WEBVTT 00:00:00.000 --> 00:00:02.899 Okay, think about your web browser. The tool 00:00:02.899 --> 00:00:05.259 you use for, well, almost everything online, 00:00:05.400 --> 00:00:08.019 it's often, maybe too often, a real productivity 00:00:08.019 --> 00:00:10.220 trap, right? You open one tab, you're searching 00:00:10.220 --> 00:00:12.439 for just one little thing, and suddenly there 00:00:12.439 --> 00:00:14.939 are five more. Then you're copying, pasting, 00:00:15.080 --> 00:00:18.559 managing maybe a dozen windows, and hours just 00:00:18.559 --> 00:00:21.539 vanish into what we call administrative friction. 00:00:21.760 --> 00:00:23.800 It's that constant drag, isn't it? That friction 00:00:23.800 --> 00:00:26.120 of just getting digital work done. But what if 00:00:26.120 --> 00:00:29.940 that very tool your browser could actually be 00:00:29.940 --> 00:00:32.920 like an intelligent assistant, one that handles 00:00:32.920 --> 00:00:35.700 the thinking part, the synthesis, the admin burden 00:00:35.700 --> 00:00:38.259 for you. That's really the core question we're 00:00:38.259 --> 00:00:40.759 digging into today in this deep dive. We're unpacking 00:00:40.759 --> 00:00:42.899 the ideas behind the Perplexity Comet browser. 00:00:43.299 --> 00:00:45.280 And our mission, really, is to show you how this 00:00:45.280 --> 00:00:47.420 tool kind of forces a necessary shift in how 00:00:47.420 --> 00:00:49.520 you work, moving you away from being just an 00:00:49.520 --> 00:00:51.479 information hunter, you know, manually clicking 00:00:51.479 --> 00:00:53.880 through links, to becoming more of a task manager, 00:00:54.219 --> 00:00:56.530 someone who delegates complex research. Yeah, 00:00:56.729 --> 00:00:59.840 exactly. We'll map out why this shift is... Frankly 00:00:59.840 --> 00:01:02.539 essential now, with generative AI everywhere, 00:01:03.079 --> 00:01:05.420 we need to define that critical difference between 00:01:05.420 --> 00:01:08.060 a standard search engine and what they're calling 00:01:08.060 --> 00:01:11.420 an answer engine. Then we'll walk through the, 00:01:11.420 --> 00:01:13.719 let's say, expert setup steps, group together 00:01:13.719 --> 00:01:16.879 logically, and finally demonstrate six really 00:01:16.879 --> 00:01:19.620 concrete workflows that can genuinely cut down 00:01:19.620 --> 00:01:22.980 on that cognitive load. OK, sounds good. Let's 00:01:22.980 --> 00:01:24.480 unpack this. We kind of have to start with the 00:01:24.480 --> 00:01:26.280 context, which involves a bit of fear, actually. 00:01:26.540 --> 00:01:28.659 The source material we looked at points to this 00:01:28.659 --> 00:01:32.019 pretty striking number. 71 % of people surveyed 00:01:32.019 --> 00:01:35.079 worry AI will take their jobs. That level of 00:01:35.079 --> 00:01:38.239 job anxiety looking at new tech can be paralyzing. 00:01:38.540 --> 00:01:41.000 It really can. But that whole narrative, it needs 00:01:41.000 --> 00:01:43.359 an immediate reframe. The sources are quite clear 00:01:43.359 --> 00:01:45.680 on this. AI isn't just replacing people wholesale. 00:01:45.920 --> 00:01:48.180 Instead, it's the people who get good at using 00:01:48.180 --> 00:01:50.700 these new AI tools quickly. They are the ones 00:01:50.700 --> 00:01:53.879 who will replace those who, well. refuse to adapt. 