WEBVTT 00:00.960 --> 00:02.370 -: Hi, I just wanna give you 00:02.370 --> 00:05.010 an inside view of the curriculum. 00:05.010 --> 00:08.040 So this is how we think about the course, 00:08.040 --> 00:09.480 and it's got a lot in here. 00:09.480 --> 00:12.000 So I'm just gonna walk you through my intent, 00:12.000 --> 00:13.290 like what we were trying to do 00:13.290 --> 00:14.730 with each checks of the course. 00:14.730 --> 00:16.080 And this might change by the way, 00:16.080 --> 00:18.840 but we keep it updated every month. 00:18.840 --> 00:21.090 I think we add new content 00:21.090 --> 00:22.800 and sometimes we film the old content 00:22.800 --> 00:24.720 as well as it gets out of date. 00:24.720 --> 00:27.210 So if you spot anything, by the way, just let us know. 00:27.210 --> 00:28.890 We have 150,000 students 00:28.890 --> 00:31.140 and every one of them gives good feedback. 00:31.140 --> 00:32.970 So we appreciate that. 00:32.970 --> 00:34.200 We start with the introduction, 00:34.200 --> 00:35.490 and this is just getting started, 00:35.490 --> 00:37.530 teaches you what is prompt engineering. 00:37.530 --> 00:41.070 We have a few different resources and things here as well. 00:41.070 --> 00:43.800 And then also the GitHub repository for the course. 00:43.800 --> 00:45.810 We try and keep everything in there. 00:45.810 --> 00:48.000 Then we get into the five principles of prompting. 00:48.000 --> 00:50.640 And this is the most important thing in the course. 00:50.640 --> 00:52.920 You'll see these techniques everywhere, 00:52.920 --> 00:55.080 but this is a checklist that you can go through 00:55.080 --> 00:57.480 and once you know how prompting works, 00:57.480 --> 00:58.560 like from this section, 00:58.560 --> 01:01.770 you'll be able to understand most of the other sections. 01:01.770 --> 01:05.340 There's also a section here on how does AI work. 01:05.340 --> 01:08.220 If you're not too interested in this, you can skip it. 01:08.220 --> 01:09.330 That's totally fine. 01:09.330 --> 01:11.070 And then same thing for this section here. 01:11.070 --> 01:14.040 We're going into each feature of ChatGPT, 01:14.040 --> 01:16.080 just giving you a bit of an introduction to them. 01:16.080 --> 01:17.850 And if you haven't seen some of these things, 01:17.850 --> 01:18.750 feel free to watch them, 01:18.750 --> 01:22.140 but you also don't feel compelled to go into it 01:22.140 --> 01:24.690 if you're already an AI expert. 01:24.690 --> 01:26.880 Now, where it starts to get really interesting 01:26.880 --> 01:29.070 is in the standard text model practices. 01:29.070 --> 01:31.320 Typically we call it a standard practice 01:31.320 --> 01:33.270 if it doesn't need code. 01:33.270 --> 01:37.136 And about 30, 40% of the videos don't really need code. 01:37.136 --> 01:41.100 And then we call it a advanced technique 01:41.100 --> 01:42.420 if you need to know how to code 01:42.420 --> 01:45.870 and we'll put the -Coding on the end of each one. 01:45.870 --> 01:47.940 There's also a section here 01:47.940 --> 01:50.160 on GPT-4 and how to use the Playground, 01:50.160 --> 01:53.640 which is needed before the advanced techniques. 01:53.640 --> 01:55.380 And then we go into LangChain. 01:55.380 --> 01:57.780 Love it or hate it, it is the most popular framework 01:57.780 --> 01:59.580 and most of the other frameworks 01:59.580 --> 02:01.110 are in some way inspired by it. 02:01.110 --> 02:03.240 So I think it's really worth learning 02:03.240 --> 02:05.040 and you can just look under the hood 02:05.040 --> 02:07.410 and see how they do certain things 02:07.410 --> 02:10.290 'cause they do follow pretty good best practices 02:10.290 --> 02:11.820 and there's a lot in here. 02:11.820 --> 02:13.320 So again, just dig into the things 02:13.320 --> 02:15.030 that you find interesting. 02:15.030 --> 02:16.050 We also go into LangGraph, 02:16.050 --> 02:17.790 which I think is actually their best product, 02:17.790 --> 02:20.640 but there's a lot more to say about that. 02:20.640 --> 02:22.230 I'll let you dig into that section. 02:22.230 --> 02:24.330 We then do some real-world projects. 02:24.330 --> 02:27.120 So these are actual things that we've done for clients, 02:27.120 --> 02:30.060 specific videos for each type of project. 02:30.060 --> 02:31.440 We also even help you build 02:31.440 --> 02:34.590 your own GPT wrapper if you need as well. 02:34.590 --> 02:35.790 So enjoy that section. 