WEBVTT 00:00:00.000 --> 00:00:01.840 You know, we often think that Google's entire 00:00:01.840 --> 00:00:05.099 AI is that little summary box that pops up when 00:00:05.099 --> 00:00:07.360 you search for something. But that little box, 00:00:07.400 --> 00:00:09.859 it's really just the front door. The sources 00:00:09.859 --> 00:00:11.880 you send over confirm something pretty surprising, 00:00:12.060 --> 00:00:15.060 which is that most of us are only using the least 00:00:15.060 --> 00:00:17.719 powerful parts of what's available. Underneath 00:00:17.719 --> 00:00:19.739 that search bar, there's this whole free toolkit 00:00:19.739 --> 00:00:23.120 for, well, automation, for content creation. 00:00:23.949 --> 00:00:27.149 deep analysis. So the question is, if you could 00:00:27.149 --> 00:00:29.089 replace half the software you're paying for right 00:00:29.089 --> 00:00:31.329 now with hidden Google tools, would you be willing 00:00:31.329 --> 00:00:33.929 to look? That's exactly the mission for today's 00:00:33.929 --> 00:00:36.369 deep dive. We've got this incredible guide to 00:00:36.369 --> 00:00:39.250 over 30 of Google's free AI tools, and we're 00:00:39.250 --> 00:00:41.130 here to give you the shortcut. And we're going 00:00:41.130 --> 00:00:43.710 way beyond just the standard chatbot. Our goal 00:00:43.710 --> 00:00:46.009 is to give you a clear map across four different 00:00:46.009 --> 00:00:48.030 layers of these free tools. We're talking about 00:00:48.030 --> 00:00:50.670 building no -code mini apps in Canvas, then advanced 00:00:50.670 --> 00:00:53.009 tools in AI Studio, a look at the experimental 00:00:53.229 --> 00:00:55.990 stuff in google labs and finally really unlocking 00:00:55.990 --> 00:00:58.289 the hidden powers inside the gemini chat you 00:00:58.289 --> 00:01:00.990 already use so yeah let's start replacing some 00:01:00.990 --> 00:01:04.090 of those clunky expensive workflows okay let's 00:01:04.090 --> 00:01:07.290 start with that first layer gemini canvas this 00:01:07.290 --> 00:01:09.730 seems to completely change what we think a chat 00:01:09.730 --> 00:01:12.620 bot can even do it absolutely does When we say 00:01:12.620 --> 00:01:15.400 mini app, we're not just asking the AI to spit 00:01:15.400 --> 00:01:17.159 out a little bit of code. No, these are functional 00:01:17.159 --> 00:01:20.540 tools. They have real user interfaces, you know, 00:01:20.540 --> 00:01:23.159 input fields, buttons. You can even upload images. 00:01:23.420 --> 00:01:25.260 And the key is that they're instantly shareable. 00:01:25.400 --> 00:01:27.719 Instantly. The power here really seems to come 00:01:27.719 --> 00:01:30.680 from how it handles vague prompts. The source 00:01:30.680 --> 00:01:33.340 material used this example of a YouTube title 00:01:33.340 --> 00:01:35.719 and thumbnail generator. You just start with 00:01:35.719 --> 00:01:37.939 something really rough, like create an app to 00:01:37.939 --> 00:01:42.459 explode YouTube titles. Yeah. handles that vagueness 00:01:42.459 --> 00:01:45.359 remarkably well. In about, what, 10 or 15 seconds, 00:01:45.420 --> 00:01:47.640 you have a working interface. There's an input 00:01:47.640 --> 00:01:50.519 box, a button to generate titles, an upload area 00:01:50.519 --> 00:01:53.319 for an image. It turns your idea into a working 00:01:53.319 --> 00:01:55.379 blueprint almost immediately. But does it really 00:01:55.379 --> 00:01:57.319 handle natural language for iterating on it? 00:01:57.400 --> 00:01:59.099 I mean, can I have that first version and then 00:01:59.099 --> 00:02:00.959 just say, add an editing box so I can change 00:02:00.959 --> 00:02:03.620 