00:01:54.040 --> 00:01:57.200 That distinction is vital. Hmm. Okay. But that 00:01:57.200 --> 00:01:59.040 sounds like it might be a technical burden. Does 00:01:59.040 --> 00:02:00.859 mastering a tool like Comet mean I need to learn 00:02:00.859 --> 00:02:03.359 programming or something? Not at all. Now, think 00:02:03.359 --> 00:02:05.359 back to when cars first showed up. The people 00:02:05.359 --> 00:02:07.939 who learned to drive cars, they found new jobs, 00:02:08.139 --> 00:02:10.520 high -leverage jobs, and the folks who insisted, 00:02:10.680 --> 00:02:13.659 no, I'm sticking with my horse. Well... They 00:02:13.659 --> 00:02:16.379 got left behind, didn't they? These AI tools 00:02:16.379 --> 00:02:18.740 are basically the car in that story. It's about 00:02:18.740 --> 00:02:21.379 upgrading your skills, adapting to modern tools. 00:02:21.479 --> 00:02:24.099 It's not about coding. Right. So if I learn how 00:02:24.099 --> 00:02:27.219 to delegate tasks effectively to this tool, this 00:02:27.219 --> 00:02:30.939 assistant, how does that change my value, you 00:02:30.939 --> 00:02:33.300 know, in the job market? It signals adaptability. 00:02:33.620 --> 00:02:35.800 It shifts you from someone who could be replaced 00:02:35.800 --> 00:02:38.419 to someone essential. Now, to really get Comet, 00:02:38.569 --> 00:02:40.710 we have to understand what it is, because it 00:02:40.710 --> 00:02:42.789 works differently than the tools we've used for, 00:02:42.789 --> 00:02:45.610 gosh, decades. We need to define the difference 00:02:45.610 --> 00:02:48.030 between a regular search engine like Google and 00:02:48.030 --> 00:02:50.889 this idea of an answer engine. OK. The chocolate 00:02:50.889 --> 00:02:52.830 cake recipe analogy you mentioned earlier makes 00:02:52.830 --> 00:02:55.050 this super clear, I think. When you ask Google 00:02:55.050 --> 00:02:57.389 the, let's call it the traditional search tool 00:02:57.389 --> 00:03:00.469 for the best chocolate cake recipe, Google acts 00:03:00.469 --> 00:03:03.699 like a librarian. Right. Precisely. A librarian 00:03:03.699 --> 00:03:07.039 who hands you a list. Maybe 10, maybe 100 websites 00:03:07.039 --> 00:03:09.919 that seem relevant. Then it's on you, the user, 00:03:10.060 --> 00:03:12.560 to click into them, read them, compare them, 00:03:12.819 --> 00:03:14.840 try to synthesize all that information yourself. 00:03:15.639 --> 00:03:19.800 That takes time, effort. It adds to that cognitive 00:03:19.800 --> 00:03:22.060 load. But when you ask an answer engine, like 00:03:22.060 --> 00:03:25.400 Comet, the same question, give me a simple chocolate 00:03:25.400 --> 00:03:28.139 cake recipe for a beginner. It acts more like 00:03:28.139 --> 00:03:30.159 a dedicated assistant. It reads a whole bunch 00:03:30.159 --> 00:03:32.400 of sources for you. It synthesizes them down 00:03:32.400 --> 00:03:35.460 into one clear, prioritized answer. Here are 00:03:35.460 --> 00:03:37.460 the ingredients, here are the steps, here are 00:03:37.460 --> 00:03:39.599 some tips. And what's absolutely crucial here 00:03:39.599 --> 00:03:41.300 is that it lists its sources. It tells you, look, 00:03:41.400 --> 00:03:43.460 this summary came from website A, website B, 00:03:43.539 --> 00:03:46.699 website C, trustworthy sites. So you skip all 00:03:46.699 --> 00:03:48.340 that tedious work of finding it yourself and 00:03:48.340 --> 00:03:50.759 you instantly get a verified, structured answer. 00:03:50.819 --> 00:03:52.800 It just cuts out a huge amount of manual effort. 