WEBVTT 00:00:00.000 --> 00:00:02.160 We spent so much time watching the horizon for 00:00:02.160 --> 00:00:05.759 the next big AI thing, right? The GPT -5 hype, 00:00:05.860 --> 00:00:08.839 the massive model announcements. We do. But honestly, 00:00:09.019 --> 00:00:11.800 I find the most impactful upgrades aren't always 00:00:11.800 --> 00:00:15.980 these huge version jumps. They're often these 00:00:15.980 --> 00:00:20.320 quieter, subtle feature additions, things that 00:00:20.320 --> 00:00:22.280 are already live in the system you're probably 00:00:22.280 --> 00:00:24.359 using. And these are the ones that really change 00:00:24.359 --> 00:00:27.239 the interaction, I suppose, moving AI from being 00:00:27.239 --> 00:00:29.679 like a slightly smarter search engine. Exactly. 00:00:29.820 --> 00:00:33.299 to a true partner in your workflow in learning. 00:00:34.060 --> 00:00:35.640 So today, we're going to do a deep dive into 00:00:35.640 --> 00:00:37.579 those kinds of updates. We're looking at three 00:00:37.579 --> 00:00:41.039 specific new features designed to make getting 00:00:41.039 --> 00:00:44.079 knowledge and handling complex tasks vastly simpler 00:00:44.079 --> 00:00:45.700 for you, provided, of course, you've got that 00:00:45.700 --> 00:00:47.920 plus subscription turned on. That's kind of key 00:00:47.920 --> 00:00:50.280 for these power tools. Gotcha. So it's like turning 00:00:50.280 --> 00:00:53.439 your AI from that smart friend who just answers 00:00:53.439 --> 00:00:56.770 questions into maybe a proactive coworker. Or 00:00:56.770 --> 00:00:59.270 even need teacher. A proactive co -worker and 00:00:59.270 --> 00:01:02.329 a surprisingly patient adaptive teacher. That's 00:01:02.329 --> 00:01:05.049 a great way to put it. So our roadmap today covers 00:01:05.049 --> 00:01:08.329 three main things. First, this personalized learning 00:01:08.329 --> 00:01:11.109 thing called study mode. OK. Then the really 00:01:11.109 --> 00:01:13.650 powerful workflow automating agent mode. Agent 00:01:13.650 --> 00:01:16.930 mode, right. And finally, how chat GPT is now 00:01:16.930 --> 00:01:18.670 starting to connect directly with your everyday 00:01:18.670 --> 00:01:21.189 apps like Google Calendar and Gmail. OK, yeah, 00:01:21.189 --> 00:01:22.829 let's unpack this. Let's start with that first 00:01:22.829 --> 00:01:25.530 one. Study and learn mode. Right. I can see this 00:01:25.530 --> 00:01:29.430 being huge for anyone facing that mountain of 00:01:29.430 --> 00:01:32.909 information, trying to tackle a new complex topic, 00:01:33.750 --> 00:01:36.950 say, macroeconomics or something. The sheer volume 00:01:36.950 --> 00:01:40.489 online can be, well, paralyzing. It absolutely 00:01:40.489 --> 00:01:43.109 creates instant overload. And this mode, I think, 00:01:43.290 --> 00:01:45.390 is a pretty elegant solution. It kind of eliminates 00:01:45.390 --> 00:01:47.450 that anxiety of just starting a complex topic. 00:01:47.650 --> 00:01:49.989 How so? Well, it doesn't just dump a long article 00:01:49.989 --> 00:01:52.430 on you. It acts more like your own personal structured 00:01:52.430 --> 00:01:54.950 teacher. and that's a major difference. Tell 00:01:54.950 --> 00:01:56.409 me more about that structure then. How does it 00:01:56.409 --> 00:01:59.450 work? The core functionality is that it creates 00:01:59.450 --> 00:02:03.349 a formal paced learning plan. It moves really 00:02:03.349 --> 00:02:06.170 systematically from the most basic ideas, the 00:02:06.170 --> 00:02:08.949 foundations, up toward more advanced knowledge. 