WEBVTT 1 00:00:01.110 --> 00:00:04.170 These days, everyone seems to be talking about 2 00:00:04.170 --> 00:00:06.600 AI tools for programming, 3 00:00:06.600 --> 00:00:10.050 things like ChatGPT, GitHub Copilot, 4 00:00:10.050 --> 00:00:13.920 the new Cursor AI Code Editor, and so on. 5 00:00:13.920 --> 00:00:18.180 There's a lot of hype in the industry around AI right now, 6 00:00:18.180 --> 00:00:21.780 but there is in fact a place for these tools 7 00:00:21.780 --> 00:00:25.140 in web development and programming in general. 8 00:00:25.140 --> 00:00:27.930 And so I think it's important to learn 9 00:00:27.930 --> 00:00:31.710 and to think about this new reality early on, 10 00:00:31.710 --> 00:00:34.800 and address the questions that you might already have 11 00:00:34.800 --> 00:00:37.770 about what this means for the future. 12 00:00:37.770 --> 00:00:40.920 So that's what this lecture is all about. 13 00:00:40.920 --> 00:00:44.730 Let me just warn you that it's gonna be pretty opinionated, 14 00:00:44.730 --> 00:00:48.270 but I believe that this is what a beginner like you 15 00:00:48.270 --> 00:00:50.133 should know and do. 16 00:00:51.120 --> 00:00:52.890 So as I already mentioned, 17 00:00:52.890 --> 00:00:57.210 we're talking about AI tools like ChatGPT, GitHub Copilot, 18 00:00:57.210 --> 00:00:59.730 or the Cursor AI Code Editor. 19 00:00:59.730 --> 00:01:01.680 And by the time you're watching this, 20 00:01:01.680 --> 00:01:03.930 I'm sure there are already a million 21 00:01:03.930 --> 00:01:06.390 new fancy tools out there. 22 00:01:06.390 --> 00:01:08.850 Now, what all of these have in common 23 00:01:08.850 --> 00:01:11.340 is that they're powered by an LLM, 24 00:01:11.340 --> 00:01:15.210 so a Large Language Model like GPT or Claude, 25 00:01:15.210 --> 00:01:16.980 in order to generate code 26 00:01:16.980 --> 00:01:19.920 based on a prompt the user gives the model, 27 00:01:19.920 --> 00:01:21.960 that's basically all they do. 28 00:01:21.960 --> 00:01:24.840 But different tools have different ways of working 29 00:01:24.840 --> 00:01:27.510 and a different user experience. 30 00:01:27.510 --> 00:01:31.680 So basically, we can ask these tools to write code snippets 31 00:01:31.680 --> 00:01:33.330 or a function for us, 32 00:01:33.330 --> 00:01:35.100 and some tools like Cursor 33 00:01:35.100 --> 00:01:39.600 can even generate entire files based on our prompts. 34 00:01:39.600 --> 00:01:41.130 Copilot and Cursor 35 00:01:41.130 --> 00:01:44.520 also give us a very powerful auto-complete feature 36 00:01:44.520 --> 00:01:47.280 that is aware of the code that you're writing, 37 00:01:47.280 --> 00:01:49.620 your comments, your clipboard, 38 00:01:49.620 --> 00:01:52.110 and also the entire code base. 39 00:01:52.110 --> 00:01:54.720 You can even chat with your code base, 40 00:01:54.720 --> 00:01:56.100 like asking questions 41 00:01:56.100 --> 00:01:58.830 about how certain parts of the app work, 42 00:01:58.830 --> 00:02:01.260 in case you don't understand something. 43 00:02:01.260 --> 00:02:04.230 And all of this works pretty well, sometimes, 44 00:02:04.230 --> 00:02:07.560 and not so well, in some other times, 45 00:02:07.560 --> 00:02:11.100 but more about potential problems later on. 46 00:02:11.100 --> 00:02:15.120 Now, when we choose to work with an AI tool in our code, 47 00:02:15.120 --> 00:02:16.590 or even just for learning 48 00:02:16.590 --> 00:02:19.020 or for writing some simple function, 49 00:02:19.020 --> 00:02:21.240 we need to follow a workflow, 50 00:02:21.240 --> 00:02:24.510 just like we follow a workflow when we're coding 51 00:02:24.510 --> 00:02:26.640 and solving problems ourselves, 52 00:02:26.640 --> 00:02:28.950 without an AI assistant. 