00:03:53.240 --> 00:03:55.939 So beyond just saving time, what's the main advantage 00:03:55.939 --> 00:03:58.400 of getting that summarized answer, but one that 00:03:58.400 --> 00:04:00.800 also includes the verifiable sources. Well, it 00:04:00.800 --> 00:04:02.800 gives you immediate clarity, and you know you 00:04:02.800 --> 00:04:05.560 can trust the information. Okay, this next part 00:04:05.560 --> 00:04:08.800 is where the real work begins. The setup. Seems 00:04:08.800 --> 00:04:10.759 most people make the mistake of just installing 00:04:10.759 --> 00:04:13.520 Comet and jumping right in. The sources really 00:04:13.520 --> 00:04:16.600 emphasize this. Spend maybe 10 minutes setting 00:04:16.600 --> 00:04:18.660 it up correctly before you use it for anything 00:04:18.660 --> 00:04:20.839 serious. It'll save you energy down the line. 00:04:20.980 --> 00:04:23.339 Yeah, absolutely. We can group the key setup 00:04:23.339 --> 00:04:25.939 steps into maybe three main buckets. The first 00:04:25.939 --> 00:04:28.459 one is all about habit and defaulting. Step one, 00:04:28.839 --> 00:04:30.620 use the desktop app. Doesn't matter if you're 00:04:30.620 --> 00:04:33.459 on Windows, Mac, Linux, use the app. And this 00:04:33.459 --> 00:04:36.040 is important, you must make it your default browser. 00:04:37.160 --> 00:04:39.779 Why is making it the default so critical? It 00:04:39.779 --> 00:04:43.079 feels a little invasive, maybe? Forcing me to 00:04:43.079 --> 00:04:45.279 change a habit that's pretty ingrained. I get 00:04:45.279 --> 00:04:46.980 that, but it's about building that essential 00:04:46.980 --> 00:04:49.600 muscle memory. If this assistant is only there 00:04:49.600 --> 00:04:51.759 when you remember to open a separate app, you're 00:04:51.759 --> 00:04:53.779 just going to default back to your old inefficient 00:04:53.779 --> 00:04:56.540 browser whenever you're in a rush. Comet needs 00:04:56.540 --> 00:04:58.579 to be the path of least resistance for it to 00:04:58.579 --> 00:05:01.759 stick. Okay, the second bucket is AI customization. 00:05:02.339 --> 00:05:04.899 You need to pick the right tool for the specific 00:05:04.899 --> 00:05:07.560 job you're doing. Comet gives you three main 00:05:07.560 --> 00:05:11.560 ways to operate. There's search for just quick 00:05:11.560 --> 00:05:14.899 daily answers. There's research for those bigger 00:05:14.899 --> 00:05:17.079 multi -step investigations where you need a proper 00:05:17.079 --> 00:05:19.379 structured summary at the end. And then there's 00:05:19.379 --> 00:05:21.680 labs which is more for turning rough ideas into 00:05:21.680 --> 00:05:24.639 complex workable outputs like drafting something 00:05:24.639 --> 00:05:26.540 substantial. And customizing what it searches 00:05:26.540 --> 00:05:28.899 seems really powerful and maybe overlooked. You 00:05:28.899 --> 00:05:30.439 don't always need the whole internet. That's 00:05:30.439 --> 00:05:33.279 exactly right. You can tell it only look at academic 00:05:33.279 --> 00:05:36.199 papers for this or search social media and read 00:05:36.199 --> 00:05:38.800 it for real user opinions or focus on finance 00:05:38.800 --> 00:05:41.300 sites for market data. You can even link it to 00:05:41.300 --> 00:05:43.740 your own connected tools. Like if you use Notion, 00:05:43.920 --> 00:05:46.600 it can search your personal notes. The key is 00:05:46.600 --> 00:05:49.160 always choose that specialized source before 00:05:49.160 --> 00:05:51.480 you hit search. And we also get control over 00:05:51.480 --> 00:05:53.779 the actual AI model running underneath the brain 00:05:53.779 --> 00:05:56.759 behind it. Which models are available and why 00:05:56.759 --> 00:05:59.259 would I pick one over another? Yes, this is key 00:05:59.259 --> 00:06:03.490 too. You can choose OpenAI's GPT -4. Right now, 00:06:03.529 --> 00:06:05.850 that's probably best for creative writing, complex 00:06:05.850 --> 00:06:09.250 analysis, or you might go for Google's Gemini 00:06:09.250 --> 00:06:12.649 2 .5 Pro, which seems to excel at really complex, 00:06:13.050 --> 00:06:16.230 multi -layered research tasks. And finally, there's 00:06:16.230 --> 00:06:18.350 Claude from Anthropic, which is known for being 00:06:18.350 --> 00:06:21.069 really good at following very specific, detailed 00:06:21.069 --> 00:06:23.170 instructions and handling long -form analysis. 