00:02:09.389 --> 00:02:11.509 So it makes sure you build those conceptual pillars 00:02:11.509 --> 00:02:14.710 first. Okay. And crucially, it delivers this 00:02:14.710 --> 00:02:19.120 info in... short, easy to digest chunks. You 00:02:19.120 --> 00:02:20.979 know, it doesn't let you get lost in some 10 00:02:20.979 --> 00:02:23.520 paragraph explanation. That pacing feels essential 00:02:23.520 --> 00:02:26.300 for actually understanding, not just, you know, 00:02:26.520 --> 00:02:29.210 skimming. And what makes it uniquely effective, 00:02:29.210 --> 00:02:31.090 I think, for long -term memory, which is the 00:02:31.090 --> 00:02:32.770 whole point of learning, right? Right. Is that 00:02:32.770 --> 00:02:35.930 it integrates active recall. It regularly stops 00:02:35.930 --> 00:02:38.689 and asks you questions, checks your understanding 00:02:38.689 --> 00:02:41.490 right then and there. Oh. And that process helps 00:02:41.490 --> 00:02:43.550 move the knowledge from short -term into your 00:02:43.550 --> 00:02:46.330 long -term storage. I love the practical application 00:02:46.330 --> 00:02:48.789 of cognitive science there. So let's use the 00:02:48.789 --> 00:02:50.810 example you mentioned, learning about the circular 00:02:50.810 --> 00:02:53.409 economy as a beginner. What's the interaction 00:02:53.409 --> 00:02:56.139 actually like? OK, so the prompt is dead simple, 00:02:56.240 --> 00:02:58.419 something like, turn on study and learn mode 00:02:58.419 --> 00:03:00.819 and teach me about the circular economy. Easy 00:03:00.819 --> 00:03:03.659 enough. And the AI responds with a clear plan, 00:03:03.719 --> 00:03:06.500 like a syllabus. Yeah. It might say, OK, we'll 00:03:06.500 --> 00:03:08.939 cover the basic intro, then compare it to the 00:03:08.939 --> 00:03:12.000 traditional linear economy, move on to main ideas 00:03:12.000 --> 00:03:14.460 like stopping waste, and finish with some real 00:03:14.460 --> 00:03:16.659 examples. This sets expectations right away. 00:03:16.659 --> 00:03:18.439 And I like that check -in step you mentioned. 00:03:18.680 --> 00:03:21.419 After it explains the basics, stop waste, reuse 00:03:21.419 --> 00:03:24.740 materials, help nature regenerate, it then asks 00:03:24.740 --> 00:03:27.280 a specific question. Multiple choice or short 00:03:27.280 --> 00:03:30.340 answer? Exactly. It might ask, what is the primary 00:03:30.340 --> 00:03:32.860 goal of the circular economy model? Yeah. And 00:03:32.860 --> 00:03:35.860 here's the good part. If you miss it, if you 00:03:35.860 --> 00:03:38.960 get the nuance wrong, it doesn't just say wrong. 00:03:39.439 --> 00:03:42.360 It gently re -explains that specific point you 00:03:42.360 --> 00:03:44.900 missed. often using a different angle or analogy. 