WEBVTT 00:00:00.000 --> 00:00:02.580 I want you to picture something. You own a Ferrari. 00:00:03.799 --> 00:00:07.459 It's this marvel of engineering, incredible speed, 00:00:07.719 --> 00:00:11.880 precision, just raw power. Right. But every single 00:00:11.880 --> 00:00:15.339 day, you only use it to drive two blocks to the 00:00:15.339 --> 00:00:17.699 grocery store. That is the perfect analogy. I 00:00:17.699 --> 00:00:20.160 mean, we have this engine of like infinite potential 00:00:20.160 --> 00:00:22.300 sitting right on our desktops, and we're just 00:00:22.300 --> 00:00:25.019 using it to write out of office emails. Today, 00:00:25.160 --> 00:00:27.100 we are going to stop driving to the grocery store. 00:00:27.440 --> 00:00:30.739 Welcome to the Deep Dive. It is January 21st, 00:00:30.739 --> 00:00:35.299 2026. And we are getting an early look at a guide 00:00:35.299 --> 00:00:37.679 dropping tomorrow. It's called 12 Advanced Chat 00:00:37.679 --> 00:00:40.659 GPT Tricks for Power Users. And it's basically 00:00:40.659 --> 00:00:43.100 the manual for the Ferrari that nobody ever bothered 00:00:43.100 --> 00:00:45.359 to read. The premise here is interesting. The 00:00:45.359 --> 00:00:47.619 author argues there's this huge gap between the 00:00:47.619 --> 00:00:49.909 tourist and the wizard. The tourist asks for 00:00:49.909 --> 00:00:52.509 a summary. The wizard, well, what is the wizard 00:00:52.509 --> 00:00:54.130 doing differently? Oh, the wizard is playing 00:00:54.130 --> 00:00:55.750 a completely different game. They're building 00:00:55.750 --> 00:00:57.929 memory systems. They have incognito modes for 00:00:57.929 --> 00:00:59.729 their thoughts. They have AI agents building 00:00:59.729 --> 00:01:01.789 software for them, you know, while they sleep. 00:01:01.950 --> 00:01:04.109 So let's bridge that gap. We have 12 tricks to 00:01:04.109 --> 00:01:06.310 get through today. Let's start with section one, 00:01:06.430 --> 00:01:08.569 which is really about the mindset. The guide 00:01:08.569 --> 00:01:12.310 says that even now, in 2026, something like 95 00:01:12.310 --> 00:01:15.450 % of users are missing these hidden features. 00:01:15.650 --> 00:01:18.629 They're still acting like it's 2023. Right. If 00:01:18.629 --> 00:01:21.150 you're still typing, summarize this or write 00:01:21.150 --> 00:01:23.269 me a blog post, you're essentially eating AI 00:01:23.269 --> 00:01:26.909 fast food. The system can do it, sure, but there's 00:01:26.909 --> 00:01:30.230 no depth. A power user knows the tool isn't just 00:01:30.230 --> 00:01:32.530 a chatbot anymore. It's more like an operating 00:01:32.530 --> 00:01:35.049 system for your work. It makes you wonder, though. 00:01:35.129 --> 00:01:37.569 We have this tool that can simulate complex reasoning, 00:01:37.730 --> 00:01:40.430 yet we just default to the easiest path. Why 00:01:40.430 --> 00:01:42.689 do we settle for surface -level use? We simply 00:01:42.689 --> 00:01:44.609 don't know what we don't know yet. Fair point. 00:01:45.129 --> 00:01:47.450 Okay, let's get into the mechanics. Trick number 00:01:47.450 --> 00:01:50.530 one is context -specific writing memory. This 00:01:50.530 --> 00:01:52.989 one solves the generic voice problem. So most 00:01:52.989 --> 00:01:55.689 people have one setting, one voice for ChatGPT, 00:01:55.790 --> 00:01:57.609 but you don't talk to your boss the same way 00:01:57.609 --> 00:01:59.890 you talk to your team. Hopefully not. That could 00:01:59.890 --> 00:02:02.420 get awkward. Exactly. So the trick here is you 00:02:02.420 --> 00:02:05.480 stop using a generic instruction like be professional. 