53 00:02:28.950 --> 00:02:31.590 And we've already talked about that workflow 54 00:02:31.590 --> 00:02:34.920 for solving problems and the different steps involved 55 00:02:34.920 --> 00:02:37.770 earlier in this section, right? 56 00:02:37.770 --> 00:02:40.830 And so let's now adapt that workflow 57 00:02:40.830 --> 00:02:42.840 for the age of AI, 58 00:02:42.840 --> 00:02:47.250 in case you decide to adopt any of these technologies. 59 00:02:47.250 --> 00:02:50.070 So just like in our earlier workflow, 60 00:02:50.070 --> 00:02:54.210 it all starts with 100% understanding the problem 61 00:02:54.210 --> 00:02:56.160 that we need to solve. 62 00:02:56.160 --> 00:02:57.810 Don't use an AI tool 63 00:02:57.810 --> 00:03:01.350 if you don't even know exactly what you need to solve, 64 00:03:01.350 --> 00:03:05.370 because this will just introduce unwanted code. 65 00:03:05.370 --> 00:03:08.610 You can also apply the divide and conquer step 66 00:03:08.610 --> 00:03:10.170 that we discussed earlier, 67 00:03:10.170 --> 00:03:12.150 to basically divide the problem 68 00:03:12.150 --> 00:03:14.130 into smaller sub-problems, 69 00:03:14.130 --> 00:03:15.930 if it's a really big one, 70 00:03:15.930 --> 00:03:18.007 like you're not gonna tell the AI, 71 00:03:18.007 --> 00:03:22.140 "Just write me a social network like Instagram, right?" 72 00:03:22.140 --> 00:03:25.980 That's way too big, and outright impossible. 73 00:03:25.980 --> 00:03:28.050 You would have to divide that one 74 00:03:28.050 --> 00:03:31.290 into thousands of small sub-problems. 75 00:03:31.290 --> 00:03:35.460 But anyway, next up, you're gonna choose which AI tool 76 00:03:35.460 --> 00:03:36.630 you want to use, 77 00:03:36.630 --> 00:03:40.080 and then you need to write a very specific prompt, 78 00:03:40.080 --> 00:03:43.410 telling it exactly what you want to get done. 79 00:03:43.410 --> 00:03:46.500 For this, it's important to feed the AI 80 00:03:46.500 --> 00:03:48.840 the exact context of the problem, 81 00:03:48.840 --> 00:03:51.000 like what topic it's about, 82 00:03:51.000 --> 00:03:53.310 the exact goal that should be achieved, 83 00:03:53.310 --> 00:03:56.250 what language it should write the solution in, 84 00:03:56.250 --> 00:04:00.000 what coding style you want, and so on. 85 00:04:00.000 --> 00:04:02.550 Then, based on this specific prompt, 86 00:04:02.550 --> 00:04:04.590 the AI will do its magic 87 00:04:04.590 --> 00:04:07.620 and generate a solution to your problem, 88 00:04:07.620 --> 00:04:10.920 and output that solution in the form of code, 89 00:04:10.920 --> 00:04:13.890 and in the language of your choice. 90 00:04:13.890 --> 00:04:17.010 So in our case, that's of course gonna be JavaScript, 91 00:04:17.010 --> 00:04:19.800 but it could also be HTML, CSS, 92 00:04:19.800 --> 00:04:23.250 Python, Java, and so one, 93 00:04:23.250 --> 00:04:26.430 and now here comes the most important part, 94 00:04:26.430 --> 00:04:28.320 where you as the developer 95 00:04:28.320 --> 00:04:30.330 need to very thoroughly 96 00:04:30.330 --> 00:04:33.900 inspect and test the generated code. 97 00:04:33.900 --> 00:04:36.810 Again, this is extremely important 98 00:04:36.810 --> 00:04:39.450 because code generated by AI 99 00:04:39.450 --> 00:04:41.790 can contain lots of bugs, 100 00:04:41.790 --> 00:04:44.490 which are sometimes, very well hidden, 101 00:04:44.490 --> 00:04:46.620 or it can also suggest a solution 102 00:04:46.620 --> 00:04:51.060 that's more complicated or convoluted than necessary. 