WEBVTT 00:00:00.000 --> 00:00:03.759 So do you still only ask AI for, you know, the 00:00:03.759 --> 00:00:06.700 weather forecast? Or maybe just drafting this 00:00:06.700 --> 00:00:08.580 really basic email you end up deleting anyway? 00:00:08.839 --> 00:00:10.919 It feels like a lot of us, well, we treat these 00:00:10.919 --> 00:00:13.339 super powerful tools like they're just fancy 00:00:13.339 --> 00:00:15.599 calculators. We're kind of missing out on their 00:00:15.599 --> 00:00:18.260 real potential as like a proper personal assistant, 00:00:18.339 --> 00:00:20.780 a teacher, even a creative partner. We're leaving 00:00:20.780 --> 00:00:22.940 hours of saved time on the table just asking 00:00:22.940 --> 00:00:26.260 questions, not really delegating tasks. Welcome 00:00:26.260 --> 00:00:29.070 to the Deep Dive. Yeah, today we're going to 00:00:29.070 --> 00:00:31.589 unpack a whole bunch of sources detailing, I 00:00:31.589 --> 00:00:34.409 think it's 24 really practical ways you can shift 00:00:34.409 --> 00:00:38.289 how you use AI, moving from just asking it stuff 00:00:38.289 --> 00:00:40.429 to actually getting it to do things for you. 00:00:40.670 --> 00:00:43.210 We've pulled out the best bits on smarter shopping, 00:00:43.509 --> 00:00:46.469 fixing things around the house instantly, optimizing 00:00:46.469 --> 00:00:48.509 your workflow, and even some pretty mind blowing 00:00:48.509 --> 00:00:50.729 research tools. And look, the goal isn't to make 00:00:50.729 --> 00:00:53.689 you some kind of tech It's really just about 00:00:53.689 --> 00:00:55.570 showing you how to weave AI into your day -to 00:00:55.570 --> 00:00:57.229 -day life. Save yourself some hours. Get rid 00:00:57.229 --> 00:01:00.310 of that boring, repetitive stuff for good. It's 00:01:00.310 --> 00:01:02.990 about action, not just ideas. Okay, yeah, let's 00:01:02.990 --> 00:01:06.109 get into it. Maybe we can start with how AI can 00:01:06.109 --> 00:01:08.810 fundamentally change... well, how you spend your 00:01:08.810 --> 00:01:12.370 money, and maybe cut down on that awful feeling 00:01:12.370 --> 00:01:14.810 buyer's remorse. You know, when you buy something 00:01:14.810 --> 00:01:17.650 big, a fridge, a laptop, whatever, and the worst 00:01:17.650 --> 00:01:19.930 feeling is buying it, and then, bam, next week, 00:01:20.030 --> 00:01:22.819 the new model comes out for the same price. Ugh. 00:01:23.159 --> 00:01:25.359 Right. That's exactly the first thing AI tackles. 00:01:25.459 --> 00:01:27.879 It can intelligently look at the predictable 00:01:27.879 --> 00:01:30.719 product cycle. We often try to Google this stuff 00:01:30.719 --> 00:01:34.260 manually, but AI can quickly scan the patterns 00:01:34.260 --> 00:01:36.920 for big companies. Apple, Samsung, you name it. 00:01:36.920 --> 00:01:38.859 So you could ask it something like, hey, I'm 00:01:38.859 --> 00:01:40.500 thinking about getting an iPad Air. When did 00:01:40.500 --> 00:01:42.620 the current one come out? How often does Apple 00:01:42.620 --> 00:01:44.939 update this line? Should I buy it now or maybe 00:01:44.939 --> 00:01:47.459 wait a bit? And the sources say AI gives you 00:01:47.459 --> 00:01:49.819 a very good guess, that's the quote, about whether 00:01:49.819 --> 00:01:51.620 waiting a few months makes sense financially. 