00:03:45.039 --> 00:03:47.740 Ah, personalized feedback. Right. Which makes 00:03:47.740 --> 00:03:49.719 it an incredible tool for actually making things 00:03:49.719 --> 00:03:52.419 stick, for retention. Yeah, I can see this being 00:03:52.419 --> 00:03:55.500 useful all over for students learning photosynthesis, 00:03:56.020 --> 00:03:58.659 professionals needing SEO basics before a meeting, 00:03:59.080 --> 00:04:01.300 or just anyone curious about how something complex 00:04:01.300 --> 00:04:03.680 works. Exactly. It manages the complexity by 00:04:03.680 --> 00:04:05.520 just giving you the very next step you need. 00:04:05.680 --> 00:04:08.180 OK, so a probing question then. Since it breaks 00:04:08.180 --> 00:04:10.800 complexity down into these small chunks, how 00:04:10.800 --> 00:04:13.699 does this mode ensure the learner still gets 00:04:13.699 --> 00:04:16.240 the big picture, you know, the overall view? 00:04:16.439 --> 00:04:18.680 It forces that continuous check on knowledge 00:04:18.680 --> 00:04:22.060 through quizzes and structures content progressively, 00:04:22.180 --> 00:04:24.399 basically from zero to mastery. That makes perfect 00:04:24.399 --> 00:04:26.899 sense. OK. Now let's shift gears from learning 00:04:26.899 --> 00:04:30.660 passively to active execution. Yes. This is where 00:04:30.660 --> 00:04:32.399 agent mode comes in, right? You said it shifts 00:04:32.399 --> 00:04:35.560 the AI from an answerer. that box you ask questions 00:04:35.560 --> 00:04:39.040 to, to a doer. And this is where it gets really 00:04:39.040 --> 00:04:40.879 interesting, you mentioned. Yeah, I think this 00:04:40.879 --> 00:04:43.160 represents maybe the biggest shift in how we 00:04:43.160 --> 00:04:46.060 interact with these systems in a long time. An 00:04:46.060 --> 00:04:49.060 AI agent, basically, is a system that can take 00:04:49.060 --> 00:04:51.740 a really big, complicated, multi -step goal, 00:04:52.699 --> 00:04:55.079 break it down into smaller, solvable tasks, and 00:04:55.079 --> 00:04:57.000 then actually execute those steps on its own. 00:04:57.379 --> 00:04:59.420 Execute using what? What tools does it have? 00:04:59.540 --> 00:05:02.569 It uses external tools. things like a web browser, 00:05:03.470 --> 00:05:05.629 a code interpreter, sometimes even a virtual 00:05:05.629 --> 00:05:08.149 computer environment. It uses these to carry 00:05:08.149 --> 00:05:10.250 out the steps it planned. And here's the kind 00:05:10.250 --> 00:05:14.269 of revolutionary part. If a step fails, maybe 00:05:14.269 --> 00:05:16.949 a website link is broken or something, the agent 00:05:16.949 --> 00:05:20.129 doesn't just stop. It doesn't. No. It recognizes 00:05:20.129 --> 00:05:23.089 the failure, adjusts its plan, and tries a different 00:05:23.089 --> 00:05:25.920 approach to still reach the final goal. It's 00:05:25.920 --> 00:05:28.339 this dynamic self -correcting loop that older 00:05:28.339 --> 00:05:31.060 systems just didn't have. That self -correction, 00:05:31.060 --> 00:05:33.720 that feels like the key insight we needed. Honestly, 00:05:33.759 --> 00:05:35.920 I still wrestle with prompt drift myself sometimes. 00:05:36.480 --> 00:05:38.860 When you try to automate multi -step things, 00:05:39.379 --> 00:05:41.720 prompt drift, it's like three steps into a long 00:05:41.720 --> 00:05:44.779 task. The AI starts forgetting the rules you 00:05:44.779 --> 00:05:47.759 set way back in the first sentence. Agents help 00:05:47.759 --> 00:05:50.439 manage that internal complexity. They keep the 00:05:50.439 --> 00:05:52.850 focus on the end goal. That admission actually 00:05:52.850 --> 00:05:54.769 really highlights the agent's power, doesn't 00:05:54.769 --> 00:05:56.730 it? It tackles stuff that takes hours manually. 