my expression to surprised without code? Yes, 00:02:03.819 --> 00:02:06.659 instantly. You just keep refining the app with 00:02:06.659 --> 00:02:09.180 plain English prompts, and that interface updates 00:02:09.180 --> 00:02:11.680 right in front of you. You're essentially automating 00:02:11.680 --> 00:02:15.219 these repetitive weekly tasks like, say, expense 00:02:15.219 --> 00:02:18.060 reports or product mockups without needing a 00:02:18.060 --> 00:02:21.000 single coding step. So how does building these 00:02:21.000 --> 00:02:24.080 custom shareable interfaces really change the 00:02:24.080 --> 00:02:26.960 user experience compared to just, you know, endlessly 00:02:26.960 --> 00:02:30.180 copy pasting in a standard chat? Well, the interface 00:02:30.180 --> 00:02:32.759 gives you that visual feedback. It elevates a 00:02:32.759 --> 00:02:36.009 simple chat into a unified. professional workflow 00:02:36.009 --> 00:02:38.770 and then when canvas starts to feel a little 00:02:38.770 --> 00:02:41.270 too simple the natural next step is ai studio 00:02:41.270 --> 00:02:43.569 this is where you get really granular control 00:02:43.569 --> 00:02:45.909 yeah i'll admit i still wrestle with prompt drift 00:02:45.909 --> 00:02:47.949 myself sometimes oh yeah where the ai just kind 00:02:47.949 --> 00:02:50.030 of forgets what you asked it to do five messages 00:02:50.030 --> 00:02:52.210 ago exactly so i really appreciate having those 00:02:52.210 --> 00:02:54.210 technical controls in studio to kind of lock 00:02:54.210 --> 00:02:56.990 things down When you first open AI Studio, the 00:02:56.990 --> 00:02:58.789 sources say it can look pretty intimidating, 00:02:58.969 --> 00:03:01.129 lots of technical terms. But you're saying it's 00:03:01.129 --> 00:03:03.289 still approachable for someone who's not a developer. 00:03:03.669 --> 00:03:05.889 It is, yeah. If you just focus on the building 00:03:05.889 --> 00:03:07.990 blocks, the real unlock here is the integrated 00:03:07.990 --> 00:03:11.289 capabilities. AI Studio lets your app pull in 00:03:11.289 --> 00:03:13.449 these other specialized Google models directly. 00:03:13.669 --> 00:03:15.949 Oh, interesting. So you can integrate tools like 00:03:15.949 --> 00:03:18.669 NanoBanana, which is for image editing, or Veo, 00:03:18.909 --> 00:03:21.349 which animates images. You can even pull in live 00:03:21.349 --> 00:03:24.110 data from Google Search and Maps. It's like stacking 00:03:24.110 --> 00:03:26.949 these powerful Lego blocks of function. So what 00:03:26.949 --> 00:03:29.349 are some of the advanced features that make that 00:03:29.349 --> 00:03:31.930 complexity really worth it? Oh, definitely the 00:03:31.930 --> 00:03:34.270 visual annotation tools. You can literally drag 00:03:34.270 --> 00:03:36.889 a box onto an image, circle an area, and just 00:03:36.889 --> 00:03:39.900 prompt, change these colors. instead of trying 00:03:39.900 --> 00:03:41.840 to describe it with text. Plus, you can adjust 00:03:41.840 --> 00:03:44.080 the temperature. Okay, temperature. That's one 00:03:44.080 --> 00:03:46.780 of those AI jargon terms. What does that actually 00:03:46.780 --> 00:03:49.039 control? Think of it like a creativity dial. 00:03:49.199 --> 00:03:52.300 High temperature means more risk -taking, more 00:03:52.300 --> 00:03:55.740 creative outputs. A low temperature gives you 00:03:55.740 --> 00:03:58.460 really reliable, logical, consistent results. 