00:01:51.689 --> 00:01:53.849 or if you're actually in a good window to buy. 00:01:54.069 --> 00:01:56.290 Mm -hmm. But it's not just release dates, is 00:01:56.290 --> 00:01:59.030 it? It goes into historical pricing, too. Exactly. 00:01:59.390 --> 00:02:01.390 For those really pricey things, you can ask it 00:02:01.390 --> 00:02:03.790 about past sales, Black Friday, Prime Day, that 00:02:03.790 --> 00:02:06.109 sort of thing. It basically becomes your historical 00:02:06.109 --> 00:02:08.750 price checker. Okay, so if I'm looking at, say, 00:02:08.889 --> 00:02:12.789 a fancy TV, I could ask, what? Like... Based 00:02:12.789 --> 00:02:16.810 on past sales for this Sony A90J TV, how much 00:02:16.810 --> 00:02:19.110 does the price usually drop during big sales 00:02:19.110 --> 00:02:21.110 events? Yeah, precisely. And it might come back 00:02:21.110 --> 00:02:23.490 saying, look, this model usually drops 20, 25 00:02:23.490 --> 00:02:25.330 percent. Boom, now you know what price to aim 00:02:25.330 --> 00:02:27.310 for and you won't overpay right now. OK, but 00:02:27.310 --> 00:02:29.729 hang on. How reliable are those predictions? 00:02:30.270 --> 00:02:31.969 I mean, things change, right? Supply chains get 00:02:31.969 --> 00:02:35.590 messed up. Is basing a buy on old data risky? 00:02:35.810 --> 00:02:37.849 That's a fair point. And the sources are clear. 00:02:37.909 --> 00:02:40.030 It's predicting probability based on tons of 00:02:40.030 --> 00:02:42.689 data. It's not like a guarantee, its strength 00:02:42.689 --> 00:02:45.150 is the sheer volume of past data it can crunch 00:02:45.150 --> 00:02:47.830 in seconds. Way better than just guessing, even 00:02:47.830 --> 00:02:49.930 if it's not a perfect crystal ball. Interesting. 00:02:50.030 --> 00:02:53.009 So if AI can kind of predict release dates and 00:02:53.009 --> 00:02:55.469 discounts, what does that really change about 00:02:55.469 --> 00:02:57.729 how we research stuff we buy? Well, knowing those 00:02:57.729 --> 00:03:00.110 cycles gives you power, right? It helps you avoid 00:03:00.110 --> 00:03:02.830 paying that new release premium unnecessarily. 00:03:02.930 --> 00:03:05.110 That could save some serious cash on big stuff. 00:03:05.389 --> 00:03:07.469 OK, let's bring it down a notch. How about using 00:03:07.469 --> 00:03:10.409 this for those annoying little problems? Right 00:03:10.409 --> 00:03:12.969 in our own homes. Oh, yeah. Moving into the home. 00:03:13.389 --> 00:03:15.830 AI is basically your universal user manual now. 00:03:15.949 --> 00:03:18.090 You know that moment, the printer jams, washing 00:03:18.090 --> 00:03:20.229 machine flashes some weird code? And the manual 00:03:20.229 --> 00:03:23.129 vanished years ago. Exactly. Or you call support 00:03:23.129 --> 00:03:25.189 and sit on hold for like half an hour listening 00:03:25.189 --> 00:03:29.530 to bad music. Yeah. So AI skips all that. Totally. 00:03:29.729 --> 00:03:31.430 It's like having a manual for everything you 00:03:31.430 --> 00:03:33.289 own, but one that actually talks you through 00:03:33.289 --> 00:03:35.990 the fix, step by step. There's a great example 00:03:35.990 --> 00:03:39.729 in the source. Quick. tech support. Say your 00:03:39.729 --> 00:03:43.650 LG washer model, WM3400CW, they mentioned shows 00:03:43.650 --> 00:03:48.349 error DE. Ask the AI, give me five specific steps 00:03:48.349 --> 00:03:53.469 to fix error code DE on my LG WM3400CW. Apparently 00:03:53.469 --> 00:03:55.789 90 % of the time it's something simple like the 00:03:55.789 --> 00:03:58.030 door isn't shut properly, AI gets you there in 00:03:58.030 --> 00:04:00.210 five minutes, no service call needed. Wow, okay. 00:04:00.449 --> 00:04:02.500 That instant fixability. Does it work in the 00:04:02.500 --> 00:04:05.180 kitchen, too? Like, beyond just basic substitutions? 