00:06:24.089 --> 00:06:26.629 All right, the third bucket, automation and delegation. 00:06:26.860 --> 00:06:28.620 You need to turn on the assistant sidebar and 00:06:28.620 --> 00:06:31.019 keep it visible. The sidebar can actually see 00:06:31.019 --> 00:06:32.720 the page you're currently looking at, and it 00:06:32.720 --> 00:06:34.319 can perform actions right there on that page 00:06:34.319 --> 00:06:36.420 for you. OK, I see the utility there, definitely. 00:06:36.600 --> 00:06:39.019 But connecting personal data, like my Gmail or 00:06:39.019 --> 00:06:41.519 my calendar, where do we draw the line on privacy? 00:06:41.600 --> 00:06:44.019 That feels a bit concerning. You have to be incredibly 00:06:44.019 --> 00:06:47.160 careful here, extremely judicious. Connect only 00:06:47.160 --> 00:06:49.639 what genuinely helps you with a specific returning 00:06:49.639 --> 00:06:55.259 task. Less is absolutely more, honestly. I still 00:06:55.259 --> 00:06:58.100 wrestle with prompt drift myself sometimes. Especially 00:06:58.100 --> 00:07:00.600 if I let these tools access way too much context, 00:07:00.759 --> 00:07:03.240 it just gets confused. So really, only connect 00:07:03.240 --> 00:07:05.000 accounts where you know you need specific task 00:07:05.000 --> 00:07:08.459 help, like maybe drafting email replies or scheduling. 00:07:08.839 --> 00:07:11.220 Not for general browsing. Right, right. And just 00:07:11.220 --> 00:07:13.120 to clarify, you mentioned prompt drift. What 00:07:13.120 --> 00:07:15.800 exactly is that again? Ah, prompt drift. It's 00:07:15.800 --> 00:07:19.100 basically when an AI model, if you give it too 00:07:19.100 --> 00:07:21.379 much background info or a really long, complicated 00:07:21.379 --> 00:07:23.560 task, it slowly starts to lose track of your 00:07:23.560 --> 00:07:25.639 original goal. It starts giving you stuff that's 00:07:25.639 --> 00:07:27.579 kind of related but not really what you asked 00:07:27.579 --> 00:07:29.779 for anymore. It's sort of... Drifts off course. 00:07:29.819 --> 00:07:31.939 Got it, that makes sense. Okay, moving on to 00:07:31.939 --> 00:07:34.420 other time savers. Workspaces sound useful, keeping 00:07:34.420 --> 00:07:37.199 tabs organized by project. Client work here, 00:07:37.339 --> 00:07:39.579 personal research there. Loretto, quick commands. 00:07:39.759 --> 00:07:41.670 Are there shortcuts we should memorize? Definitely 00:07:41.670 --> 00:07:43.889 learn the slash commands. Brief gives you a quick 00:07:43.889 --> 00:07:46.930 summary of the page you're on. Super handy. Compare 00:07:46.930 --> 00:07:49.930 shows you the differences between, say, two products 00:07:49.930 --> 00:07:52.970 or concepts clearly laid out. News pulls up a 00:07:52.970 --> 00:07:55.269 tailored list of headlines based on your interests. 00:07:55.769 --> 00:07:58.170 And crucially, you can save your own custom shortcuts 00:07:58.170 --> 00:08:00.490 for tasks you do all the time. Last point on 00:08:00.490 --> 00:08:04.509 setup. Setting the default answer depth to concise. 00:08:04.829 --> 00:08:08.170 Start short, basically. And saving two reusable 00:08:08.170 --> 00:08:10.709 prompts sounds smart for getting into that manager 00:08:10.709 --> 00:08:13.430 mindset. Exactly. One prompt should just ask 00:08:13.430 --> 00:08:15.810 for a quick summary. Any action items and points 00:08:15.810 --> 00:08:17.889 needing immediate attention. Real quick overview. 