00:05:56.790 --> 00:05:59.230 Let's use that real -world example. Finding and 00:05:59.230 --> 00:06:01.370 comparing online English -speaking courses for 00:06:01.370 --> 00:06:03.569 working folks. Okay, yeah. So the user gets it 00:06:03.569 --> 00:06:05.649 a really specific restrictive prompt. They need 00:06:05.649 --> 00:06:08.550 date of teachers, cost under $80 a month, and 00:06:08.550 --> 00:06:10.790 classes have to be after 7 p .m. local time. 00:06:10.910 --> 00:06:13.910 If you did that by hand? Yeah. Wow. You'd be 00:06:13.910 --> 00:06:16.709 clicking through maybe 20 different websites, 00:06:16.810 --> 00:06:18.949 right? Ten tabs open, probably a spreadsheet. 00:06:19.149 --> 00:06:22.269 Easily. But the agent handles it. So step one 00:06:22.269 --> 00:06:24.689 is planning and research. It figures out the 00:06:24.689 --> 00:06:27.329 right keywords, runs the Google searches. Step 00:06:27.329 --> 00:06:29.329 two is filtering and validation. It actually 00:06:29.329 --> 00:06:31.550 clicks into the search results, looks for the 00:06:31.550 --> 00:06:34.189 pricing pages, the schedule pages, and it filters 00:06:34.189 --> 00:06:36.629 out any school that doesn't meet those three 00:06:36.629 --> 00:06:40.269 specific rules. And step three? Synthesis. It 00:06:40.269 --> 00:06:42.949 builds the final comparison table, showing only 00:06:42.949 --> 00:06:45.230 the options that actually work. It's not just 00:06:45.230 --> 00:06:48.069 summarizing a search. It's doing a full multi 00:06:48.069 --> 00:06:50.779 -criteria research project. That saves hours. 00:06:50.959 --> 00:06:53.279 It eliminates hours of just slogging through 00:06:53.279 --> 00:06:56.959 websites. Oh, just imagining that level of complexity 00:06:56.959 --> 00:06:59.420 handled automatically. The searching, the checking, 00:06:59.579 --> 00:07:01.939 the reporting back. It really does save hours 00:07:01.939 --> 00:07:05.220 of manual comparison. That's that's scalability 00:07:05.220 --> 00:07:07.220 right there. We do need to be clear on security, 00:07:07.480 --> 00:07:10.500 though. Complex tasks like this, they can sometimes 00:07:10.500 --> 00:07:13.000 take a bit, maybe 10 or 15 minutes. You can let 00:07:13.000 --> 00:07:15.660 it run in the background. But crucially. The 00:07:15.660 --> 00:07:18.740 agent will pause if it hits something like a 00:07:18.740 --> 00:07:22.959 CAPUTCHA or a website login page. Ah, so it asks 00:07:22.959 --> 00:07:25.660 the user for help then. It doesn't try to guess. 00:07:26.000 --> 00:07:28.980 Absolutely. You have to manually intervene. And 00:07:28.980 --> 00:07:31.399 you should never, ever give the AI your passwords 00:07:31.399 --> 00:07:33.699 or credit card numbers or any sensitive details 00:07:33.699 --> 00:07:36.259 like that. That's a good point. And always, always 00:07:36.259 --> 00:07:38.439 double check the really critical info it brings 00:07:38.439 --> 00:07:42.019 back, like prices or dates. Check for accuracy 00:07:42.019 --> 00:07:44.620 before you act on its report. Good caveats. OK, 00:07:44.740 --> 00:07:46.959 so beyond just the searching part, what would 00:07:46.959 --> 00:07:49.879 you say is the single biggest time saver with 00:07:49.879 --> 00:07:52.300 this agent function? It autonomously searches, 00:07:52.819 --> 00:07:55.240 filters against your rules, and builds the final 00:07:55.240 --> 00:07:57.660 report, dynamically adjusting its plan if it 00:07:57.660 --> 00:08:00.600 hits roadblocks. That efficiency is just staggering. 