00:04:05.379 --> 00:04:07.039 Oh, absolutely. We've all asked, what can I use 00:04:07.039 --> 00:04:08.439 instead of butter? But yeah, it goes deeper. 00:04:08.520 --> 00:04:10.439 It gives you adjustment instructions. OK, say 00:04:10.439 --> 00:04:13.699 you're making banana bread. Recipe needs half 00:04:13.699 --> 00:04:15.740 a cup of brown sugar. You only have white sugar 00:04:15.740 --> 00:04:18.839 and honey. AI doesn't just say, use white sugar. 00:04:19.600 --> 00:04:22.079 It'll suggest, OK, use that half cup of white 00:04:22.079 --> 00:04:24.279 sugar, plus maybe a tablespoon of honey. Oh, 00:04:24.279 --> 00:04:26.689 why the honey, though? to keep the taste right 00:04:26.689 --> 00:04:29.990 but also and this is key the moisture level that 00:04:29.990 --> 00:04:31.569 the brown sugar would have given it's actually 00:04:31.569 --> 00:04:34.509 adjusting the recipe's chemistry not just swapping 00:04:34.509 --> 00:04:37.990 names it's like a proper kitchen helper and honestly 00:04:37.990 --> 00:04:40.850 i still wrestle with prompt drift myself sometimes 00:04:40.850 --> 00:04:42.790 when i try the really complex stuff you know 00:04:42.790 --> 00:04:45.370 mixing recipe ideas but for these quick home 00:04:45.370 --> 00:04:49.000 fixes the immediate results are just Yeah, incredible. 00:04:49.139 --> 00:04:51.199 That fast success really builds your confidence. 00:04:51.459 --> 00:04:53.600 So beyond just being convenient, how does getting 00:04:53.600 --> 00:04:56.040 those quick fixes at home make us feel better 00:04:56.040 --> 00:04:58.439 about using AI for, like, bigger work problems? 00:04:58.560 --> 00:05:00.740 Quick wins on simple things build that crucial 00:05:00.740 --> 00:05:03.399 trust. Makes you more willing to tackle complex 00:05:03.399 --> 00:05:05.399 tasks later. Okay, this is where I think it gets 00:05:05.399 --> 00:05:08.339 really interesting. Workflow automation. Moving 00:05:08.339 --> 00:05:10.959 past using AI like a search box and actually 00:05:10.959 --> 00:05:13.660 teaching it your specific style. Yeah, this is 00:05:13.660 --> 00:05:16.279 about ditching the repetitive grind without sounding 00:05:16.279 --> 00:05:18.660 like a robot. Think about document templates. 00:05:19.060 --> 00:05:21.699 You write that same weekly progress report or 00:05:21.699 --> 00:05:24.300 internal memo every single Monday, right? The 00:05:24.300 --> 00:05:26.300 structure is always similar. Copying and pasting 00:05:26.300 --> 00:05:32.819 is just so boring. So the AI way is what? Upload 00:05:32.819 --> 00:05:35.360 old reports. Exactly. Upload maybe three or four 00:05:35.360 --> 00:05:38.319 of your past reports, .txt files, PDFs, whatever. 