WEBVTT Kind: captions Language: en 00:00:00.000 --> 00:00:01.470 align:start position:0% You've<00:00:00.320> probably<00:00:00.680> typed<00:00:00.960> something<00:00:01.280> into 00:00:01.470 --> 00:00:01.480 align:start position:0% You've probably typed something into 00:00:01.480 --> 00:00:03.550 align:start position:0% You've probably typed something into Gemini,<00:00:02.360> got<00:00:02.560> an<00:00:02.720> answer,<00:00:03.080> and<00:00:03.200> closed<00:00:03.480> the 00:00:03.550 --> 00:00:03.560 align:start position:0% Gemini, got an answer, and closed the 00:00:03.560 --> 00:00:05.910 align:start position:0% Gemini, got an answer, and closed the tab,<00:00:04.520> the<00:00:04.600> same<00:00:04.880> way<00:00:05.000> you'd<00:00:05.200> use<00:00:05.440> any<00:00:05.680> other 00:00:05.910 --> 00:00:05.920 align:start position:0% tab, the same way you'd use any other 00:00:05.920 --> 00:00:08.190 align:start position:0% tab, the same way you'd use any other chatbot.<00:00:06.640> Here's<00:00:06.920> the<00:00:07.040> thing,<00:00:07.640> that's<00:00:07.880> maybe 00:00:08.190 --> 00:00:08.200 align:start position:0% chatbot. Here's the thing, that's maybe 00:00:08.200 --> 00:00:10.830 align:start position:0% chatbot. Here's the thing, that's maybe 10%<00:00:08.920> of<00:00:09.040> what<00:00:09.200> it<00:00:09.360> actually<00:00:09.760> does.<00:00:10.480> Use<00:00:10.680> it 00:00:10.830 --> 00:00:10.840 align:start position:0% 10% of what it actually does. Use it 00:00:10.840 --> 00:00:12.870 align:start position:0% 10% of what it actually does. Use it right,<00:00:11.480> and<00:00:11.600> it<00:00:11.720> can<00:00:11.840> quietly<00:00:12.280> take<00:00:12.560> over 00:00:12.870 --> 00:00:12.880 align:start position:0% right, and it can quietly take over 00:00:12.880 --> 00:00:14.630 align:start position:0% right, and it can quietly take over almost<00:00:13.280> half<00:00:13.560> the<00:00:13.680> busy<00:00:13.960> work<00:00:14.200> you're<00:00:14.320> still 00:00:14.630 --> 00:00:14.640 align:start position:0% almost half the busy work you're still 00:00:14.640 --> 00:00:17.030 align:start position:0% almost half the busy work you're still doing<00:00:14.920> by<00:00:15.120> hand.<00:00:15.800> I<00:00:15.880> spent<00:00:16.240> hours<00:00:16.640> mapping 00:00:17.030 --> 00:00:17.040 align:start position:0% doing by hand. I spent hours mapping 00:00:17.040 --> 00:00:19.470 align:start position:0% doing by hand. I spent hours mapping every<00:00:17.320> model,<00:00:18.040> every<00:00:18.400> mode,<00:00:19.000> and<00:00:19.200> every 00:00:19.470 --> 00:00:19.480 align:start position:0% every model, every mode, and every 00:00:19.480 --> 00:00:22.110 align:start position:0% every model, every mode, and every product<00:00:19.920> Gemini<00:00:20.400> is<00:00:20.560> quietly<00:00:21.080> wired<00:00:21.400> into, 00:00:22.110 --> 00:00:22.120 align:start position:0% product Gemini is quietly wired into, 00:00:22.120 --> 00:00:24.150 align:start position:0% product Gemini is quietly wired into, and<00:00:22.240> the<00:00:22.320> number<00:00:22.640> that<00:00:22.800> stopped<00:00:23.160> me<00:00:23.360> was<00:00:23.520> this. 00:00:24.150 --> 00:00:24.160 align:start position:0% and the number that stopped me was this. 00:00:24.160 --> 00:00:27.150 align:start position:0% and the number that stopped me was this. Gemini's<00:00:24.720> app<00:00:24.960> alone<00:00:25.360> has<00:00:25.560> over<00:00:25.760> 650<00:00:26.760> million 00:00:27.150 --> 00:00:27.160 align:start position:0% Gemini's app alone has over 650 million 00:00:27.160 --> 00:00:29.310 align:start position:0% Gemini's app alone has over 650 million monthly<00:00:27.520> users,<00:00:28.480> and<00:00:28.600> that's<00:00:28.840> before<00:00:29.200> you 00:00:29.310 --> 00:00:29.320 align:start position:0% monthly users, and that's before you 00:00:29.320 --> 00:00:31.630 align:start position:0% monthly users, and that's before you count<00:00:29.680> everyone<00:00:30.080> using<00:00:30.400> it<00:00:30.680> inside<00:00:31.000> search. 00:00:31.630 --> 00:00:31.640 align:start position:0% count everyone using it inside search. 00:00:31.640 --> 00:00:34.190 align:start position:0% count everyone using it inside search. Most<00:00:31.920> of<00:00:32.000> them<00:00:32.240> are<00:00:32.320> using<00:00:32.640> maybe<00:00:32.920> 20%<00:00:33.680> of<00:00:33.800> it, 00:00:34.190 --> 00:00:34.200 align:start position:0% Most of them are using maybe 20% of it, 00:00:34.200 --> 00:00:36.670 align:start position:0% Most of them are using maybe 20% of it, with<00:00:34.360> no<00:00:34.600> idea<00:00:34.920> the<00:00:35.040> rest<00:00:35.360> even<00:00:35.560> exists.<00:00:36.520> Look 00:00:36.670 --> 00:00:36.680 align:start position:0% with no idea the rest even exists. Look 00:00:36.680 --> 00:00:39.030 align:start position:0% with no idea the rest even exists. Look at<00:00:36.760> this<00:00:36.880> data<00:00:37.120> from<00:00:37.240> the<00:00:37.320> Census<00:00:37.680> Bureau. 00:00:39.030 --> 00:00:39.040 align:start position:0% at this data from the Census Bureau. 00:00:39.040 --> 00:00:41.390 align:start position:0% at this data from the Census Bureau. Only<00:00:39.360> about<00:00:39.720> one<00:00:39.920> in<00:00:40.040> five<00:00:40.440> US<00:00:40.720> businesses 00:00:41.390 --> 00:00:41.400 align:start position:0% Only about one in five US businesses 00:00:41.400 --> 00:00:44.390 align:start position:0% Only about one in five US businesses actually<00:00:41.840> use<00:00:42.120> AI<00:00:42.400> in<00:00:42.520> their<00:00:42.680> operations. 00:00:44.390 --> 00:00:44.400 align:start position:0% actually use AI in their operations. 00:00:44.400 --> 00:00:46.030 align:start position:0% actually use AI in their operations. So,<00:00:44.640> if<00:00:44.760> you<00:00:44.880> run<00:00:45.000> a<00:00:45.080> business,<00:00:45.800> and<00:00:45.920> you're 00:00:46.030 --> 00:00:46.040 align:start position:0% So, if you run a business, and you're 00:00:46.040 --> 00:00:48.070 align:start position:0% So, if you run a business, and you're even<00:00:46.280> thinking<00:00:46.680> about<00:00:46.960> this,<00:00:47.680> you're<00:00:47.880> ahead 00:00:48.070 --> 00:00:48.080 align:start position:0% even thinking about this, you're ahead 00:00:48.080 --> 00:00:50.390 align:start position:0% even thinking about this, you're ahead of<00:00:48.200> most<00:00:48.400> of<00:00:48.480> your<00:00:48.600> competition. 00:00:50.390 --> 00:00:50.400 align:start position:0% of most of your competition. 00:00:50.400 --> 00:00:51.710 align:start position:0% of most of your competition. What<00:00:50.640> you<00:00:50.720> might<00:00:50.920> not<00:00:51.120> know<00:00:51.480> is<00:00:51.600> that 00:00:51.710 --> 00:00:51.720 align:start position:0% What you might not know is that 00:00:51.720 --> 00:00:54.390 align:start position:0% What you might not know is that alongside<00:00:52.240> covering<00:00:52.640> AI<00:00:52.880> news,<00:00:53.880> we<00:00:54.080> work<00:00:54.280> with 00:00:54.390 --> 00:00:54.400 align:start position:0% alongside covering AI news, we work with 00:00:54.400 --> 00:00:55.950 align:start position:0% alongside covering AI news, we work with business<00:00:54.760> owners<00:00:55.040> to<00:00:55.120> help<00:00:55.320> them<00:00:55.480> implement 00:00:55.950 --> 00:00:55.960 align:start position:0% business owners to help them implement 00:00:55.960 --> 00:00:57.950 align:start position:0% business owners to help them implement AI<00:00:56.240> in<00:00:56.360> their<00:00:56.480> business. 00:00:57.950 --> 00:00:57.960 align:start position:0% AI in their business. 00:00:57.960 --> 00:00:59.510 align:start position:0% AI in their business. Our<00:00:58.240> engineering<00:00:58.720> team<00:00:58.960> gets<00:00:59.160> to<00:00:59.240> know<00:00:59.400> how 00:00:59.510 --> 00:00:59.520 align:start position:0% Our engineering team gets to know how 00:00:59.520 --> 00:01:01.430 align:start position:0% Our engineering team gets to know how your<00:00:59.680> business<00:01:00.040> runs,<00:01:00.880> then<00:01:01.080> builds<00:01:01.320> the 00:01:01.430 --> 00:01:01.440 align:start position:0% your business runs, then builds the 00:01:01.440 --> 00:01:03.070 align:start position:0% your business runs, then builds the automation<00:01:01.920> with<00:01:02.160> you. 00:01:03.070 --> 00:01:03.080 align:start position:0% automation with you. 00:01:03.080 --> 00:01:04.430 align:start position:0% automation with you. You'll<00:01:03.320> find<00:01:03.520> the<00:01:03.600> link<00:01:03.840> in<00:01:03.920> the<00:01:04.000> description 00:01:04.430 --> 00:01:04.440 align:start position:0% You'll find the link in the description 00:01:04.440 --> 00:01:05.470 align:start position:0% You'll find the link in the description below. 00:01:05.470 --> 00:01:05.480 align:start position:0% below. 00:01:05.480 --> 00:01:07.390 align:start position:0% below. Click<00:01:05.800> it,<00:01:06.280> fill<00:01:06.480> out<00:01:06.600> a<00:01:06.640> short<00:01:06.960> form<00:01:07.160> about 00:01:07.390 --> 00:01:07.400 align:start position:0% Click it, fill out a short form about 00:01:07.400 --> 00:01:09.230 align:start position:0% Click it, fill out a short form about your<00:01:07.520> business,<00:01:08.400> and<00:01:08.520> we'll<00:01:08.640> get<00:01:08.760> in<00:01:08.880> touch<00:01:09.120> to 00:01:09.230 --> 00:01:09.240 align:start position:0% your business, and we'll get in touch to 00:01:09.240 --> 00:01:11.430 align:start position:0% your business, and we'll get in touch to set<00:01:09.400> up<00:01:09.480> a<00:01:09.560> call.<00:01:10.120> So,<00:01:10.320> in<00:01:10.440> this<00:01:10.640> video,<00:01:11.280> I'm 00:01:11.430 --> 00:01:11.440 align:start position:0% set up a call. So, in this video, I'm 00:01:11.440 --> 00:01:14.270 align:start position:0% set up a call. So, in this video, I'm breaking<00:01:11.800> down<00:01:12.080> exactly<00:01:12.720> what<00:01:12.880> Gemini<00:01:13.400> is<00:01:14.080> as 00:01:14.270 --> 00:01:14.280 align:start position:0% breaking down exactly what Gemini is as 00:01:14.280 --> 00:01:16.230 align:start position:0% breaking down exactly what Gemini is as of<00:01:14.440> mid-2026, 00:01:16.230 --> 00:01:16.240 align:start position:0% of mid-2026, 00:01:16.240 --> 00:01:18.950 align:start position:0% of mid-2026, every<00:01:16.560> current<00:01:16.920> model,<00:01:17.680> every<00:01:18.000> mode,<00:01:18.720> and 00:01:18.950 --> 00:01:18.960 align:start position:0% every current model, every mode, and 00:01:18.960 --> 00:01:20.990 align:start position:0% every current model, every mode, and everywhere<00:01:19.440> Google<00:01:19.840> has<00:01:20.040> quietly<00:01:20.560> built<00:01:20.840> it 00:01:20.990 --> 00:01:21.000 align:start position:0% everywhere Google has quietly built it 00:01:21.000 --> 00:01:23.270 align:start position:0% everywhere Google has quietly built it in.<00:01:21.520> By<00:01:21.680> the<00:01:21.880> end,<00:01:22.320> you'll<00:01:22.480> know<00:01:22.640> exactly 00:01:23.270 --> 00:01:23.280 align:start position:0% in. By the end, you'll know exactly 00:01:23.280 --> 00:01:25.550 align:start position:0% in. By the end, you'll know exactly which<00:01:23.560> Gemini<00:01:24.000> tool<00:01:24.280> to<00:01:24.440> reach<00:01:24.680> for<00:01:25.120> depending 00:01:25.550 --> 00:01:25.560 align:start position:0% which Gemini tool to reach for depending 00:01:25.560 --> 00:01:27.470 align:start position:0% which Gemini tool to reach for depending on<00:01:25.680> what<00:01:25.840> you're<00:01:26.040> actually<00:01:26.400> trying<00:01:26.720> to<00:01:26.840> do, 00:01:27.470 --> 00:01:27.480 align:start position:0% on what you're actually trying to do, 00:01:27.480 --> 00:01:29.190 align:start position:0% on what you're actually trying to do, instead<00:01:27.840> of<00:01:27.920> just<00:01:28.160> typing<00:01:28.520> into<00:01:28.800> whichever 00:01:29.190 --> 00:01:29.200 align:start position:0% instead of just typing into whichever 00:01:29.200 --> 00:01:31.310 align:start position:0% instead of just typing into whichever box<00:01:29.560> is<00:01:29.680> in<00:01:29.800> front<00:01:30.040> of<00:01:30.120> you.<00:01:30.560> First,<00:01:31.120> let's 00:01:31.310 --> 00:01:31.320 align:start position:0% box is in front of you. First, let's 00:01:31.320 --> 00:01:33.230 align:start position:0% box is in front of you. First, let's clear<00:01:31.560> up<00:01:31.680> the<00:01:31.800> biggest<00:01:32.120> misconception. 00:01:33.230 --> 00:01:33.240 align:start position:0% clear up the biggest misconception. 00:01:33.240 --> 00:01:35.750 align:start position:0% clear up the biggest misconception. Gemini<00:01:33.720> isn't<00:01:34.080> one<00:01:34.280> product<00:01:34.800> at<00:01:35.000> all.<00:01:35.480> What 00:01:35.750 --> 00:01:35.760 align:start position:0% Gemini isn't one product at all. What 00:01:35.760 --> 00:01:38.230 align:start position:0% Gemini isn't one product at all. What Gemini<00:01:36.280> actually<00:01:36.840> is,<00:01:37.520> here's<00:01:37.800> the<00:01:37.880> mental 00:01:38.230 --> 00:01:38.240 align:start position:0% Gemini actually is, here's the mental 00:01:38.240 --> 00:01:40.030 align:start position:0% Gemini actually is, here's the mental model<00:01:38.560> you<00:01:38.680> need<00:01:38.920> before<00:01:39.280> any<00:01:39.480> of<00:01:39.560> this<00:01:39.760> makes 00:01:40.030 --> 00:01:40.040 align:start position:0% model you need before any of this makes 00:01:40.040 --> 00:01:42.870 align:start position:0% model you need before any of this makes sense.