00:03:58.740 --> 00:04:01.219 You get to fine -tune it. The example from the 00:04:01.219 --> 00:04:03.319 source of the children's coloring book generator 00:04:03.319 --> 00:04:06.759 really showcased that. It took the theme... space 00:04:06.759 --> 00:04:10.039 dinosaurs, and a color palette, and just produced 00:04:10.039 --> 00:04:13.280 a downloadable PDF with clean, usable line art. 00:04:13.439 --> 00:04:16.060 Right. And the real power was using those annotation 00:04:16.060 --> 00:04:19.220 tools to refine it in real time. You could just 00:04:19.220 --> 00:04:21.399 circle one of the dinosaurs and say, make this 00:04:21.399 --> 00:04:25.139 one look friendlier. It's so responsive. Beyond 00:04:25.139 --> 00:04:27.779 just content generation, what kind of really 00:04:27.779 --> 00:04:30.399 complex applications are we talking about here? 00:04:30.560 --> 00:04:33.279 You could build data visualization dashboards, 00:04:33.500 --> 00:04:36.220 fully interactive games, or really sophisticated 00:04:36.220 --> 00:04:39.459 multi -step content tools. Okay, so if we move 00:04:39.459 --> 00:04:41.220 on from building our own tools, we get to Google 00:04:41.220 --> 00:04:43.379 Labs. This is sort of the proving ground, right? 00:04:43.500 --> 00:04:46.579 At labs .google .com, are these tools actually 00:04:46.579 --> 00:04:48.740 functional, even if they're labeled experimental? 00:04:49.180 --> 00:04:50.980 Oh, absolutely. They solve real problems right 00:04:50.980 --> 00:04:53.240 now. Take Pamele, for instance. It's for marketing 00:04:53.240 --> 00:04:55.920 campaigns, and it's a huge time saver. The workflow 00:04:55.920 --> 00:04:58.079 is so simple. You just give your website URL. 00:04:58.569 --> 00:05:01.209 And the AI scans the site and generates what 00:05:01.209 --> 00:05:03.670 it calls the brand DNA. What does that mean in 00:05:03.670 --> 00:05:06.490 practice? It figures out your core values, your 00:05:06.490 --> 00:05:09.149 tone of voice, even suggests fonts and colors. 00:05:09.310 --> 00:05:11.949 Then it spits out campaign ideas and generates 00:05:11.949 --> 00:05:14.750 five or ten social media posts that actually 00:05:14.750 --> 00:05:17.589 match your brand's aesthetic. That's a huge head 00:05:17.589 --> 00:05:20.509 start on a new campaign. Saves days of brainstorming. 00:05:20.529 --> 00:05:22.949 Then there's Opal, which is more for visual thinkers. 00:05:23.050 --> 00:05:24.850 It doesn't even use text prompts. Right. Opal 00:05:24.850 --> 00:05:27.089 uses what it calls visual nodes. Think of it 00:05:27.089 --> 00:05:29.949 like... building a flow chart for your workflow, 00:05:30.089 --> 00:05:32.970 kind of like the toolmake .com, but for creating 00:05:32.970 --> 00:05:35.449 things with AI. You just drag and connect nodes 00:05:35.449 --> 00:05:38.730 like research AI news, which connects to generate 00:05:38.730 --> 00:05:40.910 images, which connects to generate web page. 00:05:41.089 --> 00:05:43.750 So it provides a kind of visual logic. You can 00:05:43.750 --> 00:05:45.649 just adjust the whole process by dragging the 00:05:45.649 --> 00:05:48.509 connections around. Exactly. So how should someone 00:05:48.509 --> 00:05:50.910 who's trying to be pragmatic approach these tools 00:05:50.910 --> 00:05:52.569 in labs? Can you trust them for professional 00:05:52.569 --> 00:05:54.980 work? I'd say use them as powerful starting points 00:05:54.980 --> 00:05:58.220 for complex tasks. They're awesome, highly functional 00:05:58.220 --> 00:06:01.839 solutions to real problems right now. All right, 00:06:01.860 --> 00:06:03.819 now let's get into the standard Gemini chat bot 00:06:03.819 --> 00:06:07.240 itself. Most people, I think, barely scratch 00:06:07.240 --> 00:06:09.360 the surface of this interface. They're missing 00:06:09.360 --> 00:06:12.279 all the hidden superpowers. We have to talk about 00:06:12.279 --> 00:06:14.079 the video analysis. That just sounds like an 00:06:14.079 --> 00:06:16.740 incredible technical feat. It really is. You 00:06:16.740 --> 00:06:18.959 can drag and drop a video file right into the 00:06:18.959 --> 00:06:22.100 chat and Gemini can analyze it. The big use case 00:06:22.100 --> 00:06:24.279 here is reverse engineering. You can literally 00:06:24.279 --> 00:06:27.300 ask the AI, watch this video, now give me a text 00:06:27.300 --> 00:06:29.860 -to -video prompt to recreate it. Wait a second. 