00:05:39.019 --> 00:05:41.120 Then you ask the AI to analyze the structure, 00:05:41.339 --> 00:05:43.459 the tone you use, the common phrases. It literally 00:05:43.459 --> 00:05:46.379 learns your voice, your style. Then you tell 00:05:46.379 --> 00:05:48.800 it, create a new template based on these examples, 00:05:48.920 --> 00:05:51.740 maybe with sections like. what I did and next 00:05:51.740 --> 00:05:53.720 steps. So next time, you just give it the bullet 00:05:53.720 --> 00:05:57.160 points. AI fills in like 90 % of it. But in your 00:05:57.160 --> 00:05:59.459 style, you get to focus on the actual substance, 00:05:59.480 --> 00:06:01.720 not the formatting. Hold on though. Uploading 00:06:01.720 --> 00:06:03.899 company reports, what about security? Right. 00:06:04.139 --> 00:06:06.240 Or, you know, AI making stuff up, hallucinating. 00:06:06.379 --> 00:06:08.500 Critical question. The sources definitely hit 00:06:08.500 --> 00:06:12.439 on this. Key thing is understanding your AI tools 00:06:12.439 --> 00:06:15.740 privacy policy. Does the data train their model 00:06:15.740 --> 00:06:18.920 or stay private? For work stuff, the advice is 00:06:18.920 --> 00:06:22.689 clear. scrub any really sensitive names or numbers 00:06:22.689 --> 00:06:25.529 before you upload. And yeah, hallucination risk 00:06:25.529 --> 00:06:28.250 is lower when it's analyzing structure versus 00:06:28.250 --> 00:06:30.310 generating pure fiction, but you always gotta 00:06:30.310 --> 00:06:33.009 double check what it puts in. That same idea, 00:06:33.170 --> 00:06:35.470 commanding it, works for calendar management 00:06:35.470 --> 00:06:38.550 too. Such a time suck, right? Forget manually 00:06:38.550 --> 00:06:41.790 clicking around. You just say, find me two hours 00:06:41.790 --> 00:06:44.550 free next week for deep work, one morning slot, 00:06:44.610 --> 00:06:47.050 one afternoon, no meetings, and just add deep 00:06:47.050 --> 00:06:49.490 work to my calendar. That's proper delegation. 00:06:49.769 --> 00:06:51.910 And meeting summaries, too, instead of making 00:06:51.910 --> 00:06:54.389 someone type up notes. Yeah. Feed it the transcript 00:06:54.389 --> 00:06:56.389 from your Zoom call or whatever. Ask it. List 00:06:56.389 --> 00:06:58.350 five action items from this meeting, who owns 00:06:58.350 --> 00:07:00.329 them, and remind me to follow up next Tuesday. 00:07:00.949 --> 00:07:02.750 Takes like three steps, turns them into one command. 00:07:03.009 --> 00:07:04.750 That's all about stacking those commands, it 00:07:04.750 --> 00:07:07.069 seems. Totally. Like travel planning. Say you're 00:07:07.069 --> 00:07:09.790 driving from Saigon to Vung Tau in Vietnam. You 00:07:09.790 --> 00:07:12.730 can ask the AI, find a really good Bangkok place. 00:07:13.269 --> 00:07:15.529 Those amazing little savory pancakes. That's 00:07:15.529 --> 00:07:18.339 actually on my route. Oh, wow. Yeah. And it doesn't 00:07:18.339 --> 00:07:20.420 just find it. It tells you how much extra time 00:07:20.420 --> 00:07:23.480 that stop will add and adjusts your ETA automatically. 