<00:01:40.720> Gemini<00:01:41.200> isn't<00:01:41.480> a<00:01:41.520> single<00:01:41.960> AI,<00:01:42.720> it's 00:01:42.870 --> 00:01:42.880 align:start position:0% sense. Gemini isn't a single AI, it's 00:01:42.880 --> 00:01:44.470 align:start position:0% sense. Gemini isn't a single AI, it's Google's<00:01:43.280> umbrella<00:01:43.720> name<00:01:44.000> for<00:01:44.160> a<00:01:44.240> whole 00:01:44.470 --> 00:01:44.480 align:start position:0% Google's umbrella name for a whole 00:01:44.480 --> 00:01:47.190 align:start position:0% Google's umbrella name for a whole platform,<00:01:45.400> a<00:01:45.480> family<00:01:45.880> of<00:01:46.000> models<00:01:46.360> underneath, 00:01:47.190 --> 00:01:47.200 align:start position:0% platform, a family of models underneath, 00:01:47.200 --> 00:01:48.870 align:start position:0% platform, a family of models underneath, and<00:01:47.320> a<00:01:47.360> set<00:01:47.560> of<00:01:47.680> products<00:01:48.080> on<00:01:48.240> top<00:01:48.560> that<00:01:48.720> let 00:01:48.870 --> 00:01:48.880 align:start position:0% and a set of products on top that let 00:01:48.880 --> 00:01:50.630 align:start position:0% and a set of products on top that let you<00:01:49.000> actually<00:01:49.400> talk<00:01:49.680> to<00:01:49.800> them.<00:01:50.240> Think<00:01:50.440> of<00:01:50.520> it 00:01:50.630 --> 00:01:50.640 align:start position:0% you actually talk to them. Think of it 00:01:50.640 --> 00:01:52.710 align:start position:0% you actually talk to them. Think of it in<00:01:50.760> two<00:01:50.960> layers.<00:01:51.800> The<00:01:51.880> bottom<00:01:52.160> layer<00:01:52.480> is<00:01:52.640> the 00:01:52.710 --> 00:01:52.720 align:start position:0% in two layers. The bottom layer is the 00:01:52.720 --> 00:01:54.830 align:start position:0% in two layers. The bottom layer is the models<00:01:53.080> themselves,<00:01:53.960> things<00:01:54.200> like<00:01:54.400> Gemini 00:01:54.830 --> 00:01:54.840 align:start position:0% models themselves, things like Gemini 00:01:54.840 --> 00:01:58.510 align:start position:0% models themselves, things like Gemini 3.6<00:01:55.640> Flash<00:01:56.280> or<00:01:56.400> Gemini<00:01:56.840> 3.1<00:01:57.560> Pro.<00:01:58.200> These<00:01:58.440> are 00:01:58.510 --> 00:01:58.520 align:start position:0% 3.6 Flash or Gemini 3.1 Pro. These are 00:01:58.520 --> 00:02:00.910 align:start position:0% 3.6 Flash or Gemini 3.1 Pro. These are the<00:01:58.680> engines<00:01:59.440> tuned<00:01:59.720> for<00:01:59.840> different<00:02:00.160> jobs. 00:02:00.910 --> 00:02:00.920 align:start position:0% the engines tuned for different jobs. 00:02:00.920 --> 00:02:02.990 align:start position:0% the engines tuned for different jobs. Some<00:02:01.160> built<00:02:01.400> for<00:02:01.520> speed,<00:02:02.360> some<00:02:02.600> for<00:02:02.760> heavy 00:02:02.990 --> 00:02:03.000 align:start position:0% Some built for speed, some for heavy 00:02:03.000 --> 00:02:05.750 align:start position:0% Some built for speed, some for heavy reasoning,<00:02:03.800> some<00:02:04.080> for<00:02:04.240> images<00:02:04.720> or<00:02:04.920> audio.<00:02:05.640> You 00:02:05.750 --> 00:02:05.760 align:start position:0% reasoning, some for images or audio. You 00:02:05.760 --> 00:02:07.110 align:start position:0% reasoning, some for images or audio. You never<00:02:05.960> see<00:02:06.120> these<00:02:06.320> names<00:02:06.640> unless<00:02:06.880> you<00:02:06.960> go 00:02:07.110 --> 00:02:07.120 align:start position:0% never see these names unless you go 00:02:07.120 --> 00:02:09.270 align:start position:0% never see these names unless you go looking.<00:02:08.000> The<00:02:08.080> top<00:02:08.360> layer<00:02:08.600> is<00:02:08.759> everything<00:02:09.119> you 00:02:09.270 --> 00:02:09.280 align:start position:0% looking. The top layer is everything you 00:02:09.280 --> 00:02:11.949 align:start position:0% looking. The top layer is everything you actually<00:02:09.640> click<00:02:09.920> on.<00:02:10.440> The<00:02:10.560> Gemini<00:02:11.040> app,<00:02:11.640> AI 00:02:11.949 --> 00:02:11.959 align:start position:0% actually click on. The Gemini app, AI 00:02:11.959 --> 00:02:14.630 align:start position:0% actually click on. The Gemini app, AI mode<00:02:12.320> inside<00:02:12.720> Google<00:02:13.000> search,<00:02:13.680> Gemini<00:02:14.200> inside 00:02:14.630 --> 00:02:14.640 align:start position:0% mode inside Google search, Gemini inside 00:02:14.640 --> 00:02:17.030 align:start position:0% mode inside Google search, Gemini inside Gmail<00:02:15.040> and<00:02:15.200> Docs,<00:02:16.040> the<00:02:16.160> voice<00:02:16.440> assistant<00:02:16.920> on 00:02:17.030 --> 00:02:17.040 align:start position:0% Gmail and Docs, the voice assistant on 00:02:17.040 --> 00:02:18.670 align:start position:0% Gmail and Docs, the voice assistant on your<00:02:17.160> phone.<00:02:17.800> All<00:02:18.000> of<00:02:18.080> those<00:02:18.320> are<00:02:18.440> just 00:02:18.670 --> 00:02:18.680 align:start position:0% your phone. All of those are just 00:02:18.680 --> 00:02:20.670 align:start position:0% your phone. All of those are just different<00:02:19.040> doors<00:02:19.440> into<00:02:19.640> the<00:02:19.720> same<00:02:20.120> underlying 00:02:20.670 --> 00:02:20.680 align:start position:0% different doors into the same underlying 00:02:20.680 --> 00:02:22.390 align:start position:0% different doors into the same underlying models.<00:02:21.440> That's<00:02:21.640> the<00:02:21.720> whole<00:02:21.920> point<00:02:22.160> of<00:02:22.240> this 00:02:22.390 --> 00:02:22.400 align:start position:0% models. That's the whole point of this 00:02:22.400 --> 00:02:24.670 align:start position:0% models. That's the whole point of this video.<00:02:23.080> Google<00:02:23.440> isn't<00:02:23.680> trying<00:02:23.959> to<00:02:24.080> build<00:02:24.440> one 00:02:24.670 --> 00:02:24.680 align:start position:0% video. Google isn't trying to build one 00:02:24.680 --> 00:02:26.550 align:start position:0% video. Google isn't trying to build one great<00:02:25.000> chatbot.<00:02:25.760> It's<00:02:25.920> trying<00:02:26.200> to<00:02:26.280> put<00:02:26.440> the 00:02:26.550 --> 00:02:26.560 align:start position:0% great chatbot. It's trying to put the 00:02:26.560 --> 00:02:28.790 align:start position:0% great chatbot. It's trying to put the same<00:02:26.920> AI<00:02:27.200> brain<00:02:27.560> behind<00:02:27.959> every<00:02:28.280> product<00:02:28.680> you 00:02:28.790 --> 00:02:28.800 align:start position:0% same AI brain behind every product you 00:02:28.800 --> 00:02:30.670 align:start position:0% same AI brain behind every product you already<00:02:29.160> use.<00:02:29.840> So,<00:02:29.959> let's<00:02:30.200> start<00:02:30.440> with<00:02:30.600> the 00:02:30.670 --> 00:02:30.680 align:start position:0% already use. So, let's start with the 00:02:30.680 --> 00:02:33.670 align:start position:0% already use. So, let's start with the brains,<00:02:31.640> the<00:02:31.800> actual<00:02:32.200> models,<00:02:33.120> because<00:02:33.400> once 00:02:33.670 --> 00:02:33.680 align:start position:0% brains, the actual models, because once 00:02:33.680 --> 00:02:35.750 align:start position:0% brains, the actual models, because once you<00:02:33.800> know<00:02:34.080> what<00:02:34.320> each<00:02:34.520> one<00:02:34.680> is<00:02:34.840> built<00:02:35.120> for, 00:02:35.750 --> 00:02:35.760 align:start position:0% you know what each one is built for, 00:02:35.760 --> 00:02:37.790 align:start position:0% you know what each one is built for, everything<00:02:36.200> else<00:02:36.440> clicks<00:02:36.760> into<00:02:36.959> place.<00:02:37.680> The 00:02:37.790 --> 00:02:37.800 align:start position:0% everything else clicks into place. The 00:02:37.800 --> 00:02:39.949 align:start position:0% everything else clicks into place. The current<00:02:38.160> model<00:02:38.480> lineup.<00:02:39.240> This<00:02:39.440> is<00:02:39.560> a<00:02:39.600> demo 00:02:39.949 --> 00:02:39.959 align:start position:0% current model lineup. This is a demo 00:02:39.959 --> 00:02:41.990 align:start position:0% current model lineup. This is a demo checklist,<00:02:40.880> so<00:02:41.040> we're<00:02:41.200> going<00:02:41.520> model<00:02:41.840> by 00:02:41.990 --> 00:02:42.000 align:start position:0% checklist, so we're going model by 00:02:42.000 --> 00:02:44.310 align:start position:0% checklist, so we're going model by model.<00:02:42.640> What<00:02:42.840> it<00:02:42.959> is,<00:02:43.520> what<00:02:43.680> it's<00:02:43.880> actually 00:02:44.310 --> 00:02:44.320 align:start position:0% model. What it is, what it's actually 00:02:44.320 --> 00:02:46.390 align:start position:0% model. What it is, what it's actually good<00:02:44.560> for,<00:02:45.160> and<00:02:45.360> where<00:02:45.560> you<00:02:45.640> can<00:02:45.800> get<00:02:46.000> it. 00:02:46.390 --> 00:02:46.400 align:start position:0% good for, and where you can get it. 00:02:46.400 --> 00:02:48.830 align:start position:0% good for, and where you can get it. Gemini<00:02:46.800> 3.7<00:02:47.640> flash. 00:02:48.830 --> 00:02:48.840 align:start position:0% Gemini 3.7 flash. 00:02:48.840 --> 00:02:51.590 align:start position:0% Gemini 3.7 flash. Launched<00:02:49.160> on<00:02:49.280> August<00:02:49.560> 13th,<00:02:50.080> 2026,<00:02:51.320> this<00:02:51.480> is 00:02:51.590 --> 00:02:51.600 align:start position:0% Launched on August 13th, 2026, this is 00:02:51.600 --> 00:02:53.430 align:start position:0% Launched on August 13th, 2026, this is Google's<00:02:51.959> newest<00:02:52.280> flash<00:02:52.600> model<00:02:52.959> and<00:02:53.080> its<00:02:53.200> most 00:02:53.430 --> 00:02:53.440 align:start position:0% Google's newest flash model and its most 00:02:53.440 --> 00:02:55.470 align:start position:0% Google's newest flash model and its most capable<00:02:53.880> workhorse<00:02:54.360> yet.<00:02:55.120> It's<00:02:55.280> built 00:02:55.470 --> 00:02:55.480 align:start position:0% capable workhorse yet. It's built 00:02:55.480 --> 00:02:58.030 align:start position:0% capable workhorse yet. It's built primarily<00:02:55.959> for<00:02:56.080> coding<00:02:56.520> and<00:02:56.720> AI<00:02:57.080> agents<00:02:57.880> with 00:02:58.030 --> 00:02:58.040 align:start position:0% primarily for coding and AI agents with 00:02:58.040 --> 00:02:59.390 align:start position:0% primarily for coding and AI agents with major<00:02:58.360> improvements<00:02:58.880> in<00:02:59.000> software 00:02:59.390 --> 00:02:59.400 align:start position:0% major improvements in software 00:02:59.400 --> 00:03:01.509 align:start position:0% major improvements in software engineering,<00:03:00.280> web<00:03:00.480> development,<00:03:01.360> and 00:03:01.509 --> 00:03:01.519 align:start position:0% engineering, web development, and 00:03:01.519 --> 00:03:04.310 align:start position:0% engineering, web development, and complex<00:03:02.080> multi-step<00:03:02.680> workflows.<00:03:03.880> Google<00:03:04.160> has 00:03:04.310 --> 00:03:04.320 align:start position:0% complex multi-step workflows. Google has 00:03:04.320 --> 00:03:05.830 align:start position:0% complex multi-step workflows. Google has already<00:03:04.600> made<00:03:04.800> it<00:03:04.920> generally<00:03:05.320> available 00:03:05.830 --> 00:03:05.840 align:start position:0% already made it generally available 00:03:05.840 --> 00:03:09.110 align:start position:0% already made it generally available through<00:03:06.000> the<00:03:06.120> Gemini<00:03:06.600> API,<00:03:07.680> positioning<00:03:08.239> 3.7 00:03:09.110 --> 00:03:09.120 align:start position:0% through the Gemini API, positioning 3.7 00:03:09.120 --> 00:03:11.550 align:start position:0% through the Gemini API, positioning 3.7 flash<00:03:09.640> as<00:03:09.800> the<00:03:09.920> new<00:03:10.160> go-to<00:03:10.600> model<00:03:11.320> when<00:03:11.480> you 00:03:11.550 --> 00:03:11.560 align:start position:0% flash as the new go-to model when you 00:03:11.560 --> 00:03:13.270 align:start position:0% flash as the new go-to model when you want<00:03:11.800> strong<00:03:12.120> intelligence<00:03:12.720> without<00:03:13.000> giving 00:03:13.270 --> 00:03:13.280 align:start position:0% want strong intelligence without giving 00:03:13.280 --> 00:03:15.070 align:start position:0% want strong intelligence without giving up<00:03:13.400> the<00:03:13.480> speed<00:03:13.920> and<00:03:14.040> efficiency<00:03:14.640> the<00:03:14.760> flash 00:03:15.070 --> 00:03:15.080 align:start position:0% up the speed and efficiency the flash 00:03:15.080 --> 00:03:18.310 align:start position:0% up the speed and efficiency the flash lineup<00:03:15.400> is<00:03:15.519> known<00:03:15.760> for.<00:03:16.239> Gemini<00:03:16.720> 3.6<00:03:17.519> flash. 00:03:18.310 --> 00:03:18.320 align:start position:0% lineup is known for. Gemini 3.6 flash. 00:03:18.320 --> 00:03:20.270 align:start position:0% lineup is known for. Gemini 3.6 flash. This<00:03:18.519> is<00:03:18.680> Google's<00:03:19.040> current<00:03:19.400> flagship, 00:03:20.270 --> 00:03:20.280 align:start position:0% This is Google's current flagship, 00:03:20.280 --> 00:03:22.310 align:start position:0% This is Google's current flagship, announced<00:03:20.720> in<00:03:20.800> a<00:03:20.840> company<00:03:21.239> blog<00:03:21.560> post<00:03:21.880> on<00:03:22.000> July 00:03:22.310 --> 00:03:22.320 align:start position:0% announced in a company blog post on July 00:03:22.320 --> 00:03:25.710 align:start position:0% announced in a company blog post on July 21st,<00:03:23.120> 2026.<00:03:24.239> It's<00:03:24.440> built<00:03:24.680> as<00:03:24.800> a<00:03:24.880> workhorse, 00:03:25.710 --> 00:03:25.720 align:start position:0% 21st, 2026. It's built as a workhorse, 00:03:25.720 --> 00:03:27.910 align:start position:0% 21st, 2026. It's built as a workhorse, strong<00:03:26.120> at<00:03:26.239> coding,<00:03:26.880> knowledge<00:03:27.320> work,<00:03:27.800> and 00:03:27.910 --> 00:03:27.920 align:start position:0% strong at coding, knowledge work, and 00:03:27.920 --> 00:03:29.910 align:start position:0% strong at coding, knowledge work, and multimodal<00:03:28.519> tasks.<00:03:29.360> And<00:03:29.480> according<00:03:29.800> to 00:03:29.910 --> 00:03:29.920 align:start position:0% multimodal tasks. And according to 00:03:29.920 --> 00:03:32.070 align:start position:0% multimodal tasks. And according to Google's<00:03:30.320> own<00:03:30.480> numbers,<00:03:31.280> it<00:03:31.440> does<00:03:31.640> the<00:03:31.760> job 00:03:32.070 --> 00:03:32.080 align:start position:0% Google's own numbers, it does the job 00:03:32.080 --> 00:03:35.110 align:start position:0% Google's own numbers, it does the job using<00:03:32.320> about<00:03:32.640> 17%<00:03:33.720> fewer<00:03:33.959> tokens<00:03:34.440> on<00:03:34.680> average 00:03:35.110 --> 00:03:35.120 align:start position:0% using about 17% fewer tokens on average 00:03:35.120 --> 00:03:37.310 align:start position:0% using about 17% fewer tokens on average than<00:03:35.239> its<00:03:35.400> predecessor.