WEBVTT 00:00.240 --> 00:05.400 Okay, so first of all, we bring up a new terminal, which we do by pressing the control button and 00:05.400 --> 00:06.280 hitting the tick. 00:06.680 --> 00:08.240 And when we do that, up it comes. 00:08.240 --> 00:11.840 You can also get there by going to the view menu and choosing terminal. 00:12.040 --> 00:17.280 So we need to CD into the two OpenAI directory. 00:17.520 --> 00:22.560 And now we need to change again into the Deep Research directory because that's where our stuff is. 00:22.600 --> 00:25.960 And you might be thinking we now type Python deepresearch dot pi. 00:26.000 --> 00:28.760 But that would not be right because we're working with UV. 00:28.960 --> 00:35.160 So you do UV and then run and deepresearch pi because that will make sure that we're in our correct 00:35.160 --> 00:36.480 virtual environment. 00:36.640 --> 00:38.160 And that is what we do now. 00:38.160 --> 00:40.720 And it's then going to pop up a screen like this. 00:40.720 --> 00:42.200 This is our deep research screen. 00:42.480 --> 00:44.280 This is our Gradio UI. 00:44.400 --> 00:46.320 And make it a tiny bit bigger. 00:47.040 --> 00:47.520 All right. 00:47.560 --> 00:49.880 What topic would you like to research. 00:49.880 --> 00:50.680 So how about this. 00:50.680 --> 01:03.160 How about we say, uh, what are some of the most exciting commercial applications of autonomous agentic 01:03.240 --> 01:09.840 AI in as of April 2025? 01:10.080 --> 01:10.520 All right. 01:10.560 --> 01:11.920 That seems like a good question. 01:12.160 --> 01:12.560 Why not? 01:12.620 --> 01:13.580 Let's give it a whirl. 01:14.300 --> 01:15.020 Run! 01:15.380 --> 01:17.500 So first of all, we get this popping up. 01:17.500 --> 01:18.300 View the trace. 01:18.300 --> 01:22.820 If we click there, it comes up and shows us it's just beginning. 01:22.860 --> 01:24.700 We're still at the very beginning stages. 01:24.980 --> 01:28.540 We'll hopefully in a moment see a bit more of what's going on. 01:28.580 --> 01:29.380 Not not yet. 01:29.380 --> 01:31.460 It's still in the early stages. 01:32.100 --> 01:38.500 What we're seeing here is an update appearing for what's going on, and that is because we are yielding 01:38.500 --> 01:39.260 these updates. 01:39.260 --> 01:46.500 And every time that we yield those updates in our run, coroutine that gets surfaced through to to our 01:46.540 --> 01:48.580 callback and through to Gradio. 01:48.580 --> 01:51.740 And that's why it's appearing here because of those yields. 01:51.780 --> 01:52.740 That's why we did it. 01:52.740 --> 01:55.020 Otherwise we'd be sitting here wondering what's going on. 01:55.340 --> 02:00.060 So it's already done the 20 searches, because they all happened in parallel, as you know, thanks 02:00.060 --> 02:02.020 to the magic of async IO. 02:02.220 --> 02:03.740 And now it is writing. 02:03.740 --> 02:07.460 It's written the report, it's now sending the email. 02:07.700 --> 02:14.820 And hopefully any moment now it will complete which will be nice and we will get to see. 02:14.820 --> 02:15.540 And here it is. 02:15.580 --> 02:20.900 And we don't need to go to our email for a change because it's all appearing here in the Gradio app 02:21.140 --> 02:24.880 and it's appearing in beautiful markdown format. 02:24.920 --> 02:33.360 It's got a nice introduction overview of autonomous Agentic AI cybersecurity, using it for threat detection, 02:33.400 --> 02:40.120 reduce workloads, customer service, obviously healthcare, automotive and transport financial services, 02:40.240 --> 02:43.920 analyzing market trends and executing trades based on real time information. 02:43.960 --> 02:46.120 Maybe we're going to be doing some of that ourselves. 02:46.400 --> 02:49.040 Logistics and supply chain okay. 02:49.080 --> 02:49.960 And ethical. 02:49.960 --> 02:50.400 Regulatory. 02:50.640 --> 02:56.320 So so now it's talking about about uh, some of the uh, sort of side effects or consequences and some 02:56.320 --> 02:58.600 conclusions and follow up questions. 02:58.760 --> 03:04.280 So there you have it, the result of our deep research right here for us to see. 03:05.080 --> 03:07.480 Let's look at the traces and see how this is going now. 03:07.800 --> 03:09.760 This is the research trace. 03:09.960 --> 03:13.120 And sure enough there of course are our agents. 03:13.320 --> 03:19.720 And at the end is our writer agent and our email agent. 03:19.720 --> 03:21.320 It successfully ran. 03:21.480 --> 03:27.920 And this is now a tool that you can use and that you can refer to and have as a deep research tool. 03:28.240 --> 03:35.520 And there's still one more thing to talk about, which is how we can now extend this and make it more 03:35.520 --> 03:36.280 powerful.