103 00:04:51.060 --> 00:04:53.400 More about that in the next slide. 104 00:04:53.400 --> 00:04:55.410 But you can already start to see 105 00:04:55.410 --> 00:04:58.800 that these tools should only be used by developers 106 00:04:58.800 --> 00:05:02.670 who already know what they are doing, right? 107 00:05:02.670 --> 00:05:06.360 But anyway, after reviewing and testing the code, 108 00:05:06.360 --> 00:05:10.800 you can use the AI to correct or improve the solution, 109 00:05:10.800 --> 00:05:13.950 by generating a new, very specific prompt 110 00:05:13.950 --> 00:05:16.710 telling the AI assistant what's wrong 111 00:05:16.710 --> 00:05:18.660 and what you need fixed. 112 00:05:18.660 --> 00:05:21.270 So this is a pretty iterative process 113 00:05:21.270 --> 00:05:22.980 until you reach a solution 114 00:05:22.980 --> 00:05:24.840 that you're happy with. 115 00:05:24.840 --> 00:05:27.660 Sometimes, this actually never happens, 116 00:05:27.660 --> 00:05:30.300 in which case, you need to abandon the idea 117 00:05:30.300 --> 00:05:33.900 of using AI for this specific case. 118 00:05:33.900 --> 00:05:36.540 But if you are satisfied with the answer 119 00:05:36.540 --> 00:05:39.750 and are 100% sure that it's correct, 120 00:05:39.750 --> 00:05:43.803 then you can integrate the solution into your code base. 121 00:05:44.820 --> 00:05:47.640 So this is how you would use an AI assistant 122 00:05:47.640 --> 00:05:49.800 to help you write some code, 123 00:05:49.800 --> 00:05:53.760 but when exactly should you actually use a tool like this 124 00:05:53.760 --> 00:05:56.910 and the code that it generates? 125 00:05:56.910 --> 00:06:00.240 Well, before you choose to use an AI assistant 126 00:06:00.240 --> 00:06:01.950 to solve any task, 127 00:06:01.950 --> 00:06:03.480 you first need to make sure 128 00:06:03.480 --> 00:06:04.890 that you actually know 129 00:06:04.890 --> 00:06:07.950 how to code yourself on your own. 130 00:06:07.950 --> 00:06:10.080 So fundamental programming skills 131 00:06:10.080 --> 00:06:12.270 are absolutely essential, 132 00:06:12.270 --> 00:06:15.750 and so that's what this course is for. 133 00:06:15.750 --> 00:06:18.420 Now, I'm giving you this lecture early on 134 00:06:18.420 --> 00:06:22.470 because there are so many questions about AI these days, 135 00:06:22.470 --> 00:06:24.360 and so I wanted to highlight the need 136 00:06:24.360 --> 00:06:27.150 for learning programming right off the gate, 137 00:06:27.150 --> 00:06:29.220 instead of relying on AI 138 00:06:29.220 --> 00:06:32.460 before you even know how to program. 139 00:06:32.460 --> 00:06:33.930 Now, in the same vein, 140 00:06:33.930 --> 00:06:37.800 you also need to be able to solve problems on your own 141 00:06:37.800 --> 00:06:41.250 and develop very good critical thinking. 142 00:06:41.250 --> 00:06:44.010 Critical thinking is especially important 143 00:06:44.010 --> 00:06:46.710 when you're dealing with AI-generated code, 144 00:06:46.710 --> 00:06:47.940 because many times, 145 00:06:47.940 --> 00:06:50.550 it contains bugs that are hard to find, 146 00:06:50.550 --> 00:06:53.610 or the code is simply not good enough. 