<00:03:36.239> Fewer<00:03:36.560> tokens<00:03:37.040> means 00:03:37.310 --> 00:03:37.320 align:start position:0% than its predecessor. Fewer tokens means 00:03:37.320 --> 00:03:39.509 align:start position:0% than its predecessor. Fewer tokens means faster<00:03:37.720> answers<00:03:38.519> and<00:03:38.640> a<00:03:38.680> lower<00:03:39.000> bill<00:03:39.360> if 00:03:39.509 --> 00:03:39.519 align:start position:0% faster answers and a lower bill if 00:03:39.519 --> 00:03:41.590 align:start position:0% faster answers and a lower bill if you're<00:03:39.640> paying<00:03:40.000> for<00:03:40.200> it<00:03:40.320> through<00:03:40.519> the<00:03:40.680> API. 00:03:41.590 --> 00:03:41.600 align:start position:0% you're paying for it through the API. 00:03:41.600 --> 00:03:44.070 align:start position:0% you're paying for it through the API. You<00:03:41.720> can<00:03:41.880> reach<00:03:42.120> it<00:03:42.280> through<00:03:42.480> the<00:03:42.600> Gemini<00:03:43.080> API, 00:03:44.070 --> 00:03:44.080 align:start position:0% You can reach it through the Gemini API, 00:03:44.080 --> 00:03:46.630 align:start position:0% You can reach it through the Gemini API, through<00:03:44.320> AI<00:03:44.600> Studio,<00:03:45.560> or<00:03:45.720> simply<00:03:46.080> by<00:03:46.280> using 00:03:46.630 --> 00:03:46.640 align:start position:0% through AI Studio, or simply by using 00:03:46.640 --> 00:03:49.630 align:start position:0% through AI Studio, or simply by using the<00:03:46.760> Gemini<00:03:47.239> app<00:03:47.680> and<00:03:47.840> searches<00:03:48.360> AI<00:03:48.680> mode.<00:03:49.360> No 00:03:49.630 --> 00:03:49.640 align:start position:0% the Gemini app and searches AI mode. No 00:03:49.640 --> 00:03:51.509 align:start position:0% the Gemini app and searches AI mode. No extra<00:03:49.959> setup<00:03:50.320> required.<00:03:51.080> If<00:03:51.239> you<00:03:51.320> only 00:03:51.509 --> 00:03:51.519 align:start position:0% extra setup required. If you only 00:03:51.519 --> 00:03:53.750 align:start position:0% extra setup required. If you only remember<00:03:51.920> one<00:03:52.160> model<00:03:52.560> name<00:03:52.800> from<00:03:52.960> this<00:03:53.080> video, 00:03:53.750 --> 00:03:53.760 align:start position:0% remember one model name from this video, 00:03:53.760 --> 00:03:55.710 align:start position:0% remember one model name from this video, make<00:03:54.000> it<00:03:54.120> this<00:03:54.360> one<00:03:54.920> because<00:03:55.200> it's<00:03:55.360> what<00:03:55.480> most 00:03:55.710 --> 00:03:55.720 align:start position:0% make it this one because it's what most 00:03:55.720 --> 00:03:57.550 align:start position:0% make it this one because it's what most of<00:03:55.840> Gemini<00:03:56.320> is<00:03:56.480> quietly<00:03:56.920> running<00:03:57.200> on<00:03:57.360> right 00:03:57.550 --> 00:03:57.560 align:start position:0% of Gemini is quietly running on right 00:03:57.560 --> 00:04:00.950 align:start position:0% of Gemini is quietly running on right now.<00:03:58.080> Gemini<00:03:58.560> 3.5<00:03:59.400> Flash.<00:04:00.160> This<00:04:00.400> one<00:04:00.640> launched 00:04:00.950 --> 00:04:00.960 align:start position:0% now. Gemini 3.5 Flash. This one launched 00:04:00.960 --> 00:04:03.230 align:start position:0% now. Gemini 3.5 Flash. This one launched back<00:04:01.280> in<00:04:01.440> May<00:04:01.720> 2026 00:04:03.230 --> 00:04:03.240 align:start position:0% back in May 2026 00:04:03.240 --> 00:04:04.790 align:start position:0% back in May 2026 and<00:04:03.360> it's<00:04:03.520> the<00:04:03.640> model<00:04:04.040> that<00:04:04.200> was<00:04:04.400> actually 00:04:04.790 --> 00:04:04.800 align:start position:0% and it's the model that was actually 00:04:04.800 --> 00:04:08.030 align:start position:0% and it's the model that was actually powering<00:04:05.360> AI<00:04:05.640> mode<00:04:06.000> in<00:04:06.160> search<00:04:06.880> before<00:04:07.240> 3.6 00:04:08.030 --> 00:04:08.040 align:start position:0% powering AI mode in search before 3.6 00:04:08.040 --> 00:04:10.310 align:start position:0% powering AI mode in search before 3.6 Flash<00:04:08.360> took<00:04:08.560> over.<00:04:09.160> Google<00:04:09.480> described<00:04:10.000> it<00:04:10.120> as 00:04:10.310 --> 00:04:10.320 align:start position:0% Flash took over. Google described it as 00:04:10.320 --> 00:04:11.949 align:start position:0% Flash took over. Google described it as frontier-level<00:04:11.120> intelligence<00:04:11.840> at 00:04:11.949 --> 00:04:11.959 align:start position:0% frontier-level intelligence at 00:04:11.959 --> 00:04:13.870 align:start position:0% frontier-level intelligence at exceptional<00:04:12.520> speed<00:04:13.240> and<00:04:13.360> by<00:04:13.520> its<00:04:13.680> own 00:04:13.870 --> 00:04:13.880 align:start position:0% exceptional speed and by its own 00:04:13.880 --> 00:04:16.390 align:start position:0% exceptional speed and by its own benchmarks,<00:04:14.680> it<00:04:14.840> pushed<00:04:15.160> output<00:04:15.600> throughput 00:04:16.390 --> 00:04:16.400 align:start position:0% benchmarks, it pushed output throughput 00:04:16.400 --> 00:04:18.349 align:start position:0% benchmarks, it pushed output throughput to<00:04:16.560> roughly<00:04:16.920> four<00:04:17.160> times<00:04:17.519> faster<00:04:17.920> than<00:04:18.120> other 00:04:18.349 --> 00:04:18.359 align:start position:0% to roughly four times faster than other 00:04:18.359 --> 00:04:20.470 align:start position:0% to roughly four times faster than other top<00:04:18.600> models<00:04:18.959> at<00:04:19.040> the<00:04:19.160> time.<00:04:19.799> It's<00:04:20.000> still<00:04:20.239> very 00:04:20.470 --> 00:04:20.480 align:start position:0% top models at the time. It's still very 00:04:20.480 --> 00:04:22.430 align:start position:0% top models at the time. It's still very much<00:04:20.760> in<00:04:20.880> active<00:04:21.239> use<00:04:21.480> across<00:04:21.840> the<00:04:21.959> Gemini 00:04:22.430 --> 00:04:22.440 align:start position:0% much in active use across the Gemini 00:04:22.440 --> 00:04:25.350 align:start position:0% much in active use across the Gemini app,<00:04:23.120> Google's<00:04:23.560> anti-gravity<00:04:24.280> platform,<00:04:25.200> and 00:04:25.350 --> 00:04:25.360 align:start position:0% app, Google's anti-gravity platform, and 00:04:25.360 --> 00:04:27.550 align:start position:0% app, Google's anti-gravity platform, and enterprise<00:04:25.880> tools.<00:04:26.680> It's<00:04:26.880> slightly<00:04:27.200> behind 00:04:27.550 --> 00:04:27.560 align:start position:0% enterprise tools. It's slightly behind 00:04:27.560 --> 00:04:30.110 align:start position:0% enterprise tools. It's slightly behind 3.6<00:04:28.320> now,<00:04:28.880> but<00:04:29.000> it's<00:04:29.160> the<00:04:29.240> model<00:04:29.640> quietly 00:04:30.110 --> 00:04:30.120 align:start position:0% 3.6 now, but it's the model quietly 00:04:30.120 --> 00:04:31.909 align:start position:0% 3.6 now, but it's the model quietly sitting<00:04:30.400> behind<00:04:30.800> a<00:04:30.840> huge<00:04:31.280> share<00:04:31.600> of<00:04:31.720> what 00:04:31.909 --> 00:04:31.919 align:start position:0% sitting behind a huge share of what 00:04:31.919 --> 00:04:34.230 align:start position:0% sitting behind a huge share of what shipped<00:04:32.280> this<00:04:32.480> year.<00:04:33.040> Gemini<00:04:33.440> 3.5 00:04:34.230 --> 00:04:34.240 align:start position:0% shipped this year. Gemini 3.5 00:04:34.240 --> 00:04:36.590 align:start position:0% shipped this year. Gemini 3.5 Flashlight.<00:04:35.120> Same<00:04:35.440> July<00:04:35.720> announcement, 00:04:36.590 --> 00:04:36.600 align:start position:0% Flashlight. Same July announcement, 00:04:36.600 --> 00:04:38.550 align:start position:0% Flashlight. Same July announcement, different<00:04:36.960> job<00:04:37.200> entirely.<00:04:38.120> This<00:04:38.320> is<00:04:38.440> the 00:04:38.550 --> 00:04:38.560 align:start position:0% different job entirely. This is the 00:04:38.560 --> 00:04:40.870 align:start position:0% different job entirely. This is the stripped-down,<00:04:39.480> high-throughput<00:04:40.200> sibling. 00:04:40.870 --> 00:04:40.880 align:start position:0% stripped-down, high-throughput sibling. 00:04:40.880 --> 00:04:43.550 align:start position:0% stripped-down, high-throughput sibling. Google<00:04:41.240> sites<00:04:41.640> roughly<00:04:42.040> 350<00:04:42.960> tokens<00:04:43.440> per 00:04:43.550 --> 00:04:43.560 align:start position:0% Google sites roughly 350 tokens per 00:04:43.560 --> 00:04:45.909 align:start position:0% Google sites roughly 350 tokens per second,<00:04:44.360> which<00:04:44.520> is<00:04:44.640> built<00:04:44.880> for<00:04:45.000> volume,<00:04:45.640> not 00:04:45.909 --> 00:04:45.919 align:start position:0% second, which is built for volume, not 00:04:45.919 --> 00:04:47.750 align:start position:0% second, which is built for volume, not depth.<00:04:46.520> You<00:04:46.680> wouldn't<00:04:46.919> use<00:04:47.080> this<00:04:47.280> for<00:04:47.440> a<00:04:47.480> hard 00:04:47.750 --> 00:04:47.760 align:start position:0% depth. You wouldn't use this for a hard 00:04:47.760 --> 00:04:49.390 align:start position:0% depth. You wouldn't use this for a hard reasoning<00:04:48.120> problem.<00:04:48.760> You'd<00:04:48.960> use<00:04:49.120> it<00:04:49.240> for 00:04:49.390 --> 00:04:49.400 align:start position:0% reasoning problem. You'd use it for 00:04:49.400 --> 00:04:51.909 align:start position:0% reasoning problem. You'd use it for background<00:04:50.000> tasks<00:04:50.560> and<00:04:50.760> agent<00:04:51.080> pipelines 00:04:51.909 --> 00:04:51.919 align:start position:0% background tasks and agent pipelines 00:04:51.919 --> 00:04:53.510 align:start position:0% background tasks and agent pipelines that<00:04:52.080> need<00:04:52.240> to<00:04:52.320> move<00:04:52.560> fast<00:04:52.919> and<00:04:53.080> cheap<00:04:53.400> at 00:04:53.510 --> 00:04:53.520 align:start position:0% that need to move fast and cheap at 00:04:53.520 --> 00:04:56.510 align:start position:0% that need to move fast and cheap at scale.<00:04:54.240> Gemini<00:04:54.600> 3.1<00:04:55.400> Pro.<00:04:56.040> Released<00:04:56.400> in 00:04:56.510 --> 00:04:56.520 align:start position:0% scale. Gemini 3.1 Pro. Released in 00:04:56.520 --> 00:04:59.150 align:start position:0% scale. Gemini 3.1 Pro. Released in February<00:04:57.000> 2026,<00:04:58.360> this<00:04:58.520> is<00:04:58.640> the<00:04:58.760> reasoning 00:04:59.150 --> 00:04:59.160 align:start position:0% February 2026, this is the reasoning 00:04:59.160 --> 00:05:01.350 align:start position:0% February 2026, this is the reasoning specialist.<00:05:00.120> Google's<00:05:00.480> benchmarks<00:05:01.040> claim 00:05:01.350 --> 00:05:01.360 align:start position:0% specialist. Google's benchmarks claim 00:05:01.360 --> 00:05:03.190 align:start position:0% specialist. Google's benchmarks claim roughly<00:05:01.760> double<00:05:02.080> the<00:05:02.200> logic<00:05:02.560> performance<00:05:03.080> of 00:05:03.190 --> 00:05:03.200 align:start position:0% roughly double the logic performance of 00:05:03.200 --> 00:05:05.510 align:start position:0% roughly double the logic performance of the<00:05:03.320> earlier<00:05:03.680> Gemini<00:05:04.080> 3<00:05:04.360> Pro.<00:05:04.960> It's<00:05:05.160> mostly 00:05:05.510 --> 00:05:05.520 align:start position:0% the earlier Gemini 3 Pro. It's mostly 00:05:05.520 --> 00:05:07.470 align:start position:0% the earlier Gemini 3 Pro. It's mostly gated<00:05:05.840> behind<00:05:06.200> preview<00:05:06.640> access<00:05:07.120> through<00:05:07.280> the 00:05:07.470 --> 00:05:07.480 align:start position:0% gated behind preview access through the 00:05:07.480 --> 00:05:11.350 align:start position:0% gated behind preview access through the API,<00:05:08.320> anti-gravity,<00:05:09.400> Vertex<00:05:09.919> AI,<00:05:10.640> and<00:05:10.800> Pro<00:05:11.160> or 00:05:11.350 --> 00:05:11.360 align:start position:0% API, anti-gravity, Vertex AI, and Pro or 00:05:11.360 --> 00:05:13.710 align:start position:0% API, anti-gravity, Vertex AI, and Pro or Ultra<00:05:11.640> subscriptions<00:05:12.400> in<00:05:12.520> the<00:05:12.640> consumer<00:05:13.120> app. 00:05:13.710 --> 00:05:13.720 align:start position:0% Ultra subscriptions in the consumer app. 00:05:13.720 --> 00:05:17.350 align:start position:0% Ultra subscriptions in the consumer app. If<00:05:13.880> 3.6<00:05:14.600> Flash<00:05:14.919> is<00:05:15.040> built<00:05:15.280> for<00:05:15.400> speed,<00:05:16.080> 3.1<00:05:16.800> Pro 00:05:17.350 --> 00:05:17.360 align:start position:0% If 3.6 Flash is built for speed, 3.1 Pro 00:05:17.360 --> 00:05:19.510 align:start position:0% If 3.6 Flash is built for speed, 3.1 Pro is<00:05:17.560> built<00:05:17.800> for<00:05:17.919> depth.<00:05:18.600> The<00:05:18.720> model<00:05:19.040> you'd<00:05:19.200> want 00:05:19.510 --> 00:05:19.520 align:start position:0% is built for depth. The model you'd want 00:05:19.520 --> 00:05:21.750 align:start position:0% is built for depth. The model you'd want on<00:05:19.640> a<00:05:19.720> genuinely<00:05:20.040> hard<00:05:20.560> problem,<00:05:21.240> not<00:05:21.440> a<00:05:21.480> quick 00:05:21.750 --> 00:05:21.760 align:start position:0% on a genuinely hard problem, not a quick 00:05:21.760 --> 00:05:23.470 align:start position:0% on a genuinely hard problem, not a quick one.