00:08:18.269 --> 00:08:20.230 The second should be more like an action checklist, 00:08:20.389 --> 00:08:23.050 maybe structured by what needs doing today, this 00:08:23.050 --> 00:08:25.370 week, this month. These act like templates, so 00:08:25.370 --> 00:08:27.430 you're ready to delegate quickly. Okay, wow. 00:08:27.449 --> 00:08:29.709 That's a lot on setup. But if I only have time 00:08:29.709 --> 00:08:32.710 for one thing today, the absolute highest leverage 00:08:32.710 --> 00:08:36.240 move. What gives the most long -term value? Making 00:08:36.240 --> 00:08:39.279 Comet your default browser. It forces the habit. 00:08:39.360 --> 00:08:41.899 That's the foundation. So once you've done that 00:08:41.899 --> 00:08:44.200 setup, Comet really starts to feel different. 00:08:44.320 --> 00:08:47.179 It goes beyond just search. It moves into actual 00:08:47.179 --> 00:08:50.159 task execution. And that's when it stops feeling 00:08:50.159 --> 00:08:52.340 like just software you use and starts feeling 00:08:52.340 --> 00:08:55.049 more like a... like a teammate, almost. Yeah, 00:08:55.169 --> 00:08:57.549 the YouTube smart fact -checking example you 00:08:57.549 --> 00:09:00.149 shared is brilliant. You're watching some dense, 00:09:00.169 --> 00:09:02.850 maybe two -hour tech review video, and you don't 00:09:02.850 --> 00:09:04.590 want to scrub through the whole timeline looking 00:09:04.590 --> 00:09:07.649 for one specific point. You just ask Comet, find 00:09:07.649 --> 00:09:10.009 the exact moment the reviewer talks about the 00:09:10.009 --> 00:09:12.220 phone's battery life. And boom, it takes you 00:09:12.220 --> 00:09:14.960 right there. Then you can follow up, ask it something 00:09:14.960 --> 00:09:17.139 like, OK, now list all the main points the reviewer 00:09:17.139 --> 00:09:19.080 made about the battery, and give me the time 00:09:19.080 --> 00:09:22.539 stamps. And it spits out this clear sourced list 00:09:22.539 --> 00:09:24.740 of bullet points. It's incredibly efficient. 00:09:25.179 --> 00:09:26.860 The adaptability is really something, too. I 00:09:26.860 --> 00:09:29.539 actually tested this. And Comet initially used 00:09:29.539 --> 00:09:32.039 a video that someone else had re -uploaded, not 00:09:32.039 --> 00:09:33.580 the official one. So I just corrected it mid 00:09:33.580 --> 00:09:35.820 -task. I said, no, use the video from the official 00:09:35.820 --> 00:09:37.899 channel. Just switch sources immediately and 00:09:37.899 --> 00:09:40.399 read the analysis. Pretty cool. But the travel 00:09:40.399 --> 00:09:42.679 plan example, that's where I had that genuine 00:09:42.679 --> 00:09:45.460 whoa moment. Imagine planning a very specific 00:09:45.460 --> 00:09:47.740 one -day trip in a city. You want to start at 00:09:47.740 --> 00:09:51.080 the central station. You specify, hit this museum, 00:09:51.179 --> 00:09:53.620 then the old town, then this park. And crucially, 00:09:53.820 --> 00:09:55.879 you say only use public transport, and you want 00:09:55.879 --> 00:09:58.179 to avoid backtracking across the city. Right. 00:09:58.460 --> 00:10:00.879 And the output isn't just a list of website links. 00:10:01.080 --> 00:10:04.070 No. Comet asks if it can connect to Google Maps, 00:10:04.250 --> 00:10:06.870 then it orders the locations logically to minimize 00:10:06.870 --> 00:10:09.190 travel time, and it gives you a structured table. 