147 00:06:53.610 --> 00:06:56.340 Always question whether the AI output 148 00:06:56.340 --> 00:06:58.590 actually makes sense. 149 00:06:58.590 --> 00:07:03.060 Finally, coding should sparkle joy and curiosity. 150 00:07:03.060 --> 00:07:06.630 If coding is boring or not exciting anyway, 151 00:07:06.630 --> 00:07:09.210 if you lack the passion for writing code, 152 00:07:09.210 --> 00:07:12.930 then maybe this isn't the right career anyway. 153 00:07:12.930 --> 00:07:17.070 So in summary, I would advise to not just use AI 154 00:07:17.070 --> 00:07:19.500 without knowing what you're doing, 155 00:07:19.500 --> 00:07:23.370 otherwise, AI will turn you into a terrible programmer 156 00:07:23.370 --> 00:07:26.610 or maybe, not even a programmer at all. 157 00:07:26.610 --> 00:07:30.180 If you rely on AI-generated code all the time, 158 00:07:30.180 --> 00:07:32.250 you're gonna grow dependent on it, 159 00:07:32.250 --> 00:07:34.020 and you'll just become someone 160 00:07:34.020 --> 00:07:36.690 who copy=pastes code left and right, 161 00:07:36.690 --> 00:07:39.690 until you have some barely working app, 162 00:07:39.690 --> 00:07:41.910 an app that you don't understand, 163 00:07:41.910 --> 00:07:44.370 can't maintain, and can't fix, 164 00:07:44.370 --> 00:07:46.020 in case anything breaks, 165 00:07:46.020 --> 00:07:49.410 which let me assure you, will break. 166 00:07:49.410 --> 00:07:54.090 So someone like this is not a programmer in my mind. 167 00:07:54.090 --> 00:07:58.020 Now, some people will go ahead and do just this, 168 00:07:58.020 --> 00:07:59.940 so they will just copy-paste, 169 00:07:59.940 --> 00:08:02.460 and they might have some success with it 170 00:08:02.460 --> 00:08:04.170 to a certain degree, 171 00:08:04.170 --> 00:08:06.330 but if you want to do things well, 172 00:08:06.330 --> 00:08:09.240 again, my advice is to learn the fundamentals 173 00:08:09.240 --> 00:08:10.290 from this course 174 00:08:10.290 --> 00:08:13.410 and become a real programmer instead. 175 00:08:13.410 --> 00:08:15.930 Now, this may all sound a bit negative, 176 00:08:15.930 --> 00:08:20.520 but, of course, it is fine to use AI in certain situations, 177 00:08:20.520 --> 00:08:22.770 I do it myself all the time. 178 00:08:22.770 --> 00:08:26.520 But instead of relying on AI to do everything, 179 00:08:26.520 --> 00:08:29.250 you should see AI as an assistant. 180 00:08:29.250 --> 00:08:33.420 AI is meant to help you, not to replace you. 181 00:08:33.420 --> 00:08:36.570 So AI tools can and should be used 182 00:08:36.570 --> 00:08:39.000 to help you save time on repetitive 183 00:08:39.000 --> 00:08:42.390 or mundane tasks like boilerplate code, 184 00:08:42.390 --> 00:08:45.903 so things that don't require you to think much anyway. 185 00:08:46.770 --> 00:08:50.340 AI assistants are also great for learning, 186 00:08:50.340 --> 00:08:52.920 for example, you can just ask it questions 187 00:08:52.920 --> 00:08:54.870 about programming topics, 188 00:08:54.870 --> 00:08:57.270 or you can ask for code solutions, 189 00:08:57.270 --> 00:09:00.030 and then analyze the output step by step 190 00:09:00.030 --> 00:09:02.520 in order to learn something new. 191 00:09:02.520 --> 00:09:05.820 This way, you're gonna find new tools and techniques 192 00:09:05.820 --> 00:09:07.020 that you can learn, 193 00:09:07.020 --> 00:09:09.840 that maybe otherwise you would've missed, 194 00:09:09.840 --> 00:09:12.750 so you wouldn't have known they even existed, 195 00:09:12.750 --> 00:09:16.293 and will actually do all this in the next lecture. 