<00:05:22.280> And<00:05:22.400> if<00:05:22.520> you're<00:05:22.680> actually<00:05:23.040> building<00:05:23.360> on 00:05:23.470 --> 00:05:23.480 align:start position:0% one. And if you're actually building on 00:05:23.480 --> 00:05:25.870 align:start position:0% one. And if you're actually building on top<00:05:23.760> of<00:05:23.880> these<00:05:24.200> through<00:05:24.360> the<00:05:24.560> API,<00:05:25.360> price<00:05:25.680> is 00:05:25.870 --> 00:05:25.880 align:start position:0% top of these through the API, price is 00:05:25.880 --> 00:05:27.950 align:start position:0% top of these through the API, price is where<00:05:26.040> the<00:05:26.160> real-world<00:05:26.680> decision<00:05:27.160> gets<00:05:27.400> made. 00:05:27.950 --> 00:05:27.960 align:start position:0% where the real-world decision gets made. 00:05:27.960 --> 00:05:29.430 align:start position:0% where the real-world decision gets made. According<00:05:28.320> to<00:05:28.440> Google's<00:05:28.800> own<00:05:29.000> published 00:05:29.430 --> 00:05:29.440 align:start position:0% According to Google's own published 00:05:29.440 --> 00:05:32.990 align:start position:0% According to Google's own published rates,<00:05:30.120> 3.6<00:05:30.840> Flash<00:05:31.480> runs<00:05:31.720> about<00:05:31.960> $1.50<00:05:32.840> per 00:05:32.990 --> 00:05:33.000 align:start position:0% rates, 3.6 Flash runs about $1.50 per 00:05:33.000 --> 00:05:35.590 align:start position:0% rates, 3.6 Flash runs about $1.50 per million<00:05:33.360> input<00:05:33.760> tokens<00:05:34.600> and<00:05:34.760> $7.50<00:05:35.440> per 00:05:35.590 --> 00:05:35.600 align:start position:0% million input tokens and $7.50 per 00:05:35.600 --> 00:05:38.230 align:start position:0% million input tokens and $7.50 per million<00:05:35.960> output<00:05:36.360> tokens.<00:05:37.120> For<00:05:37.280> context, 00:05:38.230 --> 00:05:38.240 align:start position:0% million output tokens. For context, 00:05:38.240 --> 00:05:40.150 align:start position:0% million output tokens. For context, that's<00:05:38.480> noticeably<00:05:39.080> cheaper<00:05:39.520> on<00:05:39.640> the<00:05:39.800> output 00:05:40.150 --> 00:05:40.160 align:start position:0% that's noticeably cheaper on the output 00:05:40.160 --> 00:05:44.310 align:start position:0% that's noticeably cheaper on the output side<00:05:40.760> than<00:05:40.960> GPT<00:05:41.960> 5.6<00:05:42.760> Luna's<00:05:43.240> roughly<00:05:43.600> $6<00:05:44.200> per 00:05:44.310 --> 00:05:44.320 align:start position:0% side than GPT 5.6 Luna's roughly $6 per 00:05:44.320 --> 00:05:46.990 align:start position:0% side than GPT 5.6 Luna's roughly $6 per million.<00:05:45.000> That<00:05:45.200> gap<00:05:45.560> is<00:05:45.720> exactly<00:05:46.280> why<00:05:46.480> so<00:05:46.760> many 00:05:46.990 --> 00:05:47.000 align:start position:0% million. That gap is exactly why so many 00:05:47.000 --> 00:05:49.470 align:start position:0% million. That gap is exactly why so many developers<00:05:47.680> default<00:05:48.160> to<00:05:48.280> flash<00:05:48.600> tier<00:05:48.800> models 00:05:49.470 --> 00:05:49.480 align:start position:0% developers default to flash tier models 00:05:49.480 --> 00:05:51.950 align:start position:0% developers default to flash tier models for<00:05:49.680> anything<00:05:50.080> running<00:05:50.320> at<00:05:50.440> volume.<00:05:51.200> Now,<00:05:51.880> a 00:05:51.950 --> 00:05:51.960 align:start position:0% for anything running at volume. Now, a 00:05:51.960 --> 00:05:53.670 align:start position:0% for anything running at volume. Now, a handful<00:05:52.360> of<00:05:52.480> specialty<00:05:52.960> models<00:05:53.440> worth 00:05:53.670 --> 00:05:53.680 align:start position:0% handful of specialty models worth 00:05:53.680 --> 00:05:55.910 align:start position:0% handful of specialty models worth knowing<00:05:53.960> by<00:05:54.120> name,<00:05:54.920> even<00:05:55.200> if<00:05:55.320> we<00:05:55.400> don't<00:05:55.640> dwell 00:05:55.910 --> 00:05:55.920 align:start position:0% knowing by name, even if we don't dwell 00:05:55.920 --> 00:05:58.670 align:start position:0% knowing by name, even if we don't dwell on<00:05:56.040> each<00:05:56.240> one.<00:05:56.800> Nano<00:05:57.080> Banana<00:05:57.480> 2<00:05:58.000> is<00:05:58.200> Gemini's 00:05:58.670 --> 00:05:58.680 align:start position:0% on each one. Nano Banana 2 is Gemini's 00:05:58.680 --> 00:06:00.710 align:start position:0% on each one. Nano Banana 2 is Gemini's current<00:05:59.040> image<00:05:59.360> generation<00:06:00.200> and<00:06:00.400> editing 00:06:00.710 --> 00:06:00.720 align:start position:0% current image generation and editing 00:06:00.720 --> 00:06:03.190 align:start position:0% current image generation and editing model,<00:06:01.480> replacing<00:06:02.000> the<00:06:02.120> older<00:06:02.440> Imagen<00:06:02.920> line 00:06:03.190 --> 00:06:03.200 align:start position:0% model, replacing the older Imagen line 00:06:03.200 --> 00:06:05.990 align:start position:0% model, replacing the older Imagen line entirely.<00:06:04.200> And<00:06:04.320> that's<00:06:04.560> not<00:06:04.720> a<00:06:04.760> small<00:06:05.080> detail 00:06:05.990 --> 00:06:06.000 align:start position:0% entirely. And that's not a small detail 00:06:06.000 --> 00:06:07.870 align:start position:0% entirely. And that's not a small detail because<00:06:06.280> Imagen<00:06:06.720> is<00:06:06.880> actually<00:06:07.240> shutting<00:06:07.560> down 00:06:07.870 --> 00:06:07.880 align:start position:0% because Imagen is actually shutting down 00:06:07.880 --> 00:06:10.910 align:start position:0% because Imagen is actually shutting down on<00:06:08.040> August<00:06:08.360> 17th,<00:06:09.160> 2026.<00:06:10.440> If<00:06:10.600> you've<00:06:10.720> had 00:06:10.910 --> 00:06:10.920 align:start position:0% on August 17th, 2026. If you've had 00:06:10.920 --> 00:06:13.270 align:start position:0% on August 17th, 2026. If you've had workflows<00:06:11.480> built<00:06:11.720> on<00:06:11.880> Imagen,<00:06:12.640> that<00:06:12.840> clock<00:06:13.120> is 00:06:13.270 --> 00:06:13.280 align:start position:0% workflows built on Imagen, that clock is 00:06:13.280 --> 00:06:15.190 align:start position:0% workflows built on Imagen, that clock is already<00:06:13.600> running.<00:06:14.280> There's<00:06:14.480> also<00:06:14.800> a<00:06:14.840> Nano 00:06:15.190 --> 00:06:15.200 align:start position:0% already running. There's also a Nano 00:06:15.200 --> 00:06:17.430 align:start position:0% already running. There's also a Nano Banana<00:06:15.600> 2<00:06:15.840> light<00:06:16.120> variant<00:06:16.640> built<00:06:16.920> purely<00:06:17.280> for 00:06:17.430 --> 00:06:17.440 align:start position:0% Banana 2 light variant built purely for 00:06:17.440 --> 00:06:19.909 align:start position:0% Banana 2 light variant built purely for speed,<00:06:18.120> trading<00:06:18.440> a<00:06:18.520> small<00:06:18.880> amount<00:06:19.160> of<00:06:19.280> quality 00:06:19.909 --> 00:06:19.919 align:start position:0% speed, trading a small amount of quality 00:06:19.919 --> 00:06:21.990 align:start position:0% speed, trading a small amount of quality for<00:06:20.040> much<00:06:20.320> faster,<00:06:20.800> cheaper<00:06:21.240> output<00:06:21.720> at<00:06:21.840> high 00:06:21.990 --> 00:06:22.000 align:start position:0% for much faster, cheaper output at high 00:06:22.000 --> 00:06:24.630 align:start position:0% for much faster, cheaper output at high volume.<00:06:22.760> VIO<00:06:23.120> 3.1<00:06:23.880> is<00:06:24.000> Google's<00:06:24.360> video 00:06:24.630 --> 00:06:24.640 align:start position:0% volume. VIO 3.1 is Google's video 00:06:24.640 --> 00:06:27.110 align:start position:0% volume. VIO 3.1 is Google's video generation<00:06:25.200> model,<00:06:25.760> still<00:06:26.040> in<00:06:26.160> beta,<00:06:26.919> built 00:06:27.110 --> 00:06:27.120 align:start position:0% generation model, still in beta, built 00:06:27.120 --> 00:06:29.150 align:start position:0% generation model, still in beta, built to<00:06:27.240> turn<00:06:27.440> a<00:06:27.480> text<00:06:27.800> prompt<00:06:28.120> into<00:06:28.360> a<00:06:28.440> short<00:06:28.720> clip 00:06:29.150 --> 00:06:29.160 align:start position:0% to turn a text prompt into a short clip 00:06:29.160 --> 00:06:32.350 align:start position:0% to turn a text prompt into a short clip with<00:06:29.320> matching<00:06:29.720> audio.<00:06:30.400> Gemini<00:06:30.840> audio<00:06:31.520> 3.5 00:06:32.350 --> 00:06:32.360 align:start position:0% with matching audio. Gemini audio 3.5 00:06:32.360 --> 00:06:34.350 align:start position:0% with matching audio. Gemini audio 3.5 live<00:06:32.600> translate<00:06:33.440> handles<00:06:33.800> real-time 00:06:34.350 --> 00:06:34.360 align:start position:0% live translate handles real-time 00:06:34.360 --> 00:06:36.909 align:start position:0% live translate handles real-time speech-to-speech<00:06:35.160> translation<00:06:36.360> across<00:06:36.720> more 00:06:36.909 --> 00:06:36.919 align:start position:0% speech-to-speech translation across more 00:06:36.919 --> 00:06:39.350 align:start position:0% speech-to-speech translation across more than<00:06:37.080> 70<00:06:37.520> languages,<00:06:38.560> already<00:06:38.880> built<00:06:39.160> into 00:06:39.350 --> 00:06:39.360 align:start position:0% than 70 languages, already built into 00:06:39.360 --> 00:06:41.150 align:start position:0% than 70 languages, already built into Google<00:06:39.680> Meet<00:06:39.960> and<00:06:40.120> Android.<00:06:40.840> And<00:06:40.960> if<00:06:41.040> you're 00:06:41.150 --> 00:06:41.160 align:start position:0% Google Meet and Android. And if you're 00:06:41.160 --> 00:06:42.909 align:start position:0% Google Meet and Android. And if you're curious<00:06:41.600> about<00:06:41.840> the<00:06:41.919> more<00:06:42.160> niche<00:06:42.560> end<00:06:42.760> of<00:06:42.840> the 00:06:42.909 --> 00:06:42.919 align:start position:0% curious about the more niche end of the 00:06:42.919 --> 00:06:45.070 align:start position:0% curious about the more niche end of the lineup,<00:06:43.680> there's<00:06:43.880> a<00:06:43.960> security-focused 00:06:45.070 --> 00:06:45.080 align:start position:0% lineup, there's a security-focused 00:06:45.080 --> 00:06:48.190 align:start position:0% lineup, there's a security-focused variant<00:06:45.480> called<00:06:45.760> 3.5<00:06:46.680> flash<00:06:47.040> cyber<00:06:47.360> built<00:06:48.080> to 00:06:48.190 --> 00:06:48.200 align:start position:0% variant called 3.5 flash cyber built to 00:06:48.200 --> 00:06:49.990 align:start position:0% variant called 3.5 flash cyber built to coordinate<00:06:48.760> with<00:06:48.919> vulnerability<00:06:49.600> scanning 00:06:49.990 --> 00:06:50.000 align:start position:0% coordinate with vulnerability scanning 00:06:50.000 --> 00:06:53.510 align:start position:0% coordinate with vulnerability scanning tools.<00:06:50.919> And<00:06:51.120> Lyra<00:06:51.640> 3.5,<00:06:52.800> Google's<00:06:53.200> music 00:06:53.510 --> 00:06:53.520 align:start position:0% tools. And Lyra 3.5, Google's music 00:06:53.520 --> 00:06:55.710 align:start position:0% tools. And Lyra 3.5, Google's music model,<00:06:54.240> which<00:06:54.440> can<00:06:54.560> now<00:06:54.760> generate<00:06:55.200> tracks<00:06:55.600> up 00:06:55.710 --> 00:06:55.720 align:start position:0% model, which can now generate tracks up 00:06:55.720 --> 00:06:57.990 align:start position:0% model, which can now generate tracks up to<00:06:55.840> 3<00:06:56.120> minutes<00:06:56.440> long<00:06:56.800> from<00:06:57.000> a<00:06:57.040> text<00:06:57.360> prompt. 00:06:57.990 --> 00:06:58.000 align:start position:0% to 3 minutes long from a text prompt. 00:06:58.000 --> 00:06:59.590 align:start position:0% to 3 minutes long from a text prompt. Here's<00:06:58.240> the<00:06:58.400> honest<00:06:58.720> limitation<00:06:59.360> worth 00:06:59.590 --> 00:06:59.600 align:start position:0% Here's the honest limitation worth 00:06:59.600 --> 00:07:01.550 align:start position:0% Here's the honest limitation worth naming.<00:07:00.240> Google<00:07:00.560> ships<00:07:00.919> a<00:07:01.000> lot<00:07:01.280> of<00:07:01.360> these 00:07:01.550 --> 00:07:01.560 align:start position:0% naming. Google ships a lot of these 00:07:01.560 --> 00:07:03.510 align:start position:0% naming. Google ships a lot of these models<00:07:01.919> fast,<00:07:02.720> and<00:07:02.840> the<00:07:02.919> naming<00:07:03.280> gets 00:07:03.510 --> 00:07:03.520 align:start position:0% models fast, and the naming gets 00:07:03.520 --> 00:07:07.550 align:start position:0% models fast, and the naming gets confusing<00:07:04.080> on<00:07:04.200> purpose<00:07:04.600> or<00:07:04.720> not.<00:07:05.280> 3.5,<00:07:06.360> 3.6, 00:07:07.550 --> 00:07:07.560 align:start position:0% confusing on purpose or not. 3.5, 3.6, 00:07:07.560 --> 00:07:11.150 align:start position:0% confusing on purpose or not. 3.5, 3.6, 3.1<00:07:08.240> pro,<00:07:08.919> flashlight,<00:07:09.880> flash<00:07:10.200> cyber.<00:07:11.000> If 00:07:11.150 --> 00:07:11.160 align:start position:0% 3.1 pro, flashlight, flash cyber. If 00:07:11.160 --> 00:07:12.950 align:start position:0% 3.1 pro, flashlight, flash cyber. If you're<00:07:11.280> not<00:07:11.480> building<00:07:11.840> on<00:07:11.960> top<00:07:12.240> of<00:07:12.320> the<00:07:12.480> API 00:07:12.950 --> 00:07:12.960 align:start position:0% you're not building on top of the API 00:07:12.960 --> 00:07:15.310 align:start position:0% you're not building on top of the API professionally,<00:07:14.080> you<00:07:14.240> genuinely<00:07:15.000> don't<00:07:15.120> need 00:07:15.310 --> 00:07:15.320 align:start position:0% professionally, you genuinely don't need 00:07:15.320 --> 00:07:17.470 align:start position:0% professionally, you genuinely don't need to<00:07:15.440> memorize<00:07:15.919> this<00:07:16.120> list.<00:07:16.840> You<00:07:16.960> just<00:07:17.200> need<00:07:17.400> to 00:07:17.470 --> 00:07:17.480 align:start position:0% to memorize this list. You just need to 00:07:17.480 --> 00:07:19.710 align:start position:0% to memorize this list. You just need to know<00:07:17.640> the<00:07:17.760> shape<00:07:18.080> of<00:07:18.240> it.<00:07:18.680> Fast<00:07:19.040> and<00:07:19.160> cheap, 00:07:19.710 --> 00:07:19.720 align:start position:0% know the shape of it. Fast and cheap, 00:07:19.720 --> 00:07:21.990 align:start position:0% know the shape of it. Fast and cheap, deep<00:07:19.960> reasoning,<00:07:20.640> and<00:07:20.800> multimodal.<00:07:21.760> That's 00:07:21.990 --> 00:07:22.000 align:start position:0% deep reasoning, and multimodal. That's 00:07:22.000 --> 00:07:23.870 align:start position:0% deep reasoning, and multimodal. That's really<00:07:22.240> three<00:07:22.560> categories<00:07:23.240> wearing<00:07:23.480> a<00:07:23.520> lot<00:07:23.760> of 00:07:23.870 --> 00:07:23.880 align:start position:0% really three categories wearing a lot of 00:07:23.880 --> 00:07:26.030 align:start position:0% really three categories wearing a lot of different<00:07:24.200> name<00:07:24.440> tags.