00:02:05.859 --> 00:02:08.500 Instead, you feed the AI actual examples. You 00:02:08.500 --> 00:02:10.840 paste in an email you wrote and say, analyze 00:02:10.840 --> 00:02:12.960 this for tone and style. And here's the key part. 00:02:13.180 --> 00:02:17.199 You then say, store this as team emails. So instead 00:02:17.199 --> 00:02:19.659 of prompting it every time, it just knows. If 00:02:19.659 --> 00:02:21.960 I say I'm writing to the team, it's casual. If 00:02:21.960 --> 00:02:25.360 it's for a client, it's polished. Yes. You build 00:02:25.360 --> 00:02:28.400 out these personas, a client persona that's super 00:02:28.400 --> 00:02:31.080 formal, a LinkedIn persona that's more punchy. 00:02:31.099 --> 00:02:34.039 The AI just remembers which mask to wear depending 00:02:34.039 --> 00:02:36.199 on the context. It does raise a question about 00:02:36.199 --> 00:02:38.729 authenticity. If I have an algorithm managing 00:02:38.729 --> 00:02:41.909 my personas, does this erase our authentic voice? 00:02:42.110 --> 00:02:44.270 No, it just scales our different personas efficiently. 00:02:44.710 --> 00:02:47.469 Okay, trick number two shifts from how we speak 00:02:47.469 --> 00:02:49.930 to how we listen. It's called scheduled daily 00:02:49.930 --> 00:02:52.090 reports. Yeah, this is where we see the big shift 00:02:52.090 --> 00:02:54.669 from reactive to proactive. I mean, usually you 00:02:54.669 --> 00:02:56.629 have to go to ChatGPT and ask it for something. 00:02:56.909 --> 00:03:00.090 This lets the AI nudge you. So we're inviting 00:03:00.090 --> 00:03:03.310 the machine to nag us now. In a useful way. You 00:03:03.310 --> 00:03:05.949 set a trigger. Every morning at 8 a .m., send 00:03:05.949 --> 00:03:08.810 me updates on OpenAI, Google, and Anthropic, 00:03:09.009 --> 00:03:12.310 and bam, a desktop alert. Or you could have it 00:03:12.310 --> 00:03:14.689 send you a weekly goal review every Friday afternoon. 00:03:15.009 --> 00:03:16.969 It's interesting. We spend so much of our digital 00:03:16.969 --> 00:03:19.389 lives trying to filter out notifications, and 00:03:19.389 --> 00:03:21.650 now we're programming a machine to send us more 00:03:21.650 --> 00:03:24.530 of them. Does this reduce the noise or just add 00:03:24.530 --> 00:03:27.349 to it? It filters the noise so you focus on signal. 00:03:27.610 --> 00:03:30.629 Let's move to trick number three, project -only 00:03:30.629 --> 00:03:33.629 memory. This sounds like a privacy feature. It 00:03:33.629 --> 00:03:36.330 is, but it's also a sanity feature. By default, 00:03:36.509 --> 00:03:39.310 ChatGPT remembers everything account -wide. So 00:03:39.310 --> 00:03:41.389 if you tell it about your dog's birthday in one 00:03:41.389 --> 00:03:43.810 chat, it technically knows that when you're writing 00:03:43.810 --> 00:03:45.830 a legal brief in another. Which is not great, 00:03:45.949 --> 00:03:47.710 especially if you're juggling multiple clients. 00:03:47.889 --> 00:03:50.370 Right. So project -only memory is a toggle. You 00:03:50.370 --> 00:03:52.189 find it in the gear icon when you start a new 00:03:52.189 --> 00:03:54.849 project, and it builds a walled garden. That 00:03:54.849 --> 00:03:57.189 project can only see the files and chats inside 00:03:57.189 --> 00:04:00.060 of it. Nothing else. So my data for client A 00:04:00.060 --> 00:04:02.580 and client B will never, ever cross -pollinate. 00:04:03.020 --> 00:04:04.800 Precisely. And the guide points out something 00:04:04.800 --> 00:04:07.639 really critical. You have to make this choice 00:04:07.639 --> 00:04:10.460 at the start. You can't change it later. It seems 00:04:10.460 --> 00:04:12.400 like we're constantly negotiating these boundaries 00:04:12.400 --> 00:04:15.199 with our tools. Is privacy the ultimate luxury 00:04:15.199 --> 00:04:18.699 in AI? Compartmentalization is safety. Leakage 00:04:18.699 --> 00:04:22.579 is unprofessional. Moving on. Trick number four. 00:04:22.879 --> 00:04:26.379 New personalization settings. This is the death 00:04:26.379 --> 00:04:29.360 of the old text box. You know, we used to have 00:04:29.360 --> 00:04:31.639 to type out, please be professional, or explain 00:04:31.639 --> 