196 00:09:17.250 --> 00:09:21.960 Now, okay, now assuming that you did decide to use AI, 197 00:09:21.960 --> 00:09:25.110 and you now have your AI-generated code, 198 00:09:25.110 --> 00:09:26.790 when is it actually okay 199 00:09:26.790 --> 00:09:30.360 to incorporate this code into your app? 200 00:09:30.360 --> 00:09:34.290 Well, I'm gonna start by reiterating on what I said earlier 201 00:09:34.290 --> 00:09:36.870 because it's just so important. 202 00:09:36.870 --> 00:09:39.330 You should only incorporate AI code 203 00:09:39.330 --> 00:09:42.870 if you could have written that code yourself. 204 00:09:42.870 --> 00:09:44.760 The ideal scenario is that 205 00:09:44.760 --> 00:09:47.880 before asking the AI to write the code for you, 206 00:09:47.880 --> 00:09:50.130 you have already thought of the steps 207 00:09:50.130 --> 00:09:52.830 and the JavaScript features that you might need, 208 00:09:52.830 --> 00:09:54.720 in order to solve the problem, 209 00:09:54.720 --> 00:09:58.170 and maybe even imagined the code in your mind. 210 00:09:58.170 --> 00:09:59.940 Then all the AI does 211 00:09:59.940 --> 00:10:03.090 is to actually write the code you had already imagined, 212 00:10:03.090 --> 00:10:04.740 which is gonna be a lot faster 213 00:10:04.740 --> 00:10:07.530 than you typing it out by hand. 214 00:10:07.530 --> 00:10:11.550 And this actually happens to me many times when I use AI, 215 00:10:11.550 --> 00:10:14.790 and it feels pretty magical, to be honest. 216 00:10:14.790 --> 00:10:17.190 Now, if you could have written the code yourself, 217 00:10:17.190 --> 00:10:19.650 the next guideline is pretty obvious, 218 00:10:19.650 --> 00:10:21.900 but it's still important to highlight the fact 219 00:10:21.900 --> 00:10:25.530 that you should truly understand the generated code 220 00:10:25.530 --> 00:10:29.310 before incorporating it into your code base. 221 00:10:29.310 --> 00:10:31.800 Also, as I already touched on earlier, 222 00:10:31.800 --> 00:10:36.120 you need to ensure that the code is 100% correct. 223 00:10:36.120 --> 00:10:39.600 All this will ensure that errors made by AI, 224 00:10:39.600 --> 00:10:42.420 don't make it into your code base. 225 00:10:42.420 --> 00:10:44.220 This is especially important 226 00:10:44.220 --> 00:10:47.220 for mission-critical parts of your app, 227 00:10:47.220 --> 00:10:49.200 like core business logic, 228 00:10:49.200 --> 00:10:52.320 and so I advise you to not even use AI 229 00:10:52.320 --> 00:10:55.800 for these crucial tasks at all. 230 00:10:55.800 --> 00:10:57.270 So to recap, 231 00:10:57.270 --> 00:11:00.720 it's absolutely okay to incorporate some code 232 00:11:00.720 --> 00:11:02.850 that an AI has written for you, 233 00:11:02.850 --> 00:11:04.710 if you're only using the AI 234 00:11:04.710 --> 00:11:07.890 to be faster at non-crucial tasks, 235 00:11:07.890 --> 00:11:11.010 even though, you could have done it yourself 236 00:11:11.010 --> 00:11:13.650 otherwise, you'll never be able to fix the code 237 00:11:13.650 --> 00:11:15.270 when it eventually breaks, 238 00:11:15.270 --> 00:11:17.610 and maybe even more importantly, 239 00:11:17.610 --> 00:11:20.670 you won't grow as a developer. 