WEBVTT

1
00:00:01.720 --> 00:00:10.500
Now, no matter how you use AI, and actually no matter if you use ChatGPT or Cloud or Gemini

2
00:00:10.500 --> 00:00:15.860
in the browser or if you're indirectly using some of these models through GitHub Copilot

3
00:00:15.860 --> 00:00:20.100
Chat or Cursor, you should know how to write good prompts.

4
00:00:21.620 --> 00:00:27.160
And of course I did talk about prompt engineering earlier in this course already.

5
00:00:28.140 --> 00:00:32.540
I just want to come back to this here because it really does matter and it's important

6
00:00:32.540 --> 00:00:37.180
to understand that it does matter no matter how or where you're using AI.

7
00:00:38.160 --> 00:00:45.780
You should always try to be specific regarding your intent, add useful context but not too

8
00:00:45.780 --> 00:00:53.360
much context, don't be ambiguous, add examples if it's helpful, and consider splitting

9
00:00:53.360 --> 00:00:56.140
complex tasks into multiple tasks.

10
00:00:58.060 --> 00:01:04.560
There also are other techniques, I cover way more of them in my complete ChatGPT course

11
00:01:04.560 --> 00:01:09.360
if you want to learn more about that, but these are the most important things you should

12
00:01:09.360 --> 00:01:14.040
definitely keep in mind when writing prompts no matter where you write them.

13
00:01:15.700 --> 00:01:23.280
You can of course also use advanced techniques like ask-before-answer prompting or RAG to

14
00:01:23.140 --> 00:01:29.400
include extra information for the AI to use when it generates a response.

15
00:01:30.200 --> 00:01:35.560
You can do all that, it's just important to know that prompt engineering matters and

16
00:01:35.560 --> 00:01:41.360
that you should practice writing good prompts because good prompts will yield good results

17
00:01:41.360 --> 00:01:44.440
and that's ultimately of course what you're looking for.