00:08:00.740 --> 00:08:02.360 OK, let's look at the next step in integration 00:08:02.360 --> 00:08:05.439 then, making ChatGPT part of your existing world, 00:08:05.759 --> 00:08:08.019 right? Connecting it to daily apps like Google 00:08:08.019 --> 00:08:10.879 Calendar and Gmail. Exactly. This moves the AI 00:08:10.879 --> 00:08:12.939 interaction kind of beyond just the chat window 00:08:12.939 --> 00:08:14.639 and more into your actual professional life. 00:08:15.040 --> 00:08:17.079 We're talking about connectors. Connectors, okay. 00:08:17.180 --> 00:08:19.379 Different from the old plugins. Yeah, different 00:08:19.379 --> 00:08:22.240 from those sometimes clunky plugins. Connectors 00:08:22.240 --> 00:08:24.579 feel more like built -in features. They link 00:08:24.579 --> 00:08:27.600 the AI directly to specific high -trust platforms 00:08:27.600 --> 00:08:30.120 like Google's Workspace apps. Let's start with 00:08:30.120 --> 00:08:32.409 Google Calendar. Once you activate it in the 00:08:32.409 --> 00:08:35.350 settings, your calendar basically becomes conversational. 00:08:35.669 --> 00:08:38.210 That's exactly it. You kind of stop using clicks 00:08:38.210 --> 00:08:40.610 and forms as much, and you start using natural 00:08:40.610 --> 00:08:44.580 language. You can just say, or type. Check if 00:08:44.580 --> 00:08:47.559 I'm free next Wednesday afternoon. Okay. If yes, 00:08:48.019 --> 00:08:50.179 create an event called Meeting with Client and 00:08:50.179 --> 00:08:53.399 Fat Company at 2 p .m. and send an invite to 00:08:53.399 --> 00:08:55.580 Sarah. And it handles all those steps at once. 00:08:55.820 --> 00:08:58.039 Instantly. Or even a more complex search query 00:08:58.039 --> 00:09:01.019 like, find a free two -hour slot next week for 00:09:01.019 --> 00:09:03.100 a doctor's appointment, but make sure it doesn't 00:09:03.100 --> 00:09:05.600 clash with any internal team meetings. Wow. And 00:09:05.600 --> 00:09:08.029 the automation power goes further. You can schedule 00:09:08.029 --> 00:09:10.970 recurring tasks. Ask the AI to send you, say, 00:09:11.250 --> 00:09:13.809 a daily 8 a .m. email summarizing your appointments 00:09:13.809 --> 00:09:16.289 for the day. Useful. Or even ask it to try and 00:09:16.289 --> 00:09:18.409 move a recurring team meeting if a key person 00:09:18.409 --> 00:09:20.950 is already booked during that time. The Gmail 00:09:20.950 --> 00:09:23.149 connection sounds equally transformative then. 00:09:23.519 --> 00:09:26.600 allowing the AI to read, understand, and even 00:09:26.600 --> 00:09:29.019 draft emails based on the conversation history. 00:09:29.720 --> 00:09:31.980 That saves huge amounts of context switching 00:09:31.980 --> 00:09:34.679 time. Totally. You could open a really long, 00:09:34.960 --> 00:09:37.039 dense email thread from a client and just tell 00:09:37.039 --> 00:09:40.240 it. Summarize this long work email into three 00:09:40.240 --> 00:09:43.259 action -oriented bullet points. Nice. Or you 00:09:43.259 --> 00:09:46.820 could say, draft a reply to Mr. Jones using a 00:09:46.820 --> 00:09:49.379 professional yet friendly turn, confirming the 00:09:49.379 --> 00:09:51.679 contract dates he mentioned yesterday. It pulls 00:09:51.679 --> 00:09:54.179 the context. But there's a really vital safety 00:09:54.179 --> 00:09:55.919 note here, isn't there? Something you hinted 00:09:55.919 --> 00:09:58.580 at. The system is deliberately stopped from doing 00:09:58.580 --> 00:10:01.440 one key thing with email. That's right. Super 00:10:01.440 --> 00:10:04.279 important. The system can only read, understand, 00:10:04.440 --> 00:10:07.539 and write drafts. It cannot currently send emails 00:10:07.539 --> 00:10:10.200 itself. The user always has to be the one to 00:10:10.200 --> 00:10:12.139 review the draft and physically hit the send 00:10:12.139 --> 00:10:14.059 button. That seems like a sensible guardrail. 