02:35.790 --> 02:36.990 There's a ton in there. 02:36.990 --> 02:38.567 And then we talk about Midjourney 02:38.567 --> 02:40.440 'cause that's the main model 02:40.440 --> 02:43.770 we're using for some of these simpler sections. 02:43.770 --> 02:48.330 We also talk a bit more about image prompting in here. 02:48.330 --> 02:50.370 So quite useful section, 02:50.370 --> 02:53.610 before we get into the harder image generation stuff. 02:53.610 --> 02:56.400 Reason it's harder is just because setting up 02:56.400 --> 02:59.730 Stable Diffusion in the WebUI is a bit of a pain. 02:59.730 --> 03:01.290 There, we do have some links 03:01.290 --> 03:03.750 in this lecture for you to follow. 03:03.750 --> 03:05.250 And then also how to migrate 03:05.250 --> 03:07.890 to Stable Diffusion Excel because most of the content 03:07.890 --> 03:11.550 is based on 1.5, which is still a very popular model 03:11.550 --> 03:13.140 even though it's quite old 03:13.140 --> 03:15.930 because it's the most flexible and trainable. 03:15.930 --> 03:19.143 But a lot of people are starting to switch to Excel. 03:20.190 --> 03:23.310 All this section is still useful for Excel as well. 03:23.310 --> 03:25.170 We're really teaching you the individual techniques 03:25.170 --> 03:28.140 like inpainting, ControlNet, Segment Anything. 03:28.140 --> 03:30.480 And then we have DreamBooth here as well. 03:30.480 --> 03:32.160 That's fine-tuning. 03:32.160 --> 03:34.170 And then we do some image model projects. 03:34.170 --> 03:36.540 So these are real-world projects that we've done. 03:36.540 --> 03:37.830 And then get into evals. 03:37.830 --> 03:39.210 And when you are in the real world 03:39.210 --> 03:41.460 and you're doing prompt optimization 03:41.460 --> 03:44.670 or you're a prompt engineer as your job title, 03:44.670 --> 03:46.680 you're gonna be doing a lot of this. 03:46.680 --> 03:48.510 This is probably 80% 03:48.510 --> 03:49.710 of the work is finding 03:49.710 --> 03:52.830 the right performance metrics and then optimizing to them. 03:52.830 --> 03:54.270 So take a look at that section. 03:54.270 --> 03:56.190 That's my personal favorite section. 03:56.190 --> 03:58.320 We also talk a bit about agent architectures, 03:58.320 --> 04:00.150 which are starting to get more and more important. 04:00.150 --> 04:04.680 AI agents is the next big topic after prompt engineering, 04:04.680 --> 04:05.970 but then we cover 04:05.970 --> 04:08.130 all the other kind of things that we missed. 04:08.130 --> 04:09.870 You can jump ahead to this section by the way 04:09.870 --> 04:11.850 if you want to know a bit more about Claude 04:11.850 --> 04:14.520 or dive into DALL-E or some of the other models. 04:14.520 --> 04:17.040 Like we talk a lot about the Google stuff here. 04:17.040 --> 04:19.800 In particular, Gemini to native image generation 04:19.800 --> 04:21.540 I think is worth looking at. 04:21.540 --> 04:24.000 We cover a few different tools we've tried. 04:24.000 --> 04:26.460 We haven't tried that many tools that we liked, 04:26.460 --> 04:27.293 to be honest. 04:27.293 --> 04:28.950 So this is really only the stuff 04:28.950 --> 04:32.481 that we've found useful and we keep this section updated. 04:32.481 --> 04:35.430 Some of these are some prompt optimization libraries, 04:35.430 --> 04:36.900 which are quite cool. 04:36.900 --> 04:40.230 And then we go a little bit into GitHub Copilot. 04:40.230 --> 04:41.160 So it teaches a little bit 04:41.160 --> 04:43.050 about how to prompt better 04:43.050 --> 04:46.860 when you are using AI for code generation. 04:46.860 --> 04:48.360 And then the conclusion, by the way, 04:48.360 --> 04:49.800 you get a free chapter of our book. 04:49.800 --> 04:51.960 We publish a book through O'Reilly, 04:51.960 --> 04:54.210 so feel free to download that PDF 04:54.210 --> 04:56.310 and see how different the content is. 04:56.310 --> 04:58.170 But we also give you links here 04:58.170 --> 05:00.690 to different sources of inspiration, 05:00.690 --> 05:03.390 things that inspired us when we're creating this course, 05:03.390 --> 05:06.360 'cause there is a lot of really good content out there. 05:06.360 --> 05:08.820 Cool, okay, hopefully that gives you a good understanding 05:08.820 --> 05:09.690 of what we're trying to do 05:09.690 --> 05:13.020 with the course and what's contained inside it. 05:13.020 --> 05:16.773 We do try and keep it updated, so this is subject to change.