ChatGPT, GPT-3, DALLE, MidJourney, Stable Diffusion…
What do all of these AI models have in common? Other than being way too expensive to create and very complicated to understand unless you watch my articles ;), they are all models regular people like us can use to generate amazing results or even create products through prompting.
Prompting is pretty much the only skill you now require to be a master of these new large and powerful generative models, whether it be to generate cool stories, amazing images, or any cool feature like a text summarizer or automatic video editor tool.
The results of these models are only getting better with time, but it is often hard to get exactly what you are looking for. While this is also only improving, the best way to leverage these large models’ potential is to be a better prompt engineer.
But what is a prompt engineer? Moreover, what is this prompting thing I have been talking about for the past minute or so and how can we learn this skill and get better?! Well, that’s exactly what we are going for in this video…
►Learn Prompting: https://learnprompting.org/
►Our Twitter accounts: https://twitter.com/Whats_AI https://twitter.com/SanderSchulhoff https://twitter.com/towards_AI https://twitter.com/learn_prompting
►Discord: https://www.louisbouchard.ai/learn-ai-together/
0:00
chat GPT gpt3 Dali mid-journey stable
0:04
diffusion what do all these AI models
0:06
have in common other than being way too
0:09
expensive to create and very complicated
0:11
to understand unless you watch my videos
0:13
they are all models Regular People Like
0:16
Us can use to generate amazing results
0:18
or even create products through
0:20
prompting prompting is pretty much the
0:22
only skill you now require to be a
0:25
master of these new large and Powerful
0:27
generative models whether it be to
0:29
generate cool stories amazing images or
0:32
any cool feature like a text summarizer
0:34
or automatic video editor tool the
0:37
results of these models are only getting
0:39
better with time but it's often hard to
0:41
get exactly what you are looking for
0:43
while this is also only improving the
0:45
best way to leverage these large models
0:47
potential is to be a better prompt
0:50
engineer but what is a prompt engineer
0:52
moreover what's this prompting thing I
0:55
have been talking about for the past
0:57
minute and how can we learn this skill
0:59
and get better well that's exactly what
1:02
we are going for in this video
1:04
prompting is how humans can talk to AIS
1:07
it's a way to tell the AI what you want
1:10
and how you want it usually done with
1:12
words for example you might want to
1:14
generate an image but have nothing to
1:16
show it as an example what would you do
1:19
you'd do the same thing as with an
1:21
artist you will hire try to describe
1:23
what you want with the best words
1:25
possible and some sketches if you can
1:27
draw this is the same thing with today's
1:29
most powerful AIS
1:31
as we saw in my previous video about
1:33
word embeddings AIS can understand words
1:36
though they understand them differently
1:39
than we do this is where prompt
1:41
engineering comes into play prompt
1:43
engineering is the task of finding the
1:45
prompt that gets the best results from
1:48
the AI a prompt engineer will translate
1:50
your ID into words that the AI can
1:53
understand it's a translator between
1:55
human language to AI language you can
1:58
see print engineering as the same as an
2:00
expert googler lots of people get hired
2:03
simply because they know how to use
2:05
Google well and get answers to their
2:07
questions especially programmers
2:09
compared to your dad who may struggle to
2:12
find what he wants that's because you
2:14
know what Google's search box is
2:16
expecting and which keywords are good or
2:19
not whereas your dad does not he's
2:21
typing as if he's asking a question to
2:24
another human being which the machine
2:26
cannot understand nearly as well
2:28
let's dive into some examples of how
2:31
learning prompting skills can help you
2:33
out in your AI Journey then we'll dive
2:35
into a very cool prompting course my
2:37
friend sander and I are working on to
2:39
help you improve this communication
2:41
skill and become a real prompt engineer
2:45
text prompting is a way to talk to
2:47
language models the best way to
2:49
understand this is by looking at a
2:51
simple example let's ask gpt3 to solve a
2:54
math problem we can see that it will get
2:56
the wrong answer however we can further
2:59
adapt it to the AI like the simple
3:01
addition of let's think step by step and
3:05
it will succeed image prompting is quite
3:08
similar we can use words to describe an
3:10
image to a text to image model and see
3:13
what it generates we see that as we
3:16
change different descriptors the images
3:18
that the model generates change
3:20
accordingly the code behind prompt
3:22
engineering is to optimize that we
3:25
believe it's important to learn this new
3:27
skill as large language models will only
3:29
become more available and more powerful
3:32
it's obvious that people will get paid
3:34
just to use and master them and we want
3:37
to help you reach that to this end along
3:40
with my friend sender and tours AI we
3:43
are building a course on planting called
3:44
called line prompting where you can
3:47
learn a lot about good printing
3:48
practices and cool techniques or even
3:51
tips from experts specifically oriented
3:54
toward the model you are tackling
3:55
whether it's chat GPT mid-journey or
3:58
anything else check it out to learn more
4:00
about printing and get insights from
4:03
experts in the field this is an open
4:05
source work in progress and we aim to
4:08
continually improve this course and add
4:10
cool new sections over time we'd love it
4:13
if you want to contribute and share your
4:15
knowledge as well with more people
4:17
through this platform join the learn AI
4:20
together community and Discord to learn
4:22
more about this new course and chat with
4:25
us all the links are in the description
4:27
below thank you for watching the video
4:29
and follow me sander and towards the eye
4:32
on Twitter to stay up to date with the
4:34
new sections in this course