okay. How is it working?
[00:07] (7.20s)
Tell me. I love it because what happens
[00:08] (8.56s)
now is the AI will sometimes generate
[00:10] (10.40s)
pure garbage. So, you're screen sharing
[00:11] (11.84s)
and you say, "Let us use that screen
[00:13] (13.12s)
sharing. Let them use whatever they
[00:14] (14.48s)
want." Here's what I'll say. If they
[00:15] (15.52s)
don't use a co-pilot, they usually get
[00:17] (17.36s)
creamed by someone who does. So, they
[00:19] (19.28s)
will have no choice but use a co-pilot.
[00:20] (20.88s)
Sometimes I will shadow an interview and
[00:22] (22.16s)
do the questions myself. I've never seen
[00:23] (23.52s)
them with a co-pilot and send it to the
[00:24] (24.88s)
interviewer and say, "Please mark my
[00:26] (26.32s)
assignment as well against a candidate.
[00:27] (27.92s)
I've not lost yet." If they don't have a
[00:29] (29.44s)
co-pilot, they will lose. But when they
[00:30] (30.80s)
do have a co-pilot, I love seeing the
[00:32] (32.48s)
generated code because I want to ask
[00:33] (33.92s)
them, what do you think? Is this good
[00:35] (35.44s)
code? Is this not good code? Are there
[00:36] (36.64s)
problems? And I've seen engineers, for
[00:37] (37.92s)
example, when there's something very
[00:39] (39.28s)
easy to fix. They won't fix it. They
[00:40] (40.80s)
will try to prompt to fix it. And I see,
[00:42] (42.72s)
are you really an engine? I get the
[00:44] (44.48s)
nuance of like just prompt and prompt
[00:46] (46.08s)
and prompt, but sometimes it's right
[00:47] (47.36s)
there and they will not change the one
[00:49] (49.20s)
character. They won't change it. I don't
[00:50] (50.64s)
want you to be 100% AI coder. I want you
[00:52] (52.16s)
to be like 90 or 95. I want you to be
[00:54] (54.00s)
able to go in and look at the code and
[00:55] (55.28s)
say, "Oh yeah, there's a line that's