<00:07:25.320> The<00:07:25.440> modes<00:07:25.880> you 00:07:26.030 --> 00:07:26.040 align:start position:0% different name tags. The modes you 00:07:26.040 --> 00:07:28.390 align:start position:0% different name tags. The modes you actually<00:07:26.520> interact<00:07:27.040> with,<00:07:27.440> models<00:07:28.080> are<00:07:28.240> the 00:07:28.390 --> 00:07:28.400 align:start position:0% actually interact with, models are the 00:07:28.400 --> 00:07:30.590 align:start position:0% actually interact with, models are the engine.<00:07:29.080> Modes<00:07:29.520> are<00:07:29.640> the<00:07:29.760> steering<00:07:30.120> wheel. 00:07:30.590 --> 00:07:30.600 align:start position:0% engine. Modes are the steering wheel. 00:07:30.600 --> 00:07:32.070 align:start position:0% engine. Modes are the steering wheel. Here's<00:07:30.880> where<00:07:31.040> things<00:07:31.320> get<00:07:31.520> useful<00:07:31.880> for 00:07:32.070 --> 00:07:32.080 align:start position:0% Here's where things get useful for 00:07:32.080 --> 00:07:34.750 align:start position:0% Here's where things get useful for anyone<00:07:32.440> who<00:07:32.640> isn't<00:07:32.960> a<00:07:33.000> developer.<00:07:33.880> AI<00:07:34.200> mode 00:07:34.750 --> 00:07:34.760 align:start position:0% anyone who isn't a developer. AI mode 00:07:34.760 --> 00:07:36.590 align:start position:0% anyone who isn't a developer. AI mode inside<00:07:35.200> Google<00:07:35.480> Search<00:07:35.919> turns<00:07:36.200> your<00:07:36.320> search 00:07:36.590 --> 00:07:36.600 align:start position:0% inside Google Search turns your search 00:07:36.600 --> 00:07:40.350 align:start position:0% inside Google Search turns your search bar<00:07:36.919> into<00:07:37.160> a<00:07:37.200> conversation.<00:07:38.320> As<00:07:38.560> of<00:07:38.800> IO<00:07:39.120> 2026, 00:07:40.350 --> 00:07:40.360 align:start position:0% bar into a conversation. As of IO 2026, 00:07:40.360 --> 00:07:42.950 align:start position:0% bar into a conversation. As of IO 2026, it's<00:07:40.560> globally<00:07:41.040> powered<00:07:41.360> by<00:07:41.520> Gemini<00:07:42.000> 3.5 00:07:42.950 --> 00:07:42.960 align:start position:0% it's globally powered by Gemini 3.5 00:07:42.960 --> 00:07:45.550 align:start position:0% it's globally powered by Gemini 3.5 flash<00:07:43.720> and<00:07:43.840> instead<00:07:44.160> of<00:07:44.280> 10<00:07:44.520> blue<00:07:44.760> links,<00:07:45.440> you 00:07:45.550 --> 00:07:45.560 align:start position:0% flash and instead of 10 blue links, you 00:07:45.560 --> 00:07:47.030 align:start position:0% flash and instead of 10 blue links, you get<00:07:45.720> a<00:07:45.760> written<00:07:46.080> answer<00:07:46.440> with<00:07:46.640> follow-up 00:07:47.030 --> 00:07:47.040 align:start position:0% get a written answer with follow-up 00:07:47.040 --> 00:07:49.230 align:start position:0% get a written answer with follow-up questions<00:07:47.840> and<00:07:48.000> sometimes<00:07:48.520> an<00:07:48.640> interactive 00:07:49.230 --> 00:07:49.240 align:start position:0% questions and sometimes an interactive 00:07:49.240 --> 00:07:51.150 align:start position:0% questions and sometimes an interactive widget<00:07:49.600> built<00:07:49.840> on<00:07:49.960> the<00:07:50.040> fly.<00:07:50.640> Anyone<00:07:50.960> with 00:07:51.150 --> 00:07:51.160 align:start position:0% widget built on the fly. Anyone with 00:07:51.160 --> 00:07:52.990 align:start position:0% widget built on the fly. Anyone with search<00:07:51.480> can<00:07:51.640> use<00:07:51.840> it.<00:07:52.240> No<00:07:52.440> subscription 00:07:52.990 --> 00:07:53.000 align:start position:0% search can use it. No subscription 00:07:53.000 --> 00:07:55.230 align:start position:0% search can use it. No subscription required.<00:07:53.840> Ask<00:07:54.160> something<00:07:54.480> like,<00:07:54.960> "What's<00:07:55.200> a 00:07:55.230 --> 00:07:55.240 align:start position:0% required. Ask something like, "What's a 00:07:55.240 --> 00:07:57.110 align:start position:0% required. Ask something like, "What's a quick<00:07:55.480> dinner<00:07:55.800> with<00:07:55.960> what's<00:07:56.160> in<00:07:56.240> my<00:07:56.400> fridge?" 00:07:57.110 --> 00:07:57.120 align:start position:0% quick dinner with what's in my fridge?" 00:07:57.120 --> 00:07:59.310 align:start position:0% quick dinner with what's in my fridge?" and<00:07:57.280> it<00:07:57.400> answers<00:07:57.800> in<00:07:57.920> full<00:07:58.160> sentences,<00:07:59.080> not<00:07:59.280> a 00:07:59.310 --> 00:07:59.320 align:start position:0% and it answers in full sentences, not a 00:07:59.320 --> 00:08:01.710 align:start position:0% and it answers in full sentences, not a list<00:07:59.560> of<00:07:59.680> recipe<00:08:00.040> blogs.<00:08:00.760> Deep<00:08:01.000> Think<00:08:01.400> is<00:08:01.560> the 00:08:01.710 --> 00:08:01.720 align:start position:0% list of recipe blogs. Deep Think is the 00:08:01.720 --> 00:08:04.070 align:start position:0% list of recipe blogs. Deep Think is the extra<00:08:02.080> effort<00:08:02.360> version<00:08:02.760> of<00:08:02.920> the<00:08:03.040> Gemini<00:08:03.480> app. 00:08:04.070 --> 00:08:04.080 align:start position:0% extra effort version of the Gemini app. 00:08:04.080 --> 00:08:06.110 align:start position:0% extra effort version of the Gemini app. It<00:08:04.240> spends<00:08:04.600> more<00:08:04.800> compute<00:08:05.280> per<00:08:05.520> answer<00:08:05.880> to 00:08:06.110 --> 00:08:06.120 align:start position:0% It spends more compute per answer to 00:08:06.120 --> 00:08:07.830 align:start position:0% It spends more compute per answer to reason<00:08:06.760> through<00:08:07.000> harder<00:08:07.320> problems 00:08:07.830 --> 00:08:07.840 align:start position:0% reason through harder problems 00:08:07.840 --> 00:08:10.070 align:start position:0% reason through harder problems step-by-step.<00:08:08.880> Google<00:08:09.200> gates<00:08:09.480> this<00:08:09.720> one 00:08:10.070 --> 00:08:10.080 align:start position:0% step-by-step. Google gates this one 00:08:10.080 --> 00:08:12.270 align:start position:0% step-by-step. Google gates this one behind<00:08:10.280> Google<00:08:10.640> AI<00:08:10.920> Ultra<00:08:11.720> and<00:08:11.840> it's<00:08:12.000> built 00:08:12.270 --> 00:08:12.280 align:start position:0% behind Google AI Ultra and it's built 00:08:12.280 --> 00:08:14.150 align:start position:0% behind Google AI Ultra and it's built for<00:08:12.440> genuinely<00:08:13.000> difficult<00:08:13.520> science<00:08:14.000> or 00:08:14.150 --> 00:08:14.160 align:start position:0% for genuinely difficult science or 00:08:14.160 --> 00:08:16.190 align:start position:0% for genuinely difficult science or engineering<00:08:14.680> questions,<00:08:15.520> not<00:08:15.800> everyday 00:08:16.190 --> 00:08:16.200 align:start position:0% engineering questions, not everyday 00:08:16.200 --> 00:08:18.750 align:start position:0% engineering questions, not everyday chat.<00:08:16.880> Now,<00:08:17.480> Deep<00:08:17.760> Research<00:08:18.240> is<00:08:18.400> where<00:08:18.560> this 00:08:18.750 --> 00:08:18.760 align:start position:0% chat. Now, Deep Research is where this 00:08:18.760 --> 00:08:20.830 align:start position:0% chat. Now, Deep Research is where this stops<00:08:19.120> being<00:08:19.360> a<00:08:19.440> chatbot<00:08:20.120> and<00:08:20.280> starts<00:08:20.600> being 00:08:20.830 --> 00:08:20.840 align:start position:0% stops being a chatbot and starts being 00:08:20.840 --> 00:08:22.950 align:start position:0% stops being a chatbot and starts being an<00:08:20.919> assistant.<00:08:21.680> You<00:08:21.840> give<00:08:22.000> it<00:08:22.080> a<00:08:22.160> topic<00:08:22.840> and 00:08:22.950 --> 00:08:22.960 align:start position:0% an assistant. You give it a topic and 00:08:22.960 --> 00:08:24.950 align:start position:0% an assistant. You give it a topic and instead<00:08:23.280> of<00:08:23.400> one<00:08:23.600> reply,<00:08:24.400> it<00:08:24.560> plans<00:08:24.880> a 00:08:24.950 --> 00:08:24.960 align:start position:0% instead of one reply, it plans a 00:08:24.960 --> 00:08:27.670 align:start position:0% instead of one reply, it plans a research<00:08:25.400> strategy,<00:08:26.360> opens<00:08:26.720> web<00:08:26.960> pages, 00:08:27.670 --> 00:08:27.680 align:start position:0% research strategy, opens web pages, 00:08:27.680 --> 00:08:30.590 align:start position:0% research strategy, opens web pages, reads<00:08:28.000> them<00:08:28.680> and<00:08:29.240> if<00:08:29.400> you<00:08:29.520> allow<00:08:29.840> it,<00:08:30.360> pulls 00:08:30.590 --> 00:08:30.600 align:start position:0% reads them and if you allow it, pulls 00:08:30.600 --> 00:08:32.790 align:start position:0% reads them and if you allow it, pulls from<00:08:30.760> your<00:08:30.919> own<00:08:31.120> Gmail<00:08:31.600> and<00:08:31.720> Drive,<00:08:32.080> too.<00:08:32.640> What 00:08:32.790 --> 00:08:32.800 align:start position:0% from your own Gmail and Drive, too. What 00:08:32.800 --> 00:08:34.830 align:start position:0% from your own Gmail and Drive, too. What comes<00:08:33.039> back<00:08:33.360> isn't<00:08:33.599> a<00:08:33.640> paragraph.<00:08:34.599> It's<00:08:34.760> a 00:08:34.830 --> 00:08:34.840 align:start position:0% comes back isn't a paragraph. It's a 00:08:34.840 --> 00:08:37.310 align:start position:0% comes back isn't a paragraph. It's a full<00:08:35.159> multi-page<00:08:35.760> report<00:08:36.440> inside<00:08:36.880> Gemini's 00:08:37.310 --> 00:08:37.320 align:start position:0% full multi-page report inside Gemini's 00:08:37.320 --> 00:08:39.750 align:start position:0% full multi-page report inside Gemini's canvas.<00:08:38.240> This<00:08:38.440> is<00:08:38.560> the<00:08:38.640> part<00:08:38.880> of<00:08:39.000> Gemini<00:08:39.479> that 00:08:39.750 --> 00:08:39.760 align:start position:0% canvas. This is the part of Gemini that 00:08:39.760 --> 00:08:42.469 align:start position:0% canvas. This is the part of Gemini that actually<00:08:40.240> earns<00:08:40.560> the<00:08:40.680> word<00:08:41.159> agent<00:08:42.200> and<00:08:42.360> we're 00:08:42.469 --> 00:08:42.479 align:start position:0% actually earns the word agent and we're 00:08:42.479 --> 00:08:44.310 align:start position:0% actually earns the word agent and we're coming<00:08:42.760> back<00:08:43.039> to<00:08:43.159> why<00:08:43.360> that<00:08:43.560> matters<00:08:43.960> in<00:08:44.080> a<00:08:44.120> few 00:08:44.310 --> 00:08:44.320 align:start position:0% coming back to why that matters in a few 00:08:44.320 --> 00:08:46.630 align:start position:0% coming back to why that matters in a few minutes.<00:08:45.040> Gemini<00:08:45.480> Live<00:08:45.880> is<00:08:46.040> the<00:08:46.160> voice<00:08:46.520> and 00:08:46.630 --> 00:08:46.640 align:start position:0% minutes. Gemini Live is the voice and 00:08:46.640 --> 00:08:49.110 align:start position:0% minutes. Gemini Live is the voice and camera<00:08:47.000> mode.<00:08:47.640> Say,<00:08:48.040> "Hey<00:08:48.240> Google,<00:08:48.839> let's 00:08:49.110 --> 00:08:49.120 align:start position:0% camera mode. Say, "Hey Google, let's 00:08:49.120 --> 00:08:50.550 align:start position:0% camera mode. Say, "Hey Google, let's chat"<00:08:49.720> and<00:08:49.839> you're<00:08:49.960> talking<00:08:50.320> to<00:08:50.440> it 00:08:50.550 --> 00:08:50.560 align:start position:0% chat" and you're talking to it 00:08:50.560 --> 00:08:52.630 align:start position:0% chat" and you're talking to it hands-free<00:08:51.520> with<00:08:51.640> the<00:08:51.839> option<00:08:52.160> to<00:08:52.280> point<00:08:52.520> your 00:08:52.630 --> 00:08:52.640 align:start position:0% hands-free with the option to point your 00:08:52.640 --> 00:08:54.350 align:start position:0% hands-free with the option to point your camera<00:08:53.040> at<00:08:53.160> something<00:08:53.640> and<00:08:53.800> ask<00:08:54.080> what<00:08:54.200> it's 00:08:54.350 --> 00:08:54.360 align:start position:0% camera at something and ask what it's 00:08:54.360 --> 00:08:56.710 align:start position:0% camera at something and ask what it's looking<00:08:54.680> at<00:08:55.120> live.<00:08:55.800> And<00:08:55.960> Canvas<00:08:56.480> is<00:08:56.640> the 00:08:56.710 --> 00:08:56.720 align:start position:0% looking at live. And Canvas is the 00:08:56.720 --> 00:08:59.310 align:start position:0% looking at live. And Canvas is the workspace<00:08:57.240> mode.<00:08:57.880> Type,<00:08:58.440> "Create<00:08:58.800> a<00:08:58.880> quiz<00:08:59.160> app 00:08:59.310 --> 00:08:59.320 align:start position:0% workspace mode. Type, "Create a quiz app 00:08:59.320 --> 00:09:01.710 align:start position:0% workspace mode. Type, "Create a quiz app about<00:08:59.600> planets"<00:09:00.440> and<00:09:00.560> it<00:09:00.680> writes<00:09:00.920> the<00:09:01.000> code, 00:09:01.710 --> 00:09:01.720 align:start position:0% about planets" and it writes the code, 00:09:01.720 --> 00:09:03.710 align:start position:0% about planets" and it writes the code, the<00:09:01.880> interface<00:09:02.680> and<00:09:02.800> the<00:09:02.880> content<00:09:03.400> in<00:09:03.520> one 00:09:03.710 --> 00:09:03.720 align:start position:0% the interface and the content in one 00:09:03.720 --> 00:09:06.230 align:start position:0% the interface and the content in one pass,<00:09:04.680> right<00:09:04.920> there<00:09:05.120> for<00:09:05.280> you<00:09:05.400> to<00:09:05.520> edit.<00:09:06.040> One 00:09:06.230 --> 00:09:06.240 align:start position:0% pass, right there for you to edit. One 00:09:06.240 --> 00:09:08.310 align:start position:0% pass, right there for you to edit. One more<00:09:06.440> worth<00:09:06.640> a<00:09:06.720> mention<00:09:07.360> briefly<00:09:08.040> because 00:09:08.310 --> 00:09:08.320 align:start position:0% more worth a mention briefly because 00:09:08.320 --> 00:09:10.670 align:start position:0% more worth a mention briefly because it's<00:09:08.440> still<00:09:08.720> early.