240 00:11:20.670 --> 00:11:25.530 But now it's time to finally tackle these big, big questions 241 00:11:25.530 --> 00:11:27.210 that many beginners have, 242 00:11:27.210 --> 00:11:28.687 which go something like, 243 00:11:28.687 --> 00:11:32.250 "Do I even need to grow as a developer now? 244 00:11:32.250 --> 00:11:35.580 Won't AI take my job anyway at some point? 245 00:11:35.580 --> 00:11:39.480 Or, should I even bother learning to code now?" 246 00:11:39.480 --> 00:11:42.570 Well, answering this type of question 247 00:11:42.570 --> 00:11:46.260 is one of the main goals of this lecture. 248 00:11:46.260 --> 00:11:48.787 So the short answer to the question, 249 00:11:48.787 --> 00:11:50.910 "Will AI take my job?" 250 00:11:50.910 --> 00:11:52.800 Is no. 251 00:11:52.800 --> 00:11:55.380 So case closed, right? 252 00:11:55.380 --> 00:11:58.110 Well, maybe not so fast. 253 00:11:58.110 --> 00:12:01.500 Maybe you want to understand why I'm so confident 254 00:12:01.500 --> 00:12:05.520 that developers will still be needed going forward. 255 00:12:05.520 --> 00:12:08.040 Now, I need to preface this by saying, 256 00:12:08.040 --> 00:12:10.920 that this is of course, my personal opinion, 257 00:12:10.920 --> 00:12:14.820 and that I shockingly cannot see the future, 258 00:12:14.820 --> 00:12:16.950 but I've thought about this a lot 259 00:12:16.950 --> 00:12:19.320 and here's what I think. 260 00:12:19.320 --> 00:12:22.470 So the first thing that we should address 261 00:12:22.470 --> 00:12:26.550 is that there is just a lot of hype around AI these days, 262 00:12:26.550 --> 00:12:29.010 and that's true for all industries, 263 00:12:29.010 --> 00:12:31.230 not just web development. 264 00:12:31.230 --> 00:12:32.880 Many people seem to believe 265 00:12:32.880 --> 00:12:35.460 that AI is gonna replace us soon, 266 00:12:35.460 --> 00:12:39.240 and there are actually companies trying to do just that, 267 00:12:39.240 --> 00:12:40.770 but they seem to forget 268 00:12:40.770 --> 00:12:43.290 that we've been in this situation before. 269 00:12:43.290 --> 00:12:47.040 We've had tools like Dreamweaver, or WordPress, 270 00:12:47.040 --> 00:12:49.980 or site builders like Wix in the past, 271 00:12:49.980 --> 00:12:53.880 which were supposed to make web developers obsolete. 272 00:12:53.880 --> 00:12:56.130 But did that ever happen? 273 00:12:56.130 --> 00:12:57.720 No, it didn't, 274 00:12:57.720 --> 00:13:01.440 and I'm sure that it won't happen this time around. 275 00:13:01.440 --> 00:13:05.070 Now, I do believe that AI is here to stay 276 00:13:05.070 --> 00:13:08.460 and that it will be very helpful for us developers, 277 00:13:08.460 --> 00:13:11.970 but it's clearly over-hyped at the moment. 278 00:13:11.970 --> 00:13:14.520 But this brings me to the next point, 279 00:13:14.520 --> 00:13:19.020 which is that, indeed, things will change in the future. 280 00:13:19.020 --> 00:13:21.870 No one knows how much and how fast, 281 00:13:21.870 --> 00:13:23.460 but we can say for sure 282 00:13:23.460 --> 00:13:25.830 that some portions of applications 283 00:13:25.830 --> 00:13:28.800 will be written by AI in the future. 284 00:13:28.800 --> 00:13:31.290 This is already happening right now, 285 00:13:31.290 --> 00:13:34.590 and this tendency is only gonna increase. 286 00:13:34.590 --> 00:13:38.250 But the thing is that this isn't necessarily bad, 287 00:13:38.250 --> 00:13:41.370 as we saw before, AI code can be great 288 00:13:41.370 --> 00:13:44.190 for automating boring boilerplate stuff 289 00:13:44.190 --> 00:13:45.990 that takes up a lot of time, 290 00:13:45.990 --> 00:13:49.860 or other mundane tasks in software development. 