00:06:29.899 --> 00:06:32.779 So if I see some amazing viral AI clip, I can 00:06:32.779 --> 00:06:35.660 just give it to Gemini and it analyzes everything, 00:06:35.920 --> 00:06:38.420 the composition, camera angles, lighting, dialogue, 00:06:38.519 --> 00:06:40.360 and it just spits out the rest. It gives you 00:06:40.360 --> 00:06:42.660 the full recipe. The real world test from the 00:06:42.660 --> 00:06:44.980 source material was recreating this viral video 00:06:44.980 --> 00:06:47.639 of a woman and a dinosaur. The prompt it generated 00:06:47.639 --> 00:06:51.000 led to a near perfect recreation in VO 3 .1. 00:06:51.180 --> 00:06:54.600 You just have to stop and think, whoa. Wow. Imagine 00:06:54.600 --> 00:06:57.199 scaling that. It turns every video you admire 00:06:57.199 --> 00:06:59.319 into a free masterclass in prompt engineering. 00:06:59.579 --> 00:07:02.319 That is just a staggering capability for a free 00:07:02.319 --> 00:07:06.759 tool. Okay, so next on the list, gems. What exactly 00:07:06.759 --> 00:07:09.899 are these custom Gemini versions? So gems are 00:07:09.899 --> 00:07:12.920 basically personalized, reusable AI assistance. 00:07:13.420 --> 00:07:16.980 They permanently remember a complex set of instructions. 00:07:17.240 --> 00:07:19.699 Think of them like a permanent system prompt 00:07:19.699 --> 00:07:21.620 you can create in about two minutes. And the 00:07:21.620 --> 00:07:24.300 pro tip here is to let Gemini itself rewrite 00:07:24.300 --> 00:07:26.779 your simple instructions into that more comprehensive 00:07:26.779 --> 00:07:29.360 prompt structure. Absolutely. You give it the 00:07:29.360 --> 00:07:31.300 basic idea and it'll flesh it out into a really 00:07:31.300 --> 00:07:33.259 detailed technical instruction set for itself. 00:07:33.819 --> 00:07:35.699 And the real power multiplier is when you add 00:07:35.699 --> 00:07:37.759 knowledge bases. Knowledge bases. You mean I 00:07:37.759 --> 00:07:40.199 can upload my own stuff to teach the gem of my 00:07:40.199 --> 00:07:42.600 specific style? That's it. You upload two or 00:07:42.600 --> 00:07:44.500 three of your best performing scripts or maybe 00:07:44.500 --> 00:07:47.379 some reports, and you can train a script writing 00:07:47.379 --> 00:07:49.939 gem to write in your specific voice and style. 00:07:50.120 --> 00:07:52.180 It just eliminates all that repetitive setup 00:07:52.180 --> 00:07:54.819 time. Okay, let's pivot to data analysis. We 00:07:54.819 --> 00:07:56.939 usually associate that with expensive, specialized 00:07:56.939 --> 00:07:59.959 software. What can Gemini do for just the average 00:07:59.959 --> 00:08:03.240 person with a spreadsheet? Gemini is an incredibly 00:08:03.240 --> 00:08:07.079 powerful no -code data analyst. You can upload 00:08:07.079 --> 00:08:10.819 a standard dataset as CSV or an Excel file and 00:08:10.819 --> 00:08:13.480 just make a request in natural language. No more 00:08:13.480 --> 00:08:16.060 pivot tables. So you could just ask it for, say, 00:08:16.279 --> 00:08:19.120 five clear visualizations that highlight interesting 00:08:19.120 --> 00:08:21.939 