00:04:34.360 this like I'm five. Now, under settings, there 00:04:34.360 --> 00:04:37.079 are just drop -down presets. Presets, like filters 00:04:37.079 --> 00:04:39.759 on a photo. Exactly like that. Friendly, candid, 00:04:39.860 --> 00:04:42.560 professional, you just click. But the real power 00:04:42.560 --> 00:04:44.920 move is stacking them. You choose the friendly 00:04:44.920 --> 00:04:47.360 preset. Then you fill out the about you section 00:04:47.360 --> 00:04:49.699 with your specific goals. And then you use those 00:04:49.699 --> 00:04:51.480 memory tricks we just talked about. It's all 00:04:51.480 --> 00:04:53.560 about layering context. It's about alignment. 00:04:53.680 --> 00:04:56.600 The tool stops feeling like a generic robot and 00:04:56.600 --> 00:04:58.480 starts feeling like a colleague who actually 00:04:58.480 --> 00:05:01.000 gets you. I find those labels so interesting. 00:05:01.759 --> 00:05:04.959 Friendly. Why do we feel this need to anthropomorphize 00:05:04.959 --> 00:05:07.819 the tool with friendliness? Because we work better 00:05:07.819 --> 00:05:10.639 when the vibe matches ours. All right. Trick 00:05:10.639 --> 00:05:14.449 number five. Temporary chats. This is basically 00:05:14.449 --> 00:05:17.889 the incognito mode for AI. There's a little icon 00:05:17.889 --> 00:05:20.329 in the top corner, a chat bubble with a slash 00:05:20.329 --> 00:05:22.589 through it. I see. And this does what exactly? 00:05:22.930 --> 00:05:26.649 When you click it, nothing is saved. No history, 00:05:26.810 --> 00:05:29.610 no memory is trained on it. It's just gone when 00:05:29.610 --> 00:05:31.970 you close the tab. What's the use case for that? 00:05:32.209 --> 00:05:34.730 Well, hypotheticals. Let's say you're a CEO. 00:05:34.990 --> 00:05:37.209 You might want to ask, if I were a competitor 00:05:37.209 --> 00:05:39.379 to my own company, How would I put us out of 00:05:39.379 --> 00:05:41.319 business? You do not want that stored in your 00:05:41.319 --> 00:05:44.180 long -term account memory. Right. Or if you just 00:05:44.180 --> 00:05:46.459 need to do a quick one -off math problem. Exactly. 00:05:46.519 --> 00:05:48.139 You don't want a random question about baking 00:05:48.139 --> 00:05:50.399 soda to influence your coding advice three weeks 00:05:50.399 --> 00:05:52.639 later. It suggests that memory can be a burden. 00:05:52.759 --> 00:05:55.980 So why is forgetting just as important as remembering? 00:05:56.420 --> 00:05:58.959 Freedom to experiment without polluting the long 00:05:58.959 --> 00:06:02.180 -term data. Trick number six. Company knowledge 00:06:02.180 --> 00:06:04.699 base. Now we're getting into the enterprise stuff. 00:06:04.800 --> 00:06:06.699 Yeah, this is for the business plan users. It 00:06:06.699 --> 00:06:09.379 connects ChatGPT to everything. Your Google Drive, 00:06:09.579 --> 00:06:12.180 your Slack, your HubSpot, everything. So it's 00:06:12.180 --> 00:06:14.259 not just searching the public web. It's searching 00:06:14.259 --> 00:06:16.540 our internal world. Right. You can ask, what 00:06:16.540 --> 00:06:18.480 meetings do I have tomorrow? And can you find 00:06:18.480 --> 00:06:20.540 the related files in Dropbox? It'll scan your 00:06:20.540 --> 00:06:23.459 calendar, find the meeting, go into Dropbox, 00:06:23.579 --> 00:06:26.000 find the right PDF, and summarize it for you. 00:06:26.120 --> 00:06:28.720 That used to be a 20 -minute scramble. Where 00:06:28.720 --> 00:06:31.730 is that file? Was it in Slack? Was it in an email? 