00:10:09.350 --> 00:10:12.850 It's like, oh, location, transport method, estimated 00:10:12.850 --> 00:10:15.769 travel time, and maybe some notes. Whoa. I mean, 00:10:16.029 --> 00:10:18.870 imagine scaling that level of detailed location 00:10:18.870 --> 00:10:21.330 -aware planning to manage an entire week's schedule, 00:10:21.529 --> 00:10:24.970 or like a complex client project timeline. That's 00:10:24.970 --> 00:10:27.860 incredible control. OK, if I'm delegating something 00:10:27.860 --> 00:10:30.539 that personal, like my travel plans, what happens 00:10:30.539 --> 00:10:32.440 if it gets something important wrong? Is there 00:10:32.440 --> 00:10:35.460 a risk of just relying on it too much? That's 00:10:35.460 --> 00:10:37.279 exactly where the manager mindset comes back 00:10:37.279 --> 00:10:39.940 in. You are still the boss, the final editor. 00:10:40.440 --> 00:10:42.960 If it suggests a bus route that you know is terrible, 00:10:43.059 --> 00:10:44.720 you don't throw the whole plan out. You just 00:10:44.720 --> 00:10:47.139 refine the instruction. You say, OK, public transport 00:10:47.139 --> 00:10:49.259 only, but avoid walks longer than 10 minutes. 00:10:49.419 --> 00:10:52.360 And it recalculates, finds a new route. The risk 00:10:52.360 --> 00:10:55.039 is lower because you only adjust one small instruction. 00:10:55.379 --> 00:10:57.519 You don't have to go back and manually research 00:10:57.519 --> 00:10:59.980 20 different options yourself. So thinking about 00:10:59.980 --> 00:11:01.860 that, what's the biggest shift in mindset needed 00:11:01.860 --> 00:11:04.740 to successfully delegate something complex, like 00:11:04.740 --> 00:11:06.879 detailed travel planning? You have to actually 00:11:06.879 --> 00:11:09.580 trust the tool enough to let it take the first 00:11:09.580 --> 00:11:12.259 pass, let it do the heavy lifting, and then focus 00:11:12.259 --> 00:11:14.340 your energy on checking and refining the final 00:11:14.340 --> 00:11:17.440 result. All right, let's quickly run through 00:11:17.440 --> 00:11:20.360 those six real -world workflows that the sources 00:11:20.360 --> 00:11:23.539 highlight as saving hours every week. This is 00:11:23.539 --> 00:11:25.740 where you see the practical utility shine. Okay, 00:11:26.100 --> 00:11:28.759 workflow one, finding facts. You open maybe a 00:11:28.759 --> 00:11:31.620 skeptical article online, ask comment, summarize 00:11:31.620 --> 00:11:34.500 this in five points, and then crucially, extract 00:11:34.500 --> 00:11:36.879 every single link to a scientific study mentioned 00:11:36.879 --> 00:11:39.240 in this text, instant verification. Workflow 00:11:39.240 --> 00:11:43.100 two, content planning. This is great. Analyze, 00:11:43.100 --> 00:11:44.940 say, the top five articles currently trending 00:11:44.940 --> 00:11:47.659 on a topic like saving money for young people. 00:11:49.019 --> 00:11:51.580 Ask it to identify gaps in what they cover. Draft 00:11:51.580 --> 00:11:53.440 a detailed outline that fills those gaps and 00:11:53.440 --> 00:11:55.279 then generate five different title options. Maybe 00:11:55.279 --> 00:11:57.200 a simple one, a curious one, a surprising opinion, 00:11:57.279 --> 00:11:59.179 one with data, maybe a fun one. You get a whole 00:11:59.179 --> 00:12:01.460 content strategy in like under a minute. Workflow 00:12:01.460 --> 00:12:04.500 three. Build in a buyer's guide. This usually 00:12:04.500 --> 00:12:06.860 involves so much data hunting. Let's say you 00:12:06.860 --> 00:12:09.100 want to compare three specific student laptops, 00:12:09.340 --> 00:12:12.419 all under $700. Doing that manually is easily 00:12:12.419 --> 00:12:14.379 30 minutes of just hopping between tabs. All 00:12:14.379 --> 00:12:16.879 right. Instead, you ask Comet, list only the 00:12:16.879 --> 00:12:18.860 relevant specs. Battery life, weight, keyboard 00:12:18.860 --> 00:12:21.940 quality. Focus it. Then ask for current prices, 00:12:22.139 --> 00:12:25.299 say for October 2025, and give me a recommendation 00:12:25.299 --> 00:12:28.009 for best value and best performance. And the 00:12:28.009 --> 00:12:30.330 output isn't just text, it's often a clean table. 