<00:09:09.360> Gemini<00:09:09.800> Spark,<00:09:10.600> a 00:09:10.670 --> 00:09:10.680 align:start position:0% it's still early. Gemini Spark, a 00:09:10.680 --> 00:09:14.070 align:start position:0% it's still early. Gemini Spark, a personal<00:09:11.160> agent<00:09:11.520> announced<00:09:11.960> at<00:09:12.200> IO<00:09:12.600> 2026, 00:09:14.070 --> 00:09:14.080 align:start position:0% personal agent announced at IO 2026, 00:09:14.080 --> 00:09:15.510 align:start position:0% personal agent announced at IO 2026, meant<00:09:14.320> to<00:09:14.440> run<00:09:14.640> continuously<00:09:15.360> in<00:09:15.440> the 00:09:15.510 --> 00:09:15.520 align:start position:0% meant to run continuously in the 00:09:15.520 --> 00:09:16.950 align:start position:0% meant to run continuously in the background<00:09:16.080> handling<00:09:16.480> things<00:09:16.800> like 00:09:16.950 --> 00:09:16.960 align:start position:0% background handling things like 00:09:16.960 --> 00:09:18.990 align:start position:0% background handling things like scheduling.<00:09:17.880> Right<00:09:18.120> now,<00:09:18.320> it's<00:09:18.480> limited<00:09:18.880> to 00:09:18.990 --> 00:09:19.000 align:start position:0% scheduling. Right now, it's limited to 00:09:19.000 --> 00:09:21.270 align:start position:0% scheduling. Right now, it's limited to early<00:09:19.320> Ultra<00:09:19.600> testers.<00:09:20.520> So,<00:09:20.680> treat<00:09:20.920> this<00:09:21.120> one 00:09:21.270 --> 00:09:21.280 align:start position:0% early Ultra testers. So, treat this one 00:09:21.280 --> 00:09:23.670 align:start position:0% early Ultra testers. So, treat this one as<00:09:21.440> coming,<00:09:22.080> not<00:09:22.360> here.<00:09:22.960> Quick<00:09:23.240> gut<00:09:23.440> check 00:09:23.670 --> 00:09:23.680 align:start position:0% as coming, not here. Quick gut check 00:09:23.680 --> 00:09:25.870 align:start position:0% as coming, not here. Quick gut check before<00:09:23.960> we<00:09:24.080> move<00:09:24.360> on.<00:09:24.960> If<00:09:25.120> all<00:09:25.280> of<00:09:25.400> that<00:09:25.600> sounds 00:09:25.870 --> 00:09:25.880 align:start position:0% before we move on. If all of that sounds 00:09:25.880 --> 00:09:27.710 align:start position:0% before we move on. If all of that sounds like<00:09:26.040> a<00:09:26.080> lot<00:09:26.320> of<00:09:26.400> separate<00:09:26.760> tools,<00:09:27.440> that's 00:09:27.710 --> 00:09:27.720 align:start position:0% like a lot of separate tools, that's 00:09:27.720 --> 00:09:29.870 align:start position:0% like a lot of separate tools, that's fair.<00:09:28.280> But,<00:09:28.440> notice<00:09:28.760> the<00:09:28.839> pattern.<00:09:29.600> Every 00:09:29.870 --> 00:09:29.880 align:start position:0% fair. But, notice the pattern. Every 00:09:29.880 --> 00:09:31.950 align:start position:0% fair. But, notice the pattern. Every single<00:09:30.240> one<00:09:30.400> of<00:09:30.480> these<00:09:30.720> modes<00:09:31.080> is<00:09:31.240> just<00:09:31.480> Gemini 00:09:31.950 --> 00:09:31.960 align:start position:0% single one of these modes is just Gemini 00:09:31.960 --> 00:09:35.150 align:start position:0% single one of these modes is just Gemini 3.5<00:09:32.880> or<00:09:33.000> 3.6<00:09:33.839> flash<00:09:34.280> wearing<00:09:34.560> a<00:09:34.600> different<00:09:34.920> job 00:09:35.150 --> 00:09:35.160 align:start position:0% 3.5 or 3.6 flash wearing a different job 00:09:35.160 --> 00:09:37.230 align:start position:0% 3.5 or 3.6 flash wearing a different job title.<00:09:35.800> You're<00:09:35.920> not<00:09:36.160> learning<00:09:36.480> six<00:09:36.760> different 00:09:37.230 --> 00:09:37.240 align:start position:0% title. You're not learning six different 00:09:37.240 --> 00:09:39.550 align:start position:0% title. You're not learning six different AIs,<00:09:38.160> you're<00:09:38.320> learning<00:09:38.640> six<00:09:38.920> different<00:09:39.200> ways 00:09:39.550 --> 00:09:39.560 align:start position:0% AIs, you're learning six different ways 00:09:39.560 --> 00:09:41.390 align:start position:0% AIs, you're learning six different ways to<00:09:39.680> ask<00:09:39.960> the<00:09:40.040> same<00:09:40.320> brain<00:09:40.640> for<00:09:40.760> help. 00:09:41.390 --> 00:09:41.400 align:start position:0% to ask the same brain for help. 00:09:41.400 --> 00:09:43.710 align:start position:0% to ask the same brain for help. Multimodal<00:09:42.040> in<00:09:42.160> practice.<00:09:43.080> Let's<00:09:43.280> talk<00:09:43.480> about 00:09:43.710 --> 00:09:43.720 align:start position:0% Multimodal in practice. Let's talk about 00:09:43.720 --> 00:09:45.390 align:start position:0% Multimodal in practice. Let's talk about what<00:09:43.920> multimodal<00:09:44.680> actually<00:09:45.080> means 00:09:45.390 --> 00:09:45.400 align:start position:0% what multimodal actually means 00:09:45.400 --> 00:09:46.990 align:start position:0% what multimodal actually means day-to-day,<00:09:46.160> because<00:09:46.440> it's<00:09:46.600> more<00:09:46.800> than<00:09:46.960> a 00:09:46.990 --> 00:09:47.000 align:start position:0% day-to-day, because it's more than a 00:09:47.000 --> 00:09:49.350 align:start position:0% day-to-day, because it's more than a buzzword<00:09:47.520> on<00:09:47.640> a<00:09:47.680> slide.<00:09:48.440> Gemini<00:09:48.880> reads<00:09:49.240> and 00:09:49.350 --> 00:09:49.360 align:start position:0% buzzword on a slide. Gemini reads and 00:09:49.360 --> 00:09:51.710 align:start position:0% buzzword on a slide. Gemini reads and writes<00:09:49.640> text<00:09:49.960> and<00:09:50.080> code.<00:09:50.680> Obviously,<00:09:51.480> that's 00:09:51.710 --> 00:09:51.720 align:start position:0% writes text and code. Obviously, that's 00:09:51.720 --> 00:09:53.670 align:start position:0% writes text and code. Obviously, that's the<00:09:51.800> baseline.<00:09:52.640> But<00:09:52.800> drop<00:09:53.000> a<00:09:53.080> photo<00:09:53.400> into<00:09:53.640> a 00:09:53.670 --> 00:09:53.680 align:start position:0% the baseline. But drop a photo into a 00:09:53.680 --> 00:09:56.110 align:start position:0% the baseline. But drop a photo into a chat<00:09:54.240> and<00:09:54.400> ask<00:09:54.600> it<00:09:54.720> to<00:09:54.800> caption<00:09:55.200> or<00:09:55.360> edit<00:09:55.600> it, 00:09:56.110 --> 00:09:56.120 align:start position:0% chat and ask it to caption or edit it, 00:09:56.120 --> 00:09:58.230 align:start position:0% chat and ask it to caption or edit it, and<00:09:56.280> Nano<00:09:56.560> Banana<00:09:56.960> handles<00:09:57.320> that.<00:09:57.800> Ask<00:09:58.040> it<00:09:58.160> to 00:09:58.230 --> 00:09:58.240 align:start position:0% and Nano Banana handles that. Ask it to 00:09:58.240 --> 00:10:00.470 align:start position:0% and Nano Banana handles that. Ask it to speak<00:09:58.520> an<00:09:58.640> answer<00:09:58.960> out<00:09:59.160> loud,<00:09:59.840> and<00:09:59.960> Gemini's 00:10:00.470 --> 00:10:00.480 align:start position:0% speak an answer out loud, and Gemini's 00:10:00.480 --> 00:10:02.390 align:start position:0% speak an answer out loud, and Gemini's audio<00:10:00.800> models<00:10:01.240> generate<00:10:01.680> that<00:10:01.920> voice<00:10:02.200> on<00:10:02.320> the 00:10:02.390 --> 00:10:02.400 align:start position:0% audio models generate that voice on the 00:10:02.400 --> 00:10:05.030 align:start position:0% audio models generate that voice on the spot<00:10:03.080> with<00:10:03.360> actual<00:10:03.760> control<00:10:04.240> over<00:10:04.440> tone<00:10:04.840> and 00:10:05.030 --> 00:10:05.040 align:start position:0% spot with actual control over tone and 00:10:05.040 --> 00:10:07.670 align:start position:0% spot with actual control over tone and pacing.<00:10:05.760> Ask<00:10:06.040> for<00:10:06.200> a<00:10:06.240> short<00:10:06.520> video<00:10:07.240> and<00:10:07.400> VO 00:10:07.670 --> 00:10:07.680 align:start position:0% pacing. Ask for a short video and VO 00:10:07.680 --> 00:10:09.870 align:start position:0% pacing. Ask for a short video and VO builds<00:10:07.960> one<00:10:08.120> from<00:10:08.280> scratch.<00:10:09.080> Ask<00:10:09.360> it<00:10:09.480> to<00:10:09.600> edit 00:10:09.870 --> 00:10:09.880 align:start position:0% builds one from scratch. Ask it to edit 00:10:09.880 --> 00:10:12.510 align:start position:0% builds one from scratch. Ask it to edit an<00:10:10.000> existing<00:10:10.480> clip,<00:10:11.120> swap<00:10:11.480> the<00:10:11.600> sky,<00:10:12.200> change 00:10:12.510 --> 00:10:12.520 align:start position:0% an existing clip, swap the sky, change 00:10:12.520 --> 00:10:14.470 align:start position:0% an existing clip, swap the sky, change the<00:10:12.640> style,<00:10:13.360> and<00:10:13.520> that's<00:10:13.720> a<00:10:13.800> separate<00:10:14.200> tool 00:10:14.470 --> 00:10:14.480 align:start position:0% the style, and that's a separate tool 00:10:14.480 --> 00:10:17.070 align:start position:0% the style, and that's a separate tool called<00:10:14.800> Gemini<00:10:15.320> Omni<00:10:15.960> doing<00:10:16.240> frame-by-frame 00:10:17.070 --> 00:10:17.080 align:start position:0% called Gemini Omni doing frame-by-frame 00:10:17.080 --> 00:10:19.350 align:start position:0% called Gemini Omni doing frame-by-frame editing<00:10:17.440> by<00:10:17.600> voice<00:10:17.880> command.<00:10:18.600> Inside<00:10:19.080> Google 00:10:19.350 --> 00:10:19.360 align:start position:0% editing by voice command. Inside Google 00:10:19.360 --> 00:10:21.750 align:start position:0% editing by voice command. Inside Google Docs<00:10:19.720> and<00:10:19.840> Sheets,<00:10:20.760> the<00:10:20.880> same<00:10:21.240> underlying 00:10:21.750 --> 00:10:21.760 align:start position:0% Docs and Sheets, the same underlying 00:10:21.760 --> 00:10:23.550 align:start position:0% Docs and Sheets, the same underlying models<00:10:22.200> can<00:10:22.320> draft<00:10:22.680> a<00:10:22.720> document<00:10:23.280> from<00:10:23.440> your 00:10:23.550 --> 00:10:23.560 align:start position:0% models can draft a document from your 00:10:23.560 --> 00:10:25.870 align:start position:0% models can draft a document from your meeting<00:10:23.880> notes<00:10:24.680> or<00:10:24.800> build<00:10:25.040> a<00:10:25.120> spreadsheet<00:10:25.760> out 00:10:25.870 --> 00:10:25.880 align:start position:0% meeting notes or build a spreadsheet out 00:10:25.880 --> 00:10:28.150 align:start position:0% meeting notes or build a spreadsheet out of<00:10:26.000> a<00:10:26.080> pile<00:10:26.400> of<00:10:26.520> invoices,<00:10:27.600> complete<00:10:28.000> with 00:10:28.150 --> 00:10:28.160 align:start position:0% of a pile of invoices, complete with 00:10:28.160 --> 00:10:30.390 align:start position:0% of a pile of invoices, complete with formulas<00:10:28.720> and<00:10:28.840> charts.<00:10:29.760> Not<00:10:30.000> just<00:10:30.240> raw 00:10:30.390 --> 00:10:30.400 align:start position:0% formulas and charts. Not just raw 00:10:30.400 --> 00:10:33.150 align:start position:0% formulas and charts. Not just raw numbers<00:10:30.800> dumped<00:10:31.080> into<00:10:31.280> cells.<00:10:32.080> In<00:10:32.240> Slides, 00:10:33.150 --> 00:10:33.160 align:start position:0% numbers dumped into cells. In Slides, 00:10:33.160 --> 00:10:35.030 align:start position:0% numbers dumped into cells. In Slides, hand<00:10:33.400> it<00:10:33.520> a<00:10:33.560> list<00:10:33.800> of<00:10:33.920> bullet<00:10:34.240> points<00:10:34.800> and<00:10:34.920> it 00:10:35.030 --> 00:10:35.040 align:start position:0% hand it a list of bullet points and it 00:10:35.040 --> 00:10:37.270 align:start position:0% hand it a list of bullet points and it can<00:10:35.200> lay<00:10:35.400> out<00:10:35.520> an<00:10:35.640> actual<00:10:36.040> deck,<00:10:36.800> not<00:10:37.040> just 00:10:37.270 --> 00:10:37.280 align:start position:0% can lay out an actual deck, not just 00:10:37.280 --> 00:10:39.550 align:start position:0% can lay out an actual deck, not just text<00:10:37.600> on<00:10:37.720> blank<00:10:38.000> slides.<00:10:38.840> And<00:10:38.960> through<00:10:39.160> Gemini 00:10:39.550 --> 00:10:39.560 align:start position:0% text on blank slides. And through Gemini 00:10:39.560 --> 00:10:41.550 align:start position:0% text on blank slides. And through Gemini Live's<00:10:39.880> camera<00:10:40.240> mode,<00:10:40.920> you<00:10:41.040> can<00:10:41.160> point<00:10:41.400> your 00:10:41.550 --> 00:10:41.560 align:start position:0% Live's camera mode, you can point your 00:10:41.560 --> 00:10:43.510 align:start position:0% Live's camera mode, you can point your phone<00:10:41.920> at<00:10:42.040> a<00:10:42.080> menu<00:10:42.520> in<00:10:42.640> a<00:10:42.720> language<00:10:43.160> you<00:10:43.240> don't 00:10:43.510 --> 00:10:43.520 align:start position:0% phone at a menu in a language you don't 00:10:43.520 --> 00:10:45.390 align:start position:0% phone at a menu in a language you don't speak<00:10:44.120> and<00:10:44.240> get<00:10:44.400> a<00:10:44.440> live<00:10:44.720> translation 00:10:45.390 --> 00:10:45.400 align:start position:0% speak and get a live translation 00:10:45.400 --> 00:10:47.270 align:start position:0% speak and get a live translation overlaid<00:10:45.880> on<00:10:46.000> what<00:10:46.200> you're<00:10:46.320> looking<00:10:46.680> at.<00:10:47.080> Or 00:10:47.270 --> 00:10:47.280 align:start position:0% overlaid on what you're looking at. Or 00:10:47.280 --> 00:10:49.070 align:start position:0% overlaid on what you're looking at. Or ask<00:10:47.560> it<00:10:47.640> to<00:10:47.760> identify<00:10:48.280> an<00:10:48.440> object<00:10:48.880> it's 00:10:49.070 --> 00:10:49.080 align:start position:0% ask it to identify an object it's 00:10:49.080 --> 00:10:51.110 align:start position:0% ask it to identify an object it's looking<00:10:49.360> at<00:10:49.520> through<00:10:49.680> the<00:10:49.800> lens.<00:10:50.560> No<00:10:50.760> typing 00:10:51.110 --> 00:10:51.120 align:start position:0% looking at through the lens. No typing 00:10:51.120 --> 00:10:52.710 align:start position:0% looking at through the lens. No typing involved.<00:10:51.880> Here's<00:10:52.120> the<00:10:52.200> part<00:10:52.480> worth 00:10:52.710 --> 00:10:52.720 align:start position:0% involved. Here's the part worth 00:10:52.720 --> 00:10:55.030 align:start position:0% involved. Here's the part worth remembering.<00:10:53.560> You<00:10:53.680> never<00:10:54.000> pick<00:10:54.240> the<00:10:54.360> model. 