00:06:31.949 --> 00:06:34.970 Now it's one prompt. It turns the AI into a single 00:06:34.970 --> 00:06:37.089 search guard for your entire work life. It's 00:06:37.089 --> 00:06:39.850 efficient for sure. But are we outsourcing our 00:06:39.850 --> 00:06:42.529 organizational brain? We're reclaiming time spent 00:06:42.529 --> 00:06:45.029 hunting for files. Next up, trick number seven, 00:06:45.189 --> 00:06:50.250 chat GPT apps. This is the ecosystem play. You 00:06:50.250 --> 00:06:52.250 know how you tag someone on social media with 00:06:52.250 --> 00:06:55.490 the at tip symbol? You can now do that with software 00:06:55.490 --> 00:06:57.930 right inside the chat. Give me an example. How 00:06:57.930 --> 00:07:00.410 does that work? So you type at Photoshop in the 00:07:00.410 --> 00:07:02.810 chat bar. You say, create a social media graphic 00:07:02.810 --> 00:07:05.089 for this post. It doesn't just make a flat image. 00:07:05.149 --> 00:07:07.350 It connects to Photoshop. And you can even edit 00:07:07.350 --> 00:07:10.410 the properties. Or at Notion to save your chat 00:07:10.410 --> 00:07:12.949 notes directly to a workspace. The lines between 00:07:12.949 --> 00:07:14.870 the apps are just dissolving. It's all becoming 00:07:14.870 --> 00:07:18.110 conversation. The chat bar is becoming the universal 00:07:18.110 --> 00:07:20.769 interface. You don't open apps anymore. You just 00:07:20.769 --> 00:07:22.829 call them when you need them. So is the interface 00:07:22.829 --> 00:07:25.589 itself just dissolving into conversation? Yes. 00:07:25.709 --> 00:07:28.310 The chat bar is the new operating system. Trick 00:07:28.310 --> 00:07:31.350 number eight, updating custom GPTs. This is a 00:07:31.350 --> 00:07:33.930 classic maintenance issue. You might have built 00:07:33.930 --> 00:07:36.990 an amazing custom GPT six months ago when it 00:07:36.990 --> 00:07:40.500 was running on, say, GPT -4. But now we have 00:07:40.500 --> 00:07:44.000 GPT 5 .2. And for anyone who doesn't know, a 00:07:44.000 --> 00:07:47.220 custom GPT is basically a version of ChatGPT 00:07:47.220 --> 00:07:49.339 you've trained for a specific task. Correct. 00:07:49.540 --> 00:07:51.899 But here's the catch. They don't always update 00:07:51.899 --> 00:07:54.240 themselves. You have to physically go in, click 00:07:54.240 --> 00:07:57.810 Edit, and manually select the newest model. Otherwise, 00:07:57.889 --> 00:08:00.189 you've got that Ferrari, but with a lawnmower 00:08:00.189 --> 00:08:02.029 engine inside. And you can even do this with 00:08:02.029 --> 00:08:04.329 GPTs that other people have built. Yep, there's 00:08:04.329 --> 00:08:06.290 a drop -down. If you're using a public one that 00:08:06.290 --> 00:08:08.589 feels a bit sluggish, check the model. You can 00:08:08.589 --> 00:08:10.350 often override it to the latest version yourself. 00:08:10.730 --> 00:08:13.329 It's like digital hygiene. Why is maintenance 00:08:13.329 --> 00:08:16.069 so often ignored in our digital tools? Invisible 00:08:16.069 --> 00:08:18.769 decay is easy to miss until performance drops. 00:08:19.050 --> 00:08:22.209 Trick number nine, batch image generation. This 00:08:22.209 --> 00:08:25.220 one is all about speed, pure speed. DLE is now, 00:08:25.420 --> 00:08:28.879 what, 4x faster in this 2026 update? So you don't 00:08:28.879 --> 00:08:31.579 ask for one logo idea, you ask for four at once. 00:08:32.440 --> 00:08:35.980 Minimalist, bold, abstract, and gradient. And 00:08:35.980 --> 00:08:38.259 it generates them all simultaneously. All at 00:08:38.259 --> 00:08:40.940 once. It completely changes the creative workflow. 00:08:41.120 --> 00:08:44.559 It's not... Wait, critique anymore. It's generate 00:08:44.559 --> 00:08:47.679 a batch, pick the best one, and refine from there. 00:08:47.879 --> 00:08:50.620 But does that speed devalue the creative choice? 