00:12:30.750 --> 00:12:32.690 It's immediately actionable, cross -referenced 00:12:32.690 --> 00:12:35.230 against its sources. It basically transforms 00:12:35.230 --> 00:12:37.850 all that scattered info from web pages into neat 00:12:37.850 --> 00:12:40.230 little Lego blocks of data that you can easily 00:12:40.230 --> 00:12:43.649 compare and use. Workflow 4. Learning a new skill. 00:12:44.090 --> 00:12:46.129 This is smart. Instead of just searching how 00:12:46.129 --> 00:12:48.570 to be a video editor, you tell Comet to look 00:12:48.570 --> 00:12:50.970 at current job postings for a junior video editor. 00:12:51.690 --> 00:12:53.789 What are companies actually asking for right 00:12:53.789 --> 00:12:57.179 now? It finds those core skills the market demands, 00:12:57.539 --> 00:12:59.399 then builds you a custom seven -day learning 00:12:59.399 --> 00:13:02.340 plan using only free resources it can find online. 00:13:02.840 --> 00:13:05.259 Complete with links to tutorials, articles, and 00:13:05.259 --> 00:13:07.480 even suggestions for practice projects. It's 00:13:07.480 --> 00:13:12.230 actionable, context -aware learning. Use that 00:13:12.230 --> 00:13:14.669 special news command we mentioned, but be specific. 00:13:15.049 --> 00:13:17.690 Request original reporting sources only, no aggregation 00:13:17.690 --> 00:13:20.210 if possible. Ask for 50 -word summaries. Tell 00:13:20.210 --> 00:13:22.429 it to focus strictly on, say, technology and 00:13:22.429 --> 00:13:24.370 science news, and maybe mark any stories that 00:13:24.370 --> 00:13:26.409 are still developing. You get an instant clarity 00:13:26.409 --> 00:13:28.809 and just radically cut down the noise. And workflow 00:13:28.809 --> 00:13:33.070 six. Pulling data from PDFs. We all deal with 00:13:33.070 --> 00:13:35.909 these dense documents. Open a, let's say, 50 00:13:35.909 --> 00:13:39.419 -page financial report PDF inside Comet. Ask 00:13:39.419 --> 00:13:41.539 it to. Pull out the key statistics mentioned, 00:13:42.080 --> 00:13:43.820 find three direct quotes supporting the main 00:13:43.820 --> 00:13:46.799 argument with page numbers, and write a two -sentence 00:13:46.799 --> 00:13:50.379 executive conclusion. Huh. That just eliminates 00:13:50.379 --> 00:13:52.700 potentially hours of manually searching inside 00:13:52.700 --> 00:13:55.299 those clunky documents. Okay. Thinking about 00:13:55.299 --> 00:13:57.720 the listener who feels just completely overwhelmed 00:13:57.720 --> 00:14:00.480 by the firehose of information every day, which 00:14:00.480 --> 00:14:02.720 one of these workflows offers the biggest, most 00:14:02.720 --> 00:14:04.539 immediate noise reduction? Probably workflow 00:14:04.539 --> 00:14:07.210 five. using that concise filtered news command 00:14:07.210 --> 00:14:09.990 and really demanding only primary sources. That 00:14:09.990 --> 00:14:12.169 cuts through a lot of chatter. And all this brings 00:14:12.169 --> 00:14:14.690 us back to that core idea we started with. You're 00:14:14.690 --> 00:14:16.870 not using Comet just to find information anymore. 00:14:17.009 --> 00:14:19.070 You're using it to actually complete tasks. The 00:14:19.070 --> 00:14:21.330 mental shift is, well, it's profound. You stop 00:14:21.330 --> 00:14:23.169 being the hunter, manually clicking, reading, 00:14:23.370 --> 00:14:25.610 pasting, and you become the manager, delegating 00:14:25.610 --> 00:14:28.639 all that tedious low value work. Right. And delegation, 00:14:28.860 --> 00:14:31.440 even when it's to an AI, needs sharp instructions. 