00:07:23.740 --> 00:07:26.040 Directions, food, research, culture, time, math, 00:07:26.459 --> 00:07:30.060 all in one go. OK, so when AI learns our writing 00:07:30.060 --> 00:07:32.680 style, manages our time, plans our snack stops, 00:07:33.480 --> 00:07:35.819 where do we draw the line on outsourcing? What's 00:07:35.819 --> 00:07:38.300 left for us? Well, outsourcing the boring structure 00:07:38.300 --> 00:07:40.579 frees us up, right? Let's us focus on the human 00:07:40.579 --> 00:07:43.300 stuff connection. Real insight. Mid -roll sponsor, 00:07:43.480 --> 00:07:46.139 read placeholder. Right, let's shift gears a 00:07:46.139 --> 00:07:48.519 bit. Let's talk deep research and maybe pushing 00:07:48.519 --> 00:07:51.019 creative boundaries. This seems super relevant 00:07:51.019 --> 00:07:52.779 for anyone who needs to absorb a ton of info 00:07:52.779 --> 00:07:55.720 quickly and not just by passively reading. Right, 00:07:55.860 --> 00:07:57.540 forget just reading. You can actually turn those 00:07:57.540 --> 00:07:59.779 boring documents into something engaging. Got 00:07:59.779 --> 00:08:03.980 a 50 -page PDF, dense technical manual, school 00:08:03.980 --> 00:08:08.100 paper to review, upload it, then ask the AI something 00:08:08.100 --> 00:08:11.040 like... create a short podcast script, maybe 00:08:11.040 --> 00:08:13.839 five minutes long, two hosts discussing the main 00:08:13.839 --> 00:08:16.319 points from this document, and explain the tricky 00:08:16.319 --> 00:08:19.759 ideas simply. Suddenly, learning isn't just passive 00:08:19.759 --> 00:08:22.579 reading. It's like actively reviewing content. 00:08:22.720 --> 00:08:25.100 It helps you grasp it and remember it way better. 00:08:25.240 --> 00:08:26.980 You could even get it to create personalized 00:08:26.980 --> 00:08:29.180 study guides. Like, feed it two different articles 00:08:29.180 --> 00:08:31.699 on a historical event with opposing views. Yeah. 00:08:31.879 --> 00:08:34.679 And ask it for a simple side by side. Argument 00:08:34.679 --> 00:08:37.440 A versus Argument B. It synthesizes the conflict 00:08:37.440 --> 00:08:40.360 for you. Doesn't just spit back facts. That synthesis 00:08:40.360 --> 00:08:42.580 part is powerful. And does this work for videos 00:08:42.580 --> 00:08:45.100 too? Like long lectures? Absolutely. Instead 00:08:45.100 --> 00:08:46.919 of scrubbing through a two -hour video lecture, 00:08:47.080 --> 00:08:49.519 give the AI the link. Ask it. Find the exact 00:08:49.519 --> 00:08:51.059 timestamp where the professor starts talking 00:08:51.059 --> 00:08:53.139 about the Industrial Revolution. And give me 00:08:53.139 --> 00:08:55.220 a quick bullet summary of the three main points 00:08:55.220 --> 00:08:57.779 they make in that section. The time -saving there. 00:08:58.159 --> 00:09:00.360 It feels huge. You just jump straight to the 00:09:00.360 --> 00:09:02.960 core info you actually need. Monumental time 00:09:02.960 --> 00:09:05.519 -saving. And this kind of leads us to that... 00:09:05.580 --> 00:09:07.919 that moment of wonder the source has mentioned, 00:09:08.899 --> 00:09:11.580 video translation with perfect lip sync. We're 00:09:11.580 --> 00:09:14.179 talking tools like HayGen here. Explain that 00:09:14.179 --> 00:09:16.279 again, how does that work? Okay, so you upload 00:09:16.279 --> 00:09:18.179 your video, let's say you're speaking English, 00:09:18.279 --> 00:09:20.460 but you need it in Spanish. You pick Spanish, 00:09:21.000 --> 00:09:23.639 the AI creates a new video. You're speaking fluent 00:09:23.639 --> 00:09:26.759 Spanish, but, and this is the wild part, it keeps 00:09:26.759 --> 00:09:29.659 your original voice tone. Okay. But the magic, 00:09:29.740 --> 00:09:32.679 the real leap, it analyzes how your mouth moved, 00:09:32.940 --> 