00:10:14.480 --> 00:10:17.299 So, probing question on this. Since reading emails 00:10:17.299 --> 00:10:19.620 is obviously highly sensitive, what's the primary 00:10:19.620 --> 00:10:22.379 limit imposed there for user safety? The system 00:10:22.379 --> 00:10:24.840 is deliberately prevented from sending any email 00:10:24.840 --> 00:10:27.860 on its own behalf. Always requires user action. 00:10:28.059 --> 00:10:31.879 OK, got it. So what does this all mean when put 00:10:31.879 --> 00:10:35.019 together the real power, the kind of revolutionary 00:10:35.019 --> 00:10:37.399 capability comes when you stack these features, 00:10:37.539 --> 00:10:39.220 right? Like you said, like Lego blocks of data. 00:10:39.639 --> 00:10:42.039 Exactly. It's not really three separate features. 00:10:42.100 --> 00:10:44.259 It's more like one unified system potential. 00:10:44.879 --> 00:10:46.759 We can walk through a full workflow example, 00:10:47.139 --> 00:10:48.700 something that might have taken, I don't know, 00:10:48.759 --> 00:10:51.720 a couple of hours of manual work, now potentially 00:10:51.720 --> 00:10:55.019 executed in just minutes, let's say. Organizing 00:10:55.019 --> 00:10:58.580 a two -hour online workshop. on time management 00:10:58.580 --> 00:11:00.659 skills. Okay, let's see the stack in action. 00:11:00.740 --> 00:11:03.100 How would that work? Alright, step one. You use 00:11:03.100 --> 00:11:05.740 agent mode? Yeah. You tell the agent. Research 00:11:05.740 --> 00:11:08.279 and create a detailed workshop outline on time 00:11:08.279 --> 00:11:10.940 management and include five interesting statistics 00:11:10.940 --> 00:11:13.279 about workplace distraction. Okay. The agent 00:11:13.279 --> 00:11:16.340 goes off, does the research, identifies key concepts 00:11:16.340 --> 00:11:19.480 like the Eisenhower matrix, Parkinson's law, 00:11:19.860 --> 00:11:22.659 Pomodoro technique, maybe time blocking, and 00:11:22.659 --> 00:11:25.179 builds the outline. Step two. Now we bring in 00:11:25.179 --> 00:11:28.279 the apps. Step two. Use the Gmail connection. 00:11:29.340 --> 00:11:31.320 Based on that detailed outline the agent just 00:11:31.320 --> 00:11:34.659 made, you ask the AI. Draft three customized 00:11:34.659 --> 00:11:37.080 invitation emails to potential expert speakers 00:11:37.080 --> 00:11:39.799 using this outline. Clever, pulling the context 00:11:39.799 --> 00:11:42.840 across features. Step three. Let's say a speaker 00:11:42.840 --> 00:11:45.679 confirms. Now you use the calendar connector. 