00:14:32.120 --> 00:14:33.519 You've got to think of it like working with a 00:14:33.519 --> 00:14:35.620 really capable, really smart intern. They can 00:14:35.620 --> 00:14:38.240 do amazing things, but they need crystal clear 00:14:38.240 --> 00:14:41.220 directions. Exactly. Your prompts need to be 00:14:41.220 --> 00:14:43.139 specific about the format you want, the length. 00:14:43.340 --> 00:14:45.679 Don't just say summarize. Say, summarize this 00:14:45.679 --> 00:14:47.659 article in five bullet points, keep it under 00:14:47.659 --> 00:14:50.779 100 words total, and use a skeptical tone. Be 00:14:50.779 --> 00:14:54.580 precise. And demand quality sources. That's key 00:14:54.580 --> 00:14:57.559 for reducing the risk of getting bad info. If 00:14:57.559 --> 00:14:59.480 it gives you summaries based on, I don't know, 00:14:59.539 --> 00:15:02.299 some random blogs, push back immediately. Clarify. 00:15:02.759 --> 00:15:04.980 Replace these sources with primary research papers 00:15:04.980 --> 00:15:07.460 or use only academic sources for this section. 00:15:07.840 --> 00:15:09.820 Always guide it when needed. Don't just give 00:15:09.820 --> 00:15:12.159 up and start over. Just tell it, make this section 00:15:12.159 --> 00:15:14.559 shorter or simplify the jargon here. Iterate. 00:15:14.639 --> 00:15:16.559 So the call to action here is pretty simple. 00:15:16.700 --> 00:15:19.120 Don't try to overhaul your entire workflow overnight. 00:15:19.159 --> 00:15:22.559 That's overwhelming. Just pick one specific recurring 00:15:22.559 --> 00:15:24.980 task that you tend to avoid because it feels 00:15:24.980 --> 00:15:27.919 like drudgery. Maybe it's researching a complex 00:15:27.919 --> 00:15:30.320 topic, comparing a few products before buying, 00:15:30.720 --> 00:15:32.899 drafting an executive summary of a long report. 00:15:33.360 --> 00:15:36.080 Pick one and use Comet to actually finish that 00:15:36.080 --> 00:15:39.200 task today. See how it feels. Yeah. The big idea 00:15:39.200 --> 00:15:41.980 really is this. Your browser should be empowering 00:15:41.980 --> 00:15:44.480 you. not constantly distracting you or bogging 00:15:44.480 --> 00:15:47.440 you down. For years, we just accepted that doing 00:15:47.440 --> 00:15:50.000 research meant getting lost in this endless sea 00:15:50.000 --> 00:15:52.980 of tabs. Comet is designed specifically to tackle 00:15:52.980 --> 00:15:55.179 that problem. It's probably the closest thing 00:15:55.179 --> 00:15:57.960 we have right now to a genuinely integrated assistant 00:15:57.960 --> 00:16:00.940 built right into your digital workflow. So here's 00:16:00.940 --> 00:16:03.139 a final thought to leave you with. Think about 00:16:03.139 --> 00:16:05.460 how much of your professional identity and maybe 00:16:05.460 --> 00:16:07.659 just your daily energy is currently tied up in 00:16:07.659 --> 00:16:10.519 the act of searching, just finding stuff. When 00:16:10.519 --> 00:16:12.960 that lower level task gets successfully delegated 00:16:12.960 --> 00:16:15.440 to a tool like this, what higher level, more 00:16:15.440 --> 00:16:17.740 strategic thinking could you finally dedicate 00:16:17.740 --> 00:16:20.519 your limited time and energy to? Set up Comet 00:16:20.519 --> 00:16:22.539 correctly using those steps we talked about. 00:16:22.840 --> 00:16:26.159 Try delegating just one task this week and just 00:16:26.159 --> 00:16:28.279 observe the change in your focus. Our Dean Dive 00:16:28.279 --> 00:16:28.820 is complete.