00:09:35.159 making the English sounds, and then it actually 00:09:35.159 --> 00:09:37.440 changes your mouth movements in the video, pixel 00:09:37.440 --> 00:09:40.200 by pixel, to perfectly match the new Spanish 00:09:40.200 --> 00:09:43.299 words. Whoa, wait, seriously, it changes how 00:09:43.299 --> 00:09:46.080 your lips move? Yeah, perfectly synced. Imagine 00:09:46.080 --> 00:09:48.320 scaling human connection like that, breaking 00:09:48.320 --> 00:09:51.220 language barriers for, like, global business 00:09:51.220 --> 00:09:53.870 or online education. Seamlessly, it's amazing. 00:09:53.950 --> 00:09:57.450 It's genuinely mind -blowing. So if AI can instantly 00:09:57.450 --> 00:09:59.970 break down a two -hour video or translate our 00:09:59.970 --> 00:10:02.470 words and our lit movements, what's the real 00:10:02.470 --> 00:10:05.190 bottleneck now in getting information? The bottleneck 00:10:05.190 --> 00:10:07.750 shifts, right? It's less about processing the 00:10:07.750 --> 00:10:10.610 output and more about curating the right input 00:10:10.610 --> 00:10:12.669 for the AI in the first place. That's a really 00:10:12.669 --> 00:10:16.500 powerful shift. So just to recap quickly. We 00:10:16.500 --> 00:10:18.740 saw AI move from just answering questions to 00:10:18.740 --> 00:10:21.100 being an active assistant. We looked at smarter 00:10:21.100 --> 00:10:24.220 shopping, instant home repairs, personalized 00:10:24.220 --> 00:10:26.759 workflow stuff, and that accelerated research. 00:10:27.240 --> 00:10:29.700 Yeah, covering tips from nailing discount timing 00:10:29.700 --> 00:10:31.580 to getting your calendar sorted automatically, 00:10:32.139 --> 00:10:34.679 all about turning delegation into actual action. 00:10:34.919 --> 00:10:37.559 And the big idea isn't, you know, go try all 00:10:37.559 --> 00:10:39.879 24 things right now. That's overwhelming. It's 00:10:39.879 --> 00:10:43.240 about finding one thing, one annoying, repetitive 00:10:43.240 --> 00:10:45.519 task in your life that AI could actually handle. 00:10:45.480 --> 00:10:48.360 starting today. Right. So if you hate writing 00:10:48.360 --> 00:10:51.460 that same weekly update email, maybe try the 00:10:51.460 --> 00:10:53.259 style analysis and template thing we talked about. 00:10:53.340 --> 00:10:56.200 Or if you cook a lot and always panic about substitutions, 00:10:56.600 --> 00:10:58.840 try that ingredient adjustment trick. Exactly. 00:10:59.120 --> 00:11:02.019 Start small, see how it feels having AI as an 00:11:02.019 --> 00:11:04.100 assistant, then build from there. let it do the 00:11:04.100 --> 00:11:06.940 boring work. So if AI really takes over those 00:11:06.940 --> 00:11:09.980 repetitive tasks, the comparison, shopping, the 00:11:09.980 --> 00:11:11.820 scheduling, the email templates, we get that 00:11:11.820 --> 00:11:14.620 time back. What's the one thing you are genuinely 00:11:14.620 --> 00:11:17.259 excited to reclaim that time for? Is it more 00:11:17.259 --> 00:11:19.320 focused at your job, being more present with 00:11:19.320 --> 00:11:22.019 family, or finally digging into that hobby? Two 00:11:22.019 --> 00:11:23.820 secs silence. Thanks for joining us on this deep 00:11:23.820 --> 00:11:26.200 dive exploring how to make AI your actual active 00:11:26.200 --> 00:11:27.600 assistant. Gout Loach your own music.