WEBVTT

00:00.150 --> 00:02.700
-: ChatGPT only has a certain number of tokens

00:02.700 --> 00:05.190
that it can generate as a maximum output.

00:05.190 --> 00:06.780
This could be problematic when you're looking

00:06.780 --> 00:08.940
to generate larger outputs.

00:08.940 --> 00:11.220
For example, if you wanted to write a book

00:11.220 --> 00:13.830
or you wanted to generate a very large article

00:13.830 --> 00:16.320
or blog post, these are areas

00:16.320 --> 00:19.050
where you cannot just simply put one prompt in

00:19.050 --> 00:20.790
and get the entire output.

00:20.790 --> 00:22.290
Here's a technique that you can use

00:22.290 --> 00:24.750
so you can increase the output sizes

00:24.750 --> 00:26.520
that you get whilst using ChatGPT.

00:26.520 --> 00:28.110
Firstly, we're gonna say,

00:28.110 --> 00:31.950
can you please generate a blog post outline

00:31.950 --> 00:35.440
for an article on what is the benefits

00:37.080 --> 00:38.523
of digital marketing?

00:39.780 --> 00:43.650
So rather than telling ChatGPT, can you make an article,

00:43.650 --> 00:45.390
what we're really telling it is,

00:45.390 --> 00:46.710
give me the article outline,

00:46.710 --> 00:49.050
or give me the rough structure of the larger piece

00:49.050 --> 00:51.060
of content that we're going to need to generate.

00:51.060 --> 00:54.510
Then once we have that outline, we can then tell ChatGPT

00:54.510 --> 00:57.270
to generate each part separately,

00:57.270 --> 00:59.850
because then we can make it very long and extensive.

00:59.850 --> 01:02.550
And then we'll do the same for number two and number three.

01:02.550 --> 01:04.740
And that is an easy way for us

01:04.740 --> 01:06.630
to extend the output generation

01:06.630 --> 01:09.060
that we get directly within ChatGPT.

01:09.060 --> 01:11.310
So for example, we can take the introduction,

01:11.310 --> 01:12.810
we can paste this in and say,

01:12.810 --> 01:17.810
write a really large and comprehensive introduction,

01:18.810 --> 01:20.520
and we'll basically just do that

01:20.520 --> 01:22.650
all throughout the blog post outline.

01:22.650 --> 01:26.280
And this overcomes the limitations of the maximum number

01:26.280 --> 01:29.130
of tokens you can get through one generation.

01:29.130 --> 01:30.960
Basically, once the introduction is done,

01:30.960 --> 01:32.610
you just rinse and repeat that technique

01:32.610 --> 01:34.380
for how many pieces of content

01:34.380 --> 01:36.750
and the sub pieces of structure that you have

01:36.750 --> 01:38.190
and whatever you are generating.

01:38.190 --> 01:41.700
You can use this to build eBooks, very large blog posts.

01:41.700 --> 01:43.350
You can use this to build anything

01:43.350 --> 01:46.200
that won't fit in one single ChatGPT response.

01:46.200 --> 01:48.630
This is an effective technique for you to extend

01:48.630 --> 01:50.760
and overcome the fact that there are only a certain number

01:50.760 --> 01:54.510
of tokens that ChatGPT can reply to you in a single message.

01:54.510 --> 01:56.490
Cool, so we'll just do that again and go through

01:56.490 --> 01:58.770
and we'll do the reach a global audience.

01:58.770 --> 02:01.740
And you can see here, ChatGPT is picked up on the fact

02:01.740 --> 02:04.710
that we likely want to extend that section

02:04.710 --> 02:06.960
whilst continuing the same tone and depth

02:06.960 --> 02:08.280
as the introduction.

02:08.280 --> 02:11.160
Cool, hopefully this video gives you a good technique

02:11.160 --> 02:14.853
for how you can produce larger outputs using ChatGPT.