00:10:55.030 --> 00:10:55.040 align:start position:0% remembering. You never pick the model. 00:10:55.040 --> 00:10:56.910 align:start position:0% remembering. You never pick the model. You<00:10:55.160> just<00:10:55.400> say<00:10:55.560> what<00:10:55.720> you<00:10:55.839> want.<00:10:56.400> Make<00:10:56.600> this<00:10:56.800> an 00:10:56.910 --> 00:10:56.920 align:start position:0% You just say what you want. Make this an 00:10:56.920 --> 00:10:59.710 align:start position:0% You just say what you want. Make this an infographic.<00:10:58.000> Translate<00:10:58.560> this.<00:10:59.240> Write<00:10:59.520> this 00:10:59.710 --> 00:10:59.720 align:start position:0% infographic. Translate this. Write this 00:10:59.720 --> 00:11:02.190 align:start position:0% infographic. Translate this. Write this in<00:10:59.839> Python.<00:11:00.640> And<00:11:00.800> Gemini<00:11:01.240> quietly<00:11:01.760> roots<00:11:02.120> the 00:11:02.190 --> 00:11:02.200 align:start position:0% in Python. And Gemini quietly roots the 00:11:02.200 --> 00:11:04.270 align:start position:0% in Python. And Gemini quietly roots the request<00:11:02.680> to<00:11:02.800> whichever<00:11:03.160> model<00:11:03.600> actually<00:11:04.040> does 00:11:04.270 --> 00:11:04.280 align:start position:0% request to whichever model actually does 00:11:04.280 --> 00:11:06.470 align:start position:0% request to whichever model actually does that<00:11:04.520> job.<00:11:05.200> That's<00:11:05.480> the<00:11:05.560> design<00:11:05.920> philosophy 00:11:06.470 --> 00:11:06.480 align:start position:0% that job. That's the design philosophy 00:11:06.480 --> 00:11:09.070 align:start position:0% that job. That's the design philosophy in<00:11:06.600> one<00:11:06.800> sentence.<00:11:07.760> One<00:11:08.000> platform,<00:11:08.839> and<00:11:08.960> it 00:11:09.070 --> 00:11:09.080 align:start position:0% in one sentence. One platform, and it 00:11:09.080 --> 00:11:10.870 align:start position:0% in one sentence. One platform, and it decides<00:11:09.480> the<00:11:09.600> plumbing<00:11:10.160> so<00:11:10.320> you<00:11:10.440> don't<00:11:10.640> have 00:11:10.870 --> 00:11:10.880 align:start position:0% decides the plumbing so you don't have 00:11:10.880 --> 00:11:13.470 align:start position:0% decides the plumbing so you don't have to.<00:11:11.400> Where<00:11:11.560> Gemini<00:11:12.040> actually<00:11:12.520> lives.<00:11:13.160> This<00:11:13.360> is 00:11:13.470 --> 00:11:13.480 align:start position:0% to. Where Gemini actually lives. This is 00:11:13.480 --> 00:11:15.550 align:start position:0% to. Where Gemini actually lives. This is the<00:11:13.560> part<00:11:13.839> that's<00:11:14.040> easy<00:11:14.320> to<00:11:14.440> underestimate. 00:11:15.550 --> 00:11:15.560 align:start position:0% the part that's easy to underestimate. 00:11:15.560 --> 00:11:18.110 align:start position:0% the part that's easy to underestimate. Gemini<00:11:16.000> isn't<00:11:16.320> confined<00:11:16.800> to<00:11:16.920> one<00:11:17.200> app.<00:11:17.920> It's 00:11:18.110 --> 00:11:18.120 align:start position:0% Gemini isn't confined to one app. It's 00:11:18.120 --> 00:11:19.829 align:start position:0% Gemini isn't confined to one app. It's spread<00:11:18.400> across<00:11:18.760> nearly<00:11:19.080> everything<00:11:19.520> Google 00:11:19.829 --> 00:11:19.839 align:start position:0% spread across nearly everything Google 00:11:19.839 --> 00:11:22.550 align:start position:0% spread across nearly everything Google ships.<00:11:20.600> In<00:11:20.760> Search,<00:11:21.160> it's<00:11:21.440> AI<00:11:21.680> mode,<00:11:22.240> already 00:11:22.550 --> 00:11:22.560 align:start position:0% ships. In Search, it's AI mode, already 00:11:22.560 --> 00:11:24.950 align:start position:0% ships. In Search, it's AI mode, already covered.<00:11:23.440> In<00:11:23.600> Gmail,<00:11:24.160> it's<00:11:24.320> behind<00:11:24.680> Smart 00:11:24.950 --> 00:11:24.960 align:start position:0% covered. In Gmail, it's behind Smart 00:11:24.960 --> 00:11:27.430 align:start position:0% covered. In Gmail, it's behind Smart Compose<00:11:25.480> and<00:11:25.640> auto-reply<00:11:26.240> suggestions.<00:11:27.280> In 00:11:27.430 --> 00:11:27.440 align:start position:0% Compose and auto-reply suggestions. In 00:11:27.440 --> 00:11:29.990 align:start position:0% Compose and auto-reply suggestions. In Docs,<00:11:27.880> Sheets<00:11:28.240> and<00:11:28.400> Slides,<00:11:29.280> Ultra<00:11:29.680> and<00:11:29.800> Pro 00:11:29.990 --> 00:11:30.000 align:start position:0% Docs, Sheets and Slides, Ultra and Pro 00:11:30.000 --> 00:11:30.710 align:start position:0% Docs, Sheets and Slides, Ultra and Pro Pro 00:11:30.710 --> 00:11:30.720 align:start position:0% Pro 00:11:30.720 --> 00:11:32.990 align:start position:0% Pro get<00:11:30.920> Gemini<00:11:31.320> drafting<00:11:31.880> text,<00:11:32.640> building 00:11:32.990 --> 00:11:33.000 align:start position:0% get Gemini drafting text, building 00:11:33.000 --> 00:11:35.830 align:start position:0% get Gemini drafting text, building formulas,<00:11:34.000> and<00:11:34.120> designing<00:11:34.600> slide<00:11:34.960> layouts, 00:11:35.830 --> 00:11:35.840 align:start position:0% formulas, and designing slide layouts, 00:11:35.840 --> 00:11:37.790 align:start position:0% formulas, and designing slide layouts, pulling<00:11:36.160> context<00:11:36.720> from<00:11:36.880> your<00:11:37.040> own<00:11:37.240> files<00:11:37.680> when 00:11:37.790 --> 00:11:37.800 align:start position:0% pulling context from your own files when 00:11:37.800 --> 00:11:40.110 align:start position:0% pulling context from your own files when you<00:11:37.920> let<00:11:38.120> it.<00:11:38.440> In<00:11:38.600> Drive,<00:11:39.360> it<00:11:39.480> can<00:11:39.640> find<00:11:39.960> and 00:11:40.110 --> 00:11:40.120 align:start position:0% you let it. In Drive, it can find and 00:11:40.120 --> 00:11:42.270 align:start position:0% you let it. In Drive, it can find and summarize<00:11:40.640> documents<00:11:41.160> for<00:11:41.280> you.<00:11:41.880> In<00:11:42.000> Google 00:11:42.270 --> 00:11:42.280 align:start position:0% summarize documents for you. In Google 00:11:42.280 --> 00:11:43.750 align:start position:0% summarize documents for you. In Google Meet,<00:11:42.640> it's<00:11:42.800> doing<00:11:43.040> live<00:11:43.360> caption 00:11:43.750 --> 00:11:43.760 align:start position:0% Meet, it's doing live caption 00:11:43.760 --> 00:11:46.110 align:start position:0% Meet, it's doing live caption translation.<00:11:44.760> On<00:11:44.960> Android,<00:11:45.600> especially 00:11:46.110 --> 00:11:46.120 align:start position:0% translation. On Android, especially 00:11:46.120 --> 00:11:48.350 align:start position:0% translation. On Android, especially Pixel<00:11:46.440> devices,<00:11:47.280> it's<00:11:47.480> baked<00:11:47.760> straight<00:11:48.160> into 00:11:48.350 --> 00:11:48.360 align:start position:0% Pixel devices, it's baked straight into 00:11:48.360 --> 00:11:49.910 align:start position:0% Pixel devices, it's baked straight into the<00:11:48.440> voice<00:11:48.760> assistant.<00:11:49.520> And<00:11:49.680> there's<00:11:49.840> a 00:11:49.910 --> 00:11:49.920 align:start position:0% the voice assistant. And there's a 00:11:49.920 --> 00:11:51.670 align:start position:0% the voice assistant. And there's a Chrome<00:11:50.240> extension<00:11:50.800> that<00:11:50.960> lets<00:11:51.160> the<00:11:51.280> browser 00:11:51.670 --> 00:11:51.680 align:start position:0% Chrome extension that lets the browser 00:11:51.680 --> 00:11:54.070 align:start position:0% Chrome extension that lets the browser send<00:11:51.960> page<00:11:52.240> content<00:11:52.720> straight<00:11:53.080> to<00:11:53.200> Gemini,<00:11:53.960> so 00:11:54.070 --> 00:11:54.080 align:start position:0% send page content straight to Gemini, so 00:11:54.080 --> 00:11:55.990 align:start position:0% send page content straight to Gemini, so you<00:11:54.200> can<00:11:54.360> ask<00:11:54.560> questions<00:11:55.080> about<00:11:55.360> whatever<00:11:55.720> tab 00:11:55.990 --> 00:11:56.000 align:start position:0% you can ask questions about whatever tab 00:11:56.000 --> 00:11:58.270 align:start position:0% you can ask questions about whatever tab you're<00:11:56.160> on.<00:11:56.680> And<00:11:56.800> for<00:11:56.920> developers,<00:11:57.880> all<00:11:58.080> of<00:11:58.160> it 00:11:58.270 --> 00:11:58.280 align:start position:0% you're on. And for developers, all of it 00:11:58.280 --> 00:12:00.670 align:start position:0% you're on. And for developers, all of it is<00:11:58.400> exposed<00:11:58.920> through<00:11:59.120> Google<00:11:59.520> AI<00:11:59.920> Studio<00:12:00.560> and 00:12:00.670 --> 00:12:00.680 align:start position:0% is exposed through Google AI Studio and 00:12:00.680 --> 00:12:03.270 align:start position:0% is exposed through Google AI Studio and the<00:12:00.760> Gemini<00:12:01.240> API,<00:12:02.120> plus<00:12:02.400> a<00:12:02.440> newer<00:12:02.720> platform 00:12:03.270 --> 00:12:03.280 align:start position:0% the Gemini API, plus a newer platform 00:12:03.280 --> 00:12:05.110 align:start position:0% the Gemini API, plus a newer platform called<00:12:03.600> antigravity<00:12:04.600> for<00:12:04.760> building 00:12:05.110 --> 00:12:05.120 align:start position:0% called antigravity for building 00:12:05.120 --> 00:12:07.310 align:start position:0% called antigravity for building multi-agent<00:12:05.800> workflows<00:12:06.320> on<00:12:06.440> top<00:12:06.720> of<00:12:06.840> it.<00:12:07.200> The 00:12:07.310 --> 00:12:07.320 align:start position:0% multi-agent workflows on top of it. The 00:12:07.320 --> 00:12:09.350 align:start position:0% multi-agent workflows on top of it. The strategic<00:12:07.840> point<00:12:08.120> here<00:12:08.440> isn't<00:12:08.760> subtle. 00:12:09.350 --> 00:12:09.360 align:start position:0% strategic point here isn't subtle. 00:12:09.360 --> 00:12:10.910 align:start position:0% strategic point here isn't subtle. Google<00:12:09.720> isn't<00:12:09.960> trying<00:12:10.240> to<00:12:10.320> win<00:12:10.520> the<00:12:10.600> best 00:12:10.910 --> 00:12:10.920 align:start position:0% Google isn't trying to win the best 00:12:10.920 --> 00:12:13.070 align:start position:0% Google isn't trying to win the best standalone<00:12:11.400> chatbot<00:12:11.920> argument.<00:12:12.600> It's<00:12:12.800> trying 00:12:13.070 --> 00:12:13.080 align:start position:0% standalone chatbot argument. It's trying 00:12:13.080 --> 00:12:14.590 align:start position:0% standalone chatbot argument. It's trying to<00:12:13.160> make<00:12:13.360> sure<00:12:13.520> you're<00:12:13.680> never<00:12:14.000> more<00:12:14.200> than<00:12:14.400> one 00:12:14.590 --> 00:12:14.600 align:start position:0% to make sure you're never more than one 00:12:14.600 --> 00:12:16.870 align:start position:0% to make sure you're never more than one product<00:12:15.000> away<00:12:15.240> from<00:12:15.440> Gemini,<00:12:16.280> no<00:12:16.400> matter<00:12:16.680> what 00:12:16.870 --> 00:12:16.880 align:start position:0% product away from Gemini, no matter what 00:12:16.880 --> 00:12:18.710 align:start position:0% product away from Gemini, no matter what you're<00:12:17.000> doing<00:12:17.320> on<00:12:17.440> a<00:12:17.480> Google<00:12:17.800> device<00:12:18.400> or<00:12:18.560> in<00:12:18.640> a 00:12:18.710 --> 00:12:18.720 align:start position:0% you're doing on a Google device or in a 00:12:18.720 --> 00:12:21.150 align:start position:0% you're doing on a Google device or in a Google<00:12:19.080> app.<00:12:19.680> Agents,<00:12:20.560> the<00:12:20.680> part<00:12:20.960> that 00:12:21.150 --> 00:12:21.160 align:start position:0% Google app. Agents, the part that 00:12:21.160 --> 00:12:23.270 align:start position:0% Google app. Agents, the part that actually<00:12:21.560> matters.<00:12:22.360> Now,<00:12:22.520> here's<00:12:22.800> the<00:12:22.920> shift 00:12:23.270 --> 00:12:23.280 align:start position:0% actually matters. Now, here's the shift 00:12:23.280 --> 00:12:24.950 align:start position:0% actually matters. Now, here's the shift I<00:12:23.320> promised<00:12:23.800> earlier,<00:12:24.480> the<00:12:24.600> one<00:12:24.800> that 00:12:24.950 --> 00:12:24.960 align:start position:0% I promised earlier, the one that 00:12:24.960 --> 00:12:26.870 align:start position:0% I promised earlier, the one that actually<00:12:25.360> changes<00:12:25.880> what<00:12:26.040> this<00:12:26.240> platform<00:12:26.720> is 00:12:26.870 --> 00:12:26.880 align:start position:0% actually changes what this platform is 00:12:26.880 --> 00:12:29.590 align:start position:0% actually changes what this platform is for.<00:12:27.480> Everything<00:12:27.960> so<00:12:28.200> far<00:12:28.720> has<00:12:28.920> been<00:12:29.200> ask<00:12:29.520> a 00:12:29.590 --> 00:12:29.600 align:start position:0% for. Everything so far has been ask a 00:12:29.600 --> 00:12:31.950 align:start position:0% for. Everything so far has been ask a question,<00:12:30.480> get<00:12:30.640> an<00:12:30.760> answer.<00:12:31.440> Agents<00:12:31.880> are 00:12:31.950 --> 00:12:31.960 align:start position:0% question, get an answer. Agents are 00:12:31.960 --> 00:12:33.870 align:start position:0% question, get an answer. Agents are Google<00:12:32.280> trying<00:12:32.560> to<00:12:32.640> move<00:12:32.880> Gemini<00:12:33.320> past<00:12:33.680> that 00:12:33.870 --> 00:12:33.880 align:start position:0% Google trying to move Gemini past that 00:12:33.880 --> 00:12:36.270 align:start position:0% Google trying to move Gemini past that entirely.<00:12:34.800> Deep<00:12:35.040> research<00:12:35.560> is<00:12:35.720> the<00:12:35.839> clearest 00:12:36.270 --> 00:12:36.280 align:start position:0% entirely. Deep research is the clearest 00:12:36.280 --> 00:12:39.270 align:start position:0% entirely. Deep research is the clearest example<00:12:36.839> already<00:12:37.240> live,<00:12:37.960> plan,<00:12:38.640> browse, 00:12:39.270 --> 00:12:39.280 align:start position:0% example already live, plan, browse, 00:12:39.280 --> 00:12:41.350 align:start position:0% example already live, plan, browse, synthesize,<00:12:40.240> write,<00:12:40.920> without<00:12:41.240> you 00:12:41.350 --> 00:12:41.360 align:start position:0% synthesize, write, without you 00:12:41.360 --> 00:12:43.710 align:start position:0% synthesize, write, without you babysitting<00:12:41.920> every<00:12:42.160> step.