00:08:51.000 --> 00:08:54.139 Speed allows for rapid iteration and better final 00:08:54.139 --> 00:08:57.620 choices. Now, trick number 10, canvas mode mini 00:08:57.620 --> 00:09:00.179 apps. I have to admit, when I was reading the 00:09:00.179 --> 00:09:01.440 source material, this is the one that really 00:09:01.440 --> 00:09:03.580 stopped me in my tracks. Oh, this is the wow 00:09:03.580 --> 00:09:05.799 moment. This is the one that really makes you 00:09:05.799 --> 00:09:08.870 feel like a wizard. Whoa. Just to define it for 00:09:08.870 --> 00:09:10.909 everyone listening, Canvas mode is a split screen 00:09:10.909 --> 00:09:13.409 that shows you code on one side and a live app 00:09:13.409 --> 00:09:15.450 on the other. Exactly. And you don't need to 00:09:15.450 --> 00:09:17.009 know a single line of code. You can just say, 00:09:17.029 --> 00:09:18.909 create a to -do list with a dark mode, toggle, 00:09:19.110 --> 00:09:21.990 and local storage. And it just builds it. It 00:09:21.990 --> 00:09:24.970 just appears. It appears. Code on the left, a 00:09:24.970 --> 00:09:27.389 fully working app on the right. You can click 00:09:27.389 --> 00:09:29.269 the buttons. It works. You can build mortgage 00:09:29.269 --> 00:09:32.350 calculators, little games, productivity timers. 00:09:32.769 --> 00:09:34.590 And then you can just download the code and put 00:09:34.590 --> 00:09:36.990 it on your website. That's profound. We used 00:09:36.990 --> 00:09:38.769 to need a team of engineers for that. Now all 00:09:38.769 --> 00:09:40.789 you need is a sentence. What happens when creation 00:09:40.789 --> 00:09:43.190 has zero friction? Everyone becomes a builder. 00:09:43.990 --> 00:09:46.409 Ideas become reality instantly. Trick number 00:09:46.409 --> 00:09:51.330 11. Deep research plus custom knowledge. Okay, 00:09:51.350 --> 00:09:53.370 so... Deep research has been around for a little 00:09:53.370 --> 00:09:56.649 while. It scans the web. It reads articles. But 00:09:56.649 --> 00:09:58.590 the real trick is the combination. You upload 00:09:58.590 --> 00:10:01.009 your own files first. So you're anchoring the 00:10:01.009 --> 00:10:03.409 research in your own context. Yes. Let's say 00:10:03.409 --> 00:10:05.529 you're researching market pricing. You upload 00:10:05.529 --> 00:10:07.730 your own internal pricing model spreadsheet. 00:10:08.629 --> 00:10:12.210 Then you say research competitor pricing. It 00:10:12.210 --> 00:10:14.450 doesn't just give you a generic list. It compares 00:10:14.450 --> 00:10:17.649 the world's data directly against your data. 00:10:17.809 --> 00:10:19.769 It's synthesizing the external with the internal. 00:10:19.950 --> 00:10:22.529 Context is king. Right. It stops giving you generic 00:10:22.529 --> 00:10:24.809 advice because now it knows your specific constraints. 00:10:25.090 --> 00:10:27.909 So is this the end of generic advice? Context 00:10:27.909 --> 00:10:31.190 is king. Generic answers are now obsolete. And 00:10:31.190 --> 00:10:34.389 finally, trick number 12, extended thinking mode. 00:10:34.850 --> 00:10:37.429 This is one I struggle with personally. Oh, so? 00:10:37.690 --> 00:10:40.950 And I think a lot of us do this. I get lazy. 00:10:41.330 --> 00:10:44.169 I have this bad habit of just leaving the model 00:10:44.169 --> 00:10:47.139 on auto. I'll ask it a really complex question, 00:10:47.259 --> 00:10:49.320 get a mediocre answer, and then I spend the next 00:10:49.320 --> 00:10:51.559 20 minutes trying to fix it. I call it prompt 00:10:51.559 --> 00:10:53.519 drift. I think we've all been there. I know I 00:10:53.519 --> 00:10:55.799 have. You just want the answer now, so you try 00:10:55.799 --> 00:10:57.820 to coax intelligence out of the faster mode. 00:10:57.980 --> 00:11:00.759 Right. The trick is to manually select extended 00:11:00.759 --> 00:11:04.100 thinking mode. It forces the AI to pause. It 00:11:04.100 --> 00:11:06.379 actually thinks for, you know, 30 seconds to 00:11:06.379 --> 00:11:09.159 two minutes. It plans, and it checks its work 00:11:09.159 --> 00:11:10.799 before it gives you an answer. The difference 00:11:10.799 --> 00:11:13.200 between a reflex and a reflection. Exactly. For 00:11:13.200 --> 00:11:16.149 strategy, for architecture. For any complex problem, 00:11:16.389 --> 00:11:18.649 you have to let it think. There's even a heavy 00:11:18.649 --> 00:11:21.110 mode for pro users that thinks for even longer. 