patterns. The focus is on the insight, not the 00:08:21.939 --> 00:08:23.920 chart itself. Exactly. And you're not just getting 00:08:23.920 --> 00:08:25.519 a chart back that you have to figure out on your 00:08:25.519 --> 00:08:28.420 own. No, it generates scatter plots, bar charts, 00:08:28.740 --> 00:08:31.500 correlation matrices. And this is the critical 00:08:31.500 --> 00:08:33.860 part. It provides a plain English explanation 00:08:33.860 --> 00:08:36.720 of why the data matters. It just lowers the barrier 00:08:36.720 --> 00:08:39.440 to exploring data so much. And finally, we mentioned 00:08:39.440 --> 00:08:41.440 nano banana earlier, but how does that actually 00:08:41.440 --> 00:08:44.740 work inside? chat for editing images yeah nano 00:08:44.740 --> 00:08:46.919 bananas right there in the chat the source used 00:08:46.919 --> 00:08:49.100 a screenshot from a movie and prompted change 00:08:49.100 --> 00:08:51.679 this to sunrise the result was instantaneous 00:08:51.679 --> 00:08:54.480 and it looked totally realistic it kept the composition 00:08:54.480 --> 00:08:57.019 but perfectly adjusted all the lighting and shadows 00:08:57.019 --> 00:09:00.200 that's amazing and for really complex stuff you 00:09:00.200 --> 00:09:02.659 can use a tool in labs called flow which combines 00:09:02.659 --> 00:09:05.220 imaging for creation nano banana for editing 00:09:05.220 --> 00:09:08.779 and vo for animation all in one workflow So if 00:09:08.779 --> 00:09:11.059 a user is feeling totally overwhelmed by all 00:09:11.059 --> 00:09:14.059 these new tools, what's the one hidden superpower 00:09:14.059 --> 00:09:16.399 that's going to provide the biggest daily time 00:09:16.399 --> 00:09:19.860 -saving potential? Oh, gems, hands down. They 00:09:19.860 --> 00:09:22.159 eliminate that repeated setup for personalized, 00:09:22.200 --> 00:09:25.080 repetitive workflows. Reducing that friction 00:09:25.080 --> 00:09:27.220 just makes all the difference. I mean, we've 00:09:27.220 --> 00:09:29.299 covered a really remarkable breadth of tools 00:09:29.299 --> 00:09:31.519 here, from the simple no -code apps in Canvas 00:09:31.519 --> 00:09:33.679 to these advanced specialized assistants with 00:09:33.679 --> 00:09:36.440 gems and even deep analysis of video and data. 00:09:36.960 --> 00:09:39.480 Yeah, the core insight seems to be that Google's 00:09:39.480 --> 00:09:42.379 AI power is way more decentralized and, importantly, 00:09:42.600 --> 00:09:45.419 free than most of us realize. It's offering the 00:09:45.419 --> 00:09:47.440 power of a whole automation suite without the 00:09:47.440 --> 00:09:49.460 price tag. And these aren't just cool demos. 00:09:49.580 --> 00:09:51.600 They're genuinely production ready. You got the 00:09:51.600 --> 00:09:53.679 roadmap now. So we invite you to go experiment 00:09:53.679 --> 00:09:55.679 with one of these this week. Go create your first 00:09:55.679 --> 00:09:58.440 gem that's customized for your job. Or, you know, 00:09:58.460 --> 00:10:00.580 run a simple data analysis on an old spreadsheet 00:10:00.580 --> 00:10:03.539 you've been ignoring. But remember, we only covered 00:10:03.539 --> 00:10:06.100 part one here. The source material hints at things 00:10:06.100 --> 00:10:08.059 coming down the line like Notebook LM's podcast 00:10:08.059 --> 00:10:10.600 mode and Gemini Live, which can actually see 00:10:10.600 --> 00:10:12.919 your screen and help you. So the question is, 00:10:12.940 --> 00:10:15.620 how much more essential knowledge is hidden in 00:10:15.620 --> 00:10:17.940 the everyday products you already use? And that's 00:10:17.940 --> 00:10:19.259 what we need to unpack next time.