291 00:13:49.860 --> 00:13:53.340 This means that for all the other interesting tasks, 292 00:13:53.340 --> 00:13:55.590 you will still be needed. 293 00:13:55.590 --> 00:13:57.030 You the developer, 294 00:13:57.030 --> 00:14:00.240 will always stay in control of implementing 295 00:14:00.240 --> 00:14:03.840 and of adapting the applications architecture, 296 00:14:03.840 --> 00:14:06.570 and of writing the core business logic, 297 00:14:06.570 --> 00:14:09.630 at least, that's how it should work. 298 00:14:09.630 --> 00:14:11.490 You're also the one who asks 299 00:14:11.490 --> 00:14:14.010 the AI for help in the first place, 300 00:14:14.010 --> 00:14:17.700 and you are the one who is gonna review the code, test it, 301 00:14:17.700 --> 00:14:20.880 and finally, integrated it into the app. 302 00:14:20.880 --> 00:14:25.020 So again, the developers are gonna stay in control 303 00:14:25.020 --> 00:14:26.670 and can actively decide 304 00:14:26.670 --> 00:14:31.230 how much they want to involve the AI in their projects, 305 00:14:31.230 --> 00:14:32.940 because the truth is, 306 00:14:32.940 --> 00:14:37.940 software developers do a lot more than just writing code. 307 00:14:38.010 --> 00:14:40.290 They're able to maintain an overview 308 00:14:40.290 --> 00:14:43.530 and see the bigger picture of complex projects, 309 00:14:43.530 --> 00:14:47.253 and they understand and implement design principles. 310 00:14:48.210 --> 00:14:51.660 Now, I guess all this is a bit hard for you to imagine 311 00:14:51.660 --> 00:14:53.340 as you're just starting out 312 00:14:53.340 --> 00:14:54.960 on this programming journey, 313 00:14:54.960 --> 00:14:56.550 but there are just things 314 00:14:56.550 --> 00:14:59.490 that are too complex for AI to handle, 315 00:14:59.490 --> 00:15:02.610 where it's not just writing a few lines of code, 316 00:15:02.610 --> 00:15:06.990 but really thinking hard about problems and solutions, 317 00:15:06.990 --> 00:15:11.310 and so that's where the human programmers come into play. 318 00:15:11.310 --> 00:15:15.210 These humans are also way more creative than AI, 319 00:15:15.210 --> 00:15:17.400 and they're able to actually collaborate 320 00:15:17.400 --> 00:15:18.840 with other developers, 321 00:15:18.840 --> 00:15:21.360 and maybe, even more importantly, 322 00:15:21.360 --> 00:15:23.640 collaborate with the clients. 323 00:15:23.640 --> 00:15:28.170 So it won't be easy to automate all that stuff away with AI, 324 00:15:28.170 --> 00:15:30.510 and it won't happen soon. 325 00:15:30.510 --> 00:15:31.710 On top of this, 326 00:15:31.710 --> 00:15:33.780 as I also already hinted at, 327 00:15:33.780 --> 00:15:36.660 from my experience, AI generated-code 328 00:15:36.660 --> 00:15:38.430 is still pretty buggy, 329 00:15:38.430 --> 00:15:40.740 and often, doesn't give you the best 330 00:15:40.740 --> 00:15:43.200 or most up-to-date solutions. 331 00:15:43.200 --> 00:15:47.610 It's especially bad at debugging non-trivial problems 332 00:15:47.610 --> 00:15:50.250 that span multiple functions or files, 333 00:15:50.250 --> 00:15:53.760 hallucinating solutions that don't solve anything 334 00:15:53.760 --> 00:15:56.640 and just make the code worse and worse 335 00:15:56.640 --> 00:16:00.210 each time you ask it to improve the solution. 