00:11:21.350 --> 00:11:24.090 We're so used to instant gratification. Is patience 00:11:24.090 --> 00:11:26.769 the new skill we need to relearn? Waiting 30 00:11:26.769 --> 00:11:29.190 seconds beats an hour of correcting errors. And 00:11:29.190 --> 00:11:31.769 there's a bonus trick in the guide, chat branching. 00:11:31.830 --> 00:11:33.809 This is for all the non -linear thinkers out 00:11:33.809 --> 00:11:36.389 there. You're in a conversation and you get to 00:11:36.389 --> 00:11:38.889 a fork in the road. Should I write a funny email 00:11:38.889 --> 00:11:41.250 or a serious one? And usually you have to pick 00:11:41.250 --> 00:11:45.220 a path. Not anymore. You click. branch into new 00:11:45.220 --> 00:11:48.419 chat and suddenly you have two parallel timelines 00:11:48.419 --> 00:11:51.500 so just so we're clear branting means creating 00:11:51.500 --> 00:11:54.820 a new conversation path from an existing message 00:11:54.820 --> 00:11:57.799 right in one timeline you explore the funny email 00:11:57.799 --> 00:12:00.399 in the other you explore the serious one you 00:12:00.399 --> 00:12:02.559 can explore both realities at the same time without 00:12:02.559 --> 00:12:04.879 losing your original place it's like a multiverse 00:12:04.879 --> 00:12:07.519 of thought and you can save these branches into 00:12:07.519 --> 00:12:09.820 a project to synthesize them later it's like 00:12:09.820 --> 00:12:12.799 a whiteboard strategy session but in text form 00:12:13.210 --> 00:12:15.129 Does this finally reflect how the human mind 00:12:15.129 --> 00:12:17.690 actually works? Absolutely. Nonlinear thinking 00:12:17.690 --> 00:12:20.110 is finally supported by the tool. Mid -role sponsor 00:12:20.110 --> 00:12:22.830 placeholder. So let's just try to unpack all 00:12:22.830 --> 00:12:25.769 of this. We've covered 12 tricks. Memory injection, 00:12:26.210 --> 00:12:29.730 proactive scheduling, privacy walls, app integrations, 00:12:29.830 --> 00:12:32.450 these deep thinking modes. It's a lot. It can 00:12:32.450 --> 00:12:35.070 definitely feel overwhelming. It is. But if you 00:12:35.070 --> 00:12:37.539 zoom out and look at the big picture. The theme 00:12:37.539 --> 00:12:39.740 isn't really about the software. The software 00:12:39.740 --> 00:12:42.240 is just the engine. The real theme is about the 00:12:42.240 --> 00:12:44.899 driver. Exactly. The tool didn't get smarter 00:12:44.899 --> 00:12:47.600 between yesterday and today. You get smarter 00:12:47.600 --> 00:12:50.720 by knowing these buttons exist. We started with 00:12:50.720 --> 00:12:52.899 that Ferrari analogy. The difference between 00:12:52.899 --> 00:12:54.759 the person going to the grocery store and the 00:12:54.759 --> 00:12:58.059 race car driver isn't the car itself. It's the 00:12:58.059 --> 00:13:01.179 knowledge of how to shift the gears. And knowing 00:13:01.179 --> 00:13:03.000 that you don't have to settle for the default 00:13:03.000 --> 00:13:05.559 settings, you can mold this thing to be your 00:13:05.559 --> 00:13:08.919 perfect co -pilot. So here's my challenge to 00:13:08.919 --> 00:13:10.799 you as you're listening right now. Don't try 00:13:10.799 --> 00:13:13.360 to do all 12 of these today. You'll just burn 00:13:13.360 --> 00:13:16.480 out. Pick one. Just one. Maybe it's setting up 00:13:16.480 --> 00:13:19.240 the about you personalization. Maybe it's trying 00:13:19.240 --> 00:13:21.460 that canvas mode to build a simple calculator. 00:13:21.799 --> 00:13:23.759 Yeah, just try one thing. See how it changes 00:13:23.759 --> 00:13:26.000 the friction in your day -to -day workflow. The 00:13:26.000 --> 00:13:28.120 gap between the tourist and the wizard isn't 00:13:28.120 --> 00:13:31.659 intelligence. It's just curiosity. Go find the 00:13:31.659 --> 00:13:33.320 hidden buttons. See you in the deep end.