WEBVTT

00:00.000 --> 00:01.440
-: Welcome back, and in this video

00:01.440 --> 00:03.750
what we're gonna be talking about is why it's important

00:03.750 --> 00:06.600
to give detailed instructions to ChatGPT.

00:06.600 --> 00:08.610
Having a look at ChatGPT,

00:08.610 --> 00:10.710
in this specific chat history,

00:10.710 --> 00:12.630
we've got a similar mark down table

00:12.630 --> 00:14.580
for some different types of prompts,

00:14.580 --> 00:16.020
with some lower quality prompts,

00:16.020 --> 00:17.850
and high quality prompts.

00:17.850 --> 00:20.617
As you can see on the left, the lower quality prompts are,

00:20.617 --> 00:23.700
"Can you, for example, give me some book recommendations?"

00:23.700 --> 00:26.791
A more high quality prompt, or a more detailed prompt,

00:26.791 --> 00:27.727
would be,

00:27.727 --> 00:31.080
"Could you recommend any recent science fiction novels

00:31.080 --> 00:33.690
that explore themes of artificial intelligence

00:33.690 --> 00:35.490
and humanity's future?"

00:35.490 --> 00:37.530
This is a good example of where we've got

00:37.530 --> 00:39.810
a very vague prompt, which doesn't really give

00:39.810 --> 00:42.270
any information about what you're looking for

00:42.270 --> 00:44.100
in particular detail.

00:44.100 --> 00:48.030
And, on the right hand side, we have a more high quality,

00:48.030 --> 00:50.010
more in detail prompt.

00:50.010 --> 00:52.057
So, let's also ask ChatGPT,

00:52.057 --> 00:55.050
"Why is it important that we give details to prompts?"

00:55.050 --> 00:56.670
There's a couple of different reasons.

00:56.670 --> 00:59.010
One is, the more specific your question or prompt,

00:59.010 --> 01:01.050
the more targeted and accurate the response.

01:01.050 --> 01:02.790
Also, it gives ChatGPT,

01:02.790 --> 01:04.800
and other large language models,

01:04.800 --> 01:08.190
the contextual understanding to understand your entire scope

01:08.190 --> 01:09.870
of the task or problem.

01:09.870 --> 01:11.640
It's also much more efficient,

01:11.640 --> 01:14.040
and you'll find yourself not having to do lots

01:14.040 --> 01:15.480
of extra prompt hacks,

01:15.480 --> 01:17.370
and the quality of the response you get back

01:17.370 --> 01:18.510
is much greater.

01:18.510 --> 01:21.630
Make sure that you spend more time on your prompts

01:21.630 --> 01:22.923
and make them detailed.