336 00:16:00.210 --> 00:16:02.370 I've seen this happening a few times, 337 00:16:02.370 --> 00:16:05.190 where the most state-of-the-art AI models 338 00:16:05.190 --> 00:16:07.800 become completely useless. 339 00:16:07.800 --> 00:16:10.530 So yeah, I think we're pretty safe, 340 00:16:10.530 --> 00:16:13.650 especially when it comes to debugging. 341 00:16:13.650 --> 00:16:16.410 Another aspect that's interesting to consider 342 00:16:16.410 --> 00:16:18.630 is the fact that, right now, 343 00:16:18.630 --> 00:16:21.630 the world actually needs more software 344 00:16:21.630 --> 00:16:25.170 than can be built by us humans right now. 345 00:16:25.170 --> 00:16:29.220 So how many times have you used some terrible website 346 00:16:29.220 --> 00:16:30.930 from a local business, 347 00:16:30.930 --> 00:16:33.600 or used a non-functioning web app 348 00:16:33.600 --> 00:16:37.290 to make a restaurant reservation or something like that? 349 00:16:37.290 --> 00:16:38.520 I know I have, 350 00:16:38.520 --> 00:16:42.960 and if we recruit AI to help us automating the boring parts, 351 00:16:42.960 --> 00:16:45.900 maybe we can improve much of the bad, 352 00:16:45.900 --> 00:16:48.150 or even non-existing software, 353 00:16:48.150 --> 00:16:50.520 which is gonna benefit everyone. 354 00:16:50.520 --> 00:16:54.840 So maybe if AI makes us like 10 times faster 355 00:16:54.840 --> 00:16:57.240 at some distant point in the future, 356 00:16:57.240 --> 00:17:01.590 it might enable us to write 10 times more software. 357 00:17:01.590 --> 00:17:04.770 So those are the kinds of positive effects 358 00:17:04.770 --> 00:17:07.740 that many people don't even think about, 359 00:17:07.740 --> 00:17:10.950 and there are probably many more positive effects 360 00:17:10.950 --> 00:17:14.100 that we can't even imagine yet. 361 00:17:14.100 --> 00:17:15.810 So in conclusion, 362 00:17:15.810 --> 00:17:19.410 AI will help us, not replace us. 363 00:17:19.410 --> 00:17:22.530 It might take some boring coding tasks away, 364 00:17:22.530 --> 00:17:26.070 allowing us to focus on what really matters, 365 00:17:26.070 --> 00:17:28.170 building amazing software, 366 00:17:28.170 --> 00:17:31.830 and building a lot more software faster. 367 00:17:31.830 --> 00:17:34.890 This means, there won't be fewer jobs in the future, 368 00:17:34.890 --> 00:17:38.370 but actually, maybe even more. 369 00:17:38.370 --> 00:17:42.330 So please keep learning JavaScript or React, 370 00:17:42.330 --> 00:17:45.120 or whatever it is you're doing right now, 371 00:17:45.120 --> 00:17:48.030 and don't stop this whole AI narrative 372 00:17:48.030 --> 00:17:49.860 from progressing on your journey 373 00:17:49.860 --> 00:17:51.720 of learning web development 374 00:17:51.720 --> 00:17:54.480 and building amazing software. 375 00:17:54.480 --> 00:17:58.350 Plus, learning programming teaches you foundational skills 376 00:17:58.350 --> 00:18:00.900 like logic and reasoning. 377 00:18:00.900 --> 00:18:03.960 Basically, it teaches you how to think, 378 00:18:03.960 --> 00:18:07.050 and those skills will always be relevant, 379 00:18:07.050 --> 00:18:10.530 and maybe even more so in the future. 380 00:18:10.530 --> 00:18:12.090 And with this being said, 381 00:18:12.090 --> 00:18:15.180 let's now actually try our hands at using AI 382 00:18:15.180 --> 00:18:16.860 to solve a challenge, 383 00:18:16.860 --> 00:18:19.200 following the process and the guidelines 384 00:18:19.200 --> 00:18:21.813 that we just talked about in this video.