00:11:46.000 --> 00:11:48.639 You say, schedule the final event, workshop, 00:11:48.980 --> 00:11:51.299 time management skills for next Tuesday at 10 00:11:51.299 --> 00:11:54.139 a .m. and invite the speaker, Dr. Emily Carter, 00:11:54.460 --> 00:11:56.740 and my internal team. And the AI handles the 00:11:56.740 --> 00:11:59.480 invites, time zones, everything. Yep, finds the 00:11:59.480 --> 00:12:01.460 email, sets up the calendar event, done. OK, 00:12:01.500 --> 00:12:04.659 what's step four? Step four, personal prep. You 00:12:04.659 --> 00:12:06.460 realize you're a bit rusty on one of the topics 00:12:06.460 --> 00:12:08.820 on the outline, so you use study mode. Hey, chat 00:12:08.820 --> 00:12:10.840 GPT, turn on study mode and teach me about the 00:12:10.840 --> 00:12:12.720 Eisenhower matrix. Right, full circle. Right, 00:12:12.740 --> 00:12:14.279 and then it quizzes you on the four quadrants. 00:12:14.860 --> 00:12:17.659 Do it, plan it, delegate, minimize, eliminate. 00:12:18.000 --> 00:12:21.539 Wow. OK, so with just a few natural language 00:12:21.539 --> 00:12:25.059 commands. No coding, no complex setup. You've 00:12:25.059 --> 00:12:28.159 gone from just an idea to detailed research, 00:12:28.500 --> 00:12:30.740 targeted communication, scheduling, and even 00:12:30.740 --> 00:12:32.679 personal learning on the topic. That's true. 00:12:32.960 --> 00:12:36.519 End -to -end automation of a small but non -trivial 00:12:36.519 --> 00:12:40.090 project. So, recapping here. We've really seen 00:12:40.090 --> 00:12:43.350 that AI is rapidly growing into, well, a genuinely 00:12:43.350 --> 00:12:46.490 helpful assistant, and a patient teacher, and 00:12:46.490 --> 00:12:49.389 a powerful project manager, too. It's moved way 00:12:49.389 --> 00:12:51.710 beyond that simple Q &A box we kind of started 00:12:51.710 --> 00:12:54.429 with a couple years ago. Absolutely. And the 00:12:54.429 --> 00:12:57.090 three features we talked about, study mode, agent 00:12:57.090 --> 00:12:59.230 mode, and those app connectors, these aren't 00:12:59.230 --> 00:13:02.049 just theoretical ideas down the road. They are 00:13:02.049 --> 00:13:05.110 immediately practical tools available right now 00:13:05.110 --> 00:13:07.980 for daily work and learning. again, provided 00:13:07.980 --> 00:13:10.379 you're on that plus or maybe an enterprise tier. 00:13:10.779 --> 00:13:13.940 And maybe the takeaway for you listening is don't 00:13:13.940 --> 00:13:16.679 feel overwhelmed by how fast this is all changing. 00:13:16.820 --> 00:13:19.120 Good point. Start small. You know, if you have 00:13:19.120 --> 00:13:21.000 the subscription, maybe try the calendar connector 00:13:21.000 --> 00:13:23.820 first. Ask it to summarize your schedule for 00:13:23.820 --> 00:13:26.980 tomorrow or ask study mode to teach you the basics 00:13:26.980 --> 00:13:28.919 of something you've always been curious about. 00:13:29.080 --> 00:13:31.360 Just experiment a little. That's how you really 00:13:31.360 --> 00:13:33.600 start to leverage this power, I think. Definitely. 00:13:34.039 --> 00:13:35.659 And maybe a final thought to leave you with. 00:13:36.120 --> 00:13:38.799 Now that AI can actively manage and even act 00:13:38.799 --> 00:13:40.960 upon your schedule, your communications, and 00:13:40.960 --> 00:13:44.220 your whole knowledge base, the next really fundamental 00:13:44.220 --> 00:13:47.519 question to consider is, how do we actually redefine 00:13:47.519 --> 00:13:49.779 productivity when the process itself starts getting 00:13:49.779 --> 00:13:52.539 automated like this? What kinds of tasks, what 00:13:52.539 --> 00:13:55.299 creative or conceptual or relational work suddenly 00:13:55.299 --> 00:13:57.940 becomes truly worthy of your limited human time 00:13:57.940 --> 00:13:58.519 and focus?