<00:12:42.839> Spark<00:12:43.360> is<00:12:43.480> the 00:12:43.710 --> 00:12:43.720 align:start position:0% babysitting every step. Spark is the 00:12:43.720 --> 00:12:45.470 align:start position:0% babysitting every step. Spark is the early,<00:12:44.080> still<00:12:44.480> limited<00:12:44.880> attempt<00:12:45.320> at<00:12:45.440> a 00:12:45.470 --> 00:12:45.480 align:start position:0% early, still limited attempt at a 00:12:45.480 --> 00:12:47.350 align:start position:0% early, still limited attempt at a persistent<00:12:46.080> personal<00:12:46.560> agent<00:12:47.080> running 00:12:47.350 --> 00:12:47.360 align:start position:0% persistent personal agent running 00:12:47.360 --> 00:12:49.470 align:start position:0% persistent personal agent running continuously<00:12:48.120> in<00:12:48.240> the<00:12:48.320> background.<00:12:49.240> And<00:12:49.360> on 00:12:49.470 --> 00:12:49.480 align:start position:0% continuously in the background. And on 00:12:49.480 --> 00:12:51.550 align:start position:0% continuously in the background. And on the<00:12:49.560> developer<00:12:50.040> side,<00:12:50.600> antigravity<00:12:51.360> lets 00:12:51.550 --> 00:12:51.560 align:start position:0% the developer side, antigravity lets 00:12:51.560 --> 00:12:53.430 align:start position:0% the developer side, antigravity lets companies<00:12:52.080> build<00:12:52.320> coordinated<00:12:52.960> teams<00:12:53.320> of 00:12:53.430 --> 00:12:53.440 align:start position:0% companies build coordinated teams of 00:12:53.440 --> 00:12:55.829 align:start position:0% companies build coordinated teams of sub-agents.<00:12:54.400> Google's<00:12:54.800> own<00:12:55.000> blog<00:12:55.320> post<00:12:55.680> gave 00:12:55.829 --> 00:12:55.839 align:start position:0% sub-agents. Google's own blog post gave 00:12:55.839 --> 00:12:57.350 align:start position:0% sub-agents. Google's own blog post gave an<00:12:55.920> example<00:12:56.400> of<00:12:56.520> businesses<00:12:57.080> running 00:12:57.350 --> 00:12:57.360 align:start position:0% an example of businesses running 00:12:57.360 --> 00:12:59.350 align:start position:0% an example of businesses running parallel<00:12:57.839> agents<00:12:58.280> to<00:12:58.400> analyze<00:12:58.920> data<00:12:59.200> at 00:12:59.350 --> 00:12:59.360 align:start position:0% parallel agents to analyze data at 00:12:59.360 --> 00:13:01.590 align:start position:0% parallel agents to analyze data at scale,<00:13:00.280> rather<00:13:00.600> than<00:13:00.800> one<00:13:01.000> model<00:13:01.320> doing 00:13:01.590 --> 00:13:01.600 align:start position:0% scale, rather than one model doing 00:13:01.600 --> 00:13:03.430 align:start position:0% scale, rather than one model doing everything<00:13:02.000> sequentially.<00:13:02.960> Picture<00:13:03.320> the 00:13:03.430 --> 00:13:03.440 align:start position:0% everything sequentially. Picture the 00:13:03.440 --> 00:13:05.950 align:start position:0% everything sequentially. Picture the difference<00:13:03.800> in<00:13:03.920> practice.<00:13:04.800> The<00:13:04.960> old<00:13:05.200> way,<00:13:05.800> you 00:13:05.950 --> 00:13:05.960 align:start position:0% difference in practice. The old way, you 00:13:05.960 --> 00:13:08.110 align:start position:0% difference in practice. The old way, you ask<00:13:06.240> Gemini,<00:13:07.080> "What<00:13:07.280> should<00:13:07.520> I<00:13:07.560> know<00:13:07.800> before<00:13:08.080> a 00:13:08.110 --> 00:13:08.120 align:start position:0% ask Gemini, "What should I know before a 00:13:08.120 --> 00:13:09.790 align:start position:0% ask Gemini, "What should I know before a trip<00:13:08.360> to<00:13:08.480> Japan?"<00:13:09.240> and<00:13:09.360> it<00:13:09.440> gives<00:13:09.640> you<00:13:09.760> a 00:13:09.790 --> 00:13:09.800 align:start position:0% trip to Japan?" and it gives you a 00:13:09.800 --> 00:13:12.590 align:start position:0% trip to Japan?" and it gives you a paragraph.<00:13:10.640> The<00:13:10.800> agent<00:13:11.120> way,<00:13:11.600> you<00:13:11.760> say,<00:13:12.320> "Plan 00:13:12.590 --> 00:13:12.600 align:start position:0% paragraph. The agent way, you say, "Plan 00:13:12.600 --> 00:13:14.550 align:start position:0% paragraph. The agent way, you say, "Plan my<00:13:12.720> trip<00:13:13.000> to<00:13:13.120> Japan."<00:13:14.040> and<00:13:14.160> it<00:13:14.280> checks 00:13:14.550 --> 00:13:14.560 align:start position:0% my trip to Japan." and it checks 00:13:14.560 --> 00:13:17.070 align:start position:0% my trip to Japan." and it checks flights,<00:13:15.360> compares<00:13:15.760> hotel<00:13:16.160> options,<00:13:16.960> and 00:13:17.070 --> 00:13:17.080 align:start position:0% flights, compares hotel options, and 00:13:17.080 --> 00:13:19.230 align:start position:0% flights, compares hotel options, and comes<00:13:17.360> back<00:13:17.600> with<00:13:17.760> an<00:13:17.920> actual<00:13:18.320> itinerary, 00:13:19.230 --> 00:13:19.240 align:start position:0% comes back with an actual itinerary, 00:13:19.240 --> 00:13:20.870 align:start position:0% comes back with an actual itinerary, pausing<00:13:19.640> to<00:13:19.720> confirm<00:13:20.160> with<00:13:20.320> you<00:13:20.480> before<00:13:20.760> it 00:13:20.870 --> 00:13:20.880 align:start position:0% pausing to confirm with you before it 00:13:20.880 --> 00:13:22.550 align:start position:0% pausing to confirm with you before it books<00:13:21.120> anything.<00:13:21.880> That's<00:13:22.120> the<00:13:22.240> same 00:13:22.550 --> 00:13:22.560 align:start position:0% books anything. That's the same 00:13:22.560 --> 00:13:24.630 align:start position:0% books anything. That's the same underlying<00:13:23.080> model,<00:13:23.760> just<00:13:23.960> given<00:13:24.240> permission 00:13:24.630 --> 00:13:24.640 align:start position:0% underlying model, just given permission 00:13:24.640 --> 00:13:26.390 align:start position:0% underlying model, just given permission to<00:13:24.760> take<00:13:25.000> more<00:13:25.200> than<00:13:25.360> one<00:13:25.560> step<00:13:26.080> before 00:13:26.390 --> 00:13:26.400 align:start position:0% to take more than one step before 00:13:26.400 --> 00:13:28.230 align:start position:0% to take more than one step before handing<00:13:26.720> control<00:13:27.080> back<00:13:27.360> to<00:13:27.480> you.<00:13:27.960> None<00:13:28.160> of 00:13:28.230 --> 00:13:28.240 align:start position:0% handing control back to you. None of 00:13:28.240 --> 00:13:30.110 align:start position:0% handing control back to you. None of this<00:13:28.440> is<00:13:28.560> science<00:13:28.880> fiction<00:13:29.200> anymore,<00:13:29.960> and 00:13:30.110 --> 00:13:30.120 align:start position:0% this is science fiction anymore, and 00:13:30.120 --> 00:13:31.950 align:start position:0% this is science fiction anymore, and none<00:13:30.280> of<00:13:30.400> it<00:13:30.520> is<00:13:30.640> fully<00:13:30.960> finished,<00:13:31.360> either. 00:13:31.950 --> 00:13:31.960 align:start position:0% none of it is fully finished, either. 00:13:31.960 --> 00:13:33.990 align:start position:0% none of it is fully finished, either. That's<00:13:32.160> the<00:13:32.320> honest<00:13:32.640> read.<00:13:33.200> Deep<00:13:33.480> Research 00:13:33.990 --> 00:13:34.000 align:start position:0% That's the honest read. Deep Research 00:13:34.000 --> 00:13:36.430 align:start position:0% That's the honest read. Deep Research genuinely<00:13:34.600> works<00:13:34.880> today.<00:13:35.560> Spark<00:13:35.920> is<00:13:36.040> still<00:13:36.320> in 00:13:36.430 --> 00:13:36.440 align:start position:0% genuinely works today. Spark is still in 00:13:36.440 --> 00:13:38.510 align:start position:0% genuinely works today. Spark is still in early<00:13:36.720> testing,<00:13:37.480> but<00:13:37.640> the<00:13:37.720> direction<00:13:38.280> is 00:13:38.510 --> 00:13:38.520 align:start position:0% early testing, but the direction is 00:13:38.520 --> 00:13:40.790 align:start position:0% early testing, but the direction is unmistakable.<00:13:39.600> Google<00:13:39.920> wants<00:13:40.200> Gemini<00:13:40.640> to 00:13:40.790 --> 00:13:40.800 align:start position:0% unmistakable. Google wants Gemini to 00:13:40.800 --> 00:13:42.950 align:start position:0% unmistakable. Google wants Gemini to eventually<00:13:41.280> take<00:13:41.480> a<00:13:41.560> task,<00:13:42.400> break<00:13:42.640> it<00:13:42.760> into 00:13:42.950 --> 00:13:42.960 align:start position:0% eventually take a task, break it into 00:13:42.960 --> 00:13:45.510 align:start position:0% eventually take a task, break it into steps,<00:13:43.760> and<00:13:43.920> execute<00:13:44.400> most<00:13:44.680> of<00:13:44.760> them<00:13:45.160> without 00:13:45.510 --> 00:13:45.520 align:start position:0% steps, and execute most of them without 00:13:45.520 --> 00:13:47.390 align:start position:0% steps, and execute most of them without you<00:13:45.640> typing<00:13:46.000> a<00:13:46.080> follow-up<00:13:46.640> for<00:13:46.840> every<00:13:47.080> single 00:13:47.390 --> 00:13:47.400 align:start position:0% you typing a follow-up for every single 00:13:47.400 --> 00:13:49.430 align:start position:0% you typing a follow-up for every single one.<00:13:48.000> What<00:13:48.320> actually<00:13:48.760> makes<00:13:49.040> Gemini 00:13:49.430 --> 00:13:49.440 align:start position:0% one. What actually makes Gemini 00:13:49.440 --> 00:13:51.270 align:start position:0% one. What actually makes Gemini different?<00:13:50.120> So,<00:13:50.320> how<00:13:50.440> does<00:13:50.640> this<00:13:50.800> stack<00:13:51.120> up 00:13:51.270 --> 00:13:51.280 align:start position:0% different? So, how does this stack up 00:13:51.280 --> 00:13:52.990 align:start position:0% different? So, how does this stack up against<00:13:51.640> everyone<00:13:52.040> else<00:13:52.320> building<00:13:52.640> the<00:13:52.760> same 00:13:52.990 --> 00:13:53.000 align:start position:0% against everyone else building the same 00:13:53.000 --> 00:13:55.550 align:start position:0% against everyone else building the same kind<00:13:53.280> of<00:13:53.400> thing?<00:13:54.040> Let's<00:13:54.240> be<00:13:54.360> balanced<00:13:54.840> here, 00:13:55.550 --> 00:13:55.560 align:start position:0% kind of thing? Let's be balanced here, 00:13:55.560 --> 00:13:57.510 align:start position:0% kind of thing? Let's be balanced here, because<00:13:55.800> Google's<00:13:56.160> advantages<00:13:56.720> are<00:13:56.840> real, 00:13:57.510 --> 00:13:57.520 align:start position:0% because Google's advantages are real, 00:13:57.520 --> 00:13:59.710 align:start position:0% because Google's advantages are real, but<00:13:57.680> so<00:13:57.880> are<00:13:58.000> its<00:13:58.160> weak<00:13:58.400> spots.<00:13:59.160> The<00:13:59.240> clearest 00:13:59.710 --> 00:13:59.720 align:start position:0% but so are its weak spots. The clearest 00:13:59.720 --> 00:14:02.110 align:start position:0% but so are its weak spots. The clearest edge<00:13:59.960> is<00:14:00.120> data.<00:14:00.800> Gemini<00:14:01.160> can<00:14:01.360> pull<00:14:01.640> from<00:14:01.880> live 00:14:02.110 --> 00:14:02.120 align:start position:0% edge is data. Gemini can pull from live 00:14:02.120 --> 00:14:05.430 align:start position:0% edge is data. Gemini can pull from live search<00:14:02.400> results,<00:14:03.200> Maps,<00:14:03.800> Gmail,<00:14:04.440> and<00:14:04.600> Drive 00:14:05.430 --> 00:14:05.440 align:start position:0% search results, Maps, Gmail, and Drive 00:14:05.440 --> 00:14:08.190 align:start position:0% search results, Maps, Gmail, and Drive in<00:14:05.600> ways<00:14:05.920> that<00:14:06.040> a<00:14:06.120> closed<00:14:06.760> sandbox<00:14:07.440> chatbot 00:14:08.190 --> 00:14:08.200 align:start position:0% in ways that a closed sandbox chatbot 00:14:08.200 --> 00:14:09.950 align:start position:0% in ways that a closed sandbox chatbot simply<00:14:08.560> can't<00:14:08.880> match<00:14:09.240> without<00:14:09.560> plugins 00:14:09.950 --> 00:14:09.960 align:start position:0% simply can't match without plugins 00:14:09.960 --> 00:14:12.430 align:start position:0% simply can't match without plugins bolted<00:14:10.320> on.<00:14:10.840> The<00:14:10.960> second<00:14:11.320> edge<00:14:11.560> is<00:14:11.760> reach. 00:14:12.430 --> 00:14:12.440 align:start position:0% bolted on. The second edge is reach. 00:14:12.440 --> 00:14:14.270 align:start position:0% bolted on. The second edge is reach. Every<00:14:12.760> Android<00:14:13.200> phone<00:14:13.560> is<00:14:13.760> a<00:14:13.800> potential 00:14:14.270 --> 00:14:14.280 align:start position:0% Every Android phone is a potential 00:14:14.280 --> 00:14:16.470 align:start position:0% Every Android phone is a potential Gemini<00:14:14.680> client,<00:14:15.440> and<00:14:15.640> every<00:14:15.960> Workspace 00:14:16.470 --> 00:14:16.480 align:start position:0% Gemini client, and every Workspace 00:14:16.480 --> 00:14:18.070 align:start position:0% Gemini client, and every Workspace business<00:14:16.840> account<00:14:17.320> already<00:14:17.760> has<00:14:17.960> it 00:14:18.070 --> 00:14:18.080 align:start position:0% business account already has it 00:14:18.080 --> 00:14:20.510 align:start position:0% business account already has it available.<00:14:18.920> No<00:14:19.120> competitor<00:14:19.880> has<00:14:20.080> that<00:14:20.240> kind 00:14:20.510 --> 00:14:20.520 align:start position:0% available. No competitor has that kind 00:14:20.520 --> 00:14:22.470 align:start position:0% available. No competitor has that kind of<00:14:20.640> built-in<00:14:21.000> distribution.<00:14:22.000> And<00:14:22.120> on<00:14:22.280> raw 00:14:22.470 --> 00:14:22.480 align:start position:0% of built-in distribution. And on raw 00:14:22.480 --> 00:14:25.230 align:start position:0% of built-in distribution. And on raw benchmarks,<00:14:23.400> Gemini<00:14:23.800> 3<00:14:24.080> Pro<00:14:24.400> topped<00:14:24.800> the<00:14:25.080> LM 00:14:25.230 --> 00:14:25.240 align:start position:0% benchmarks, Gemini 3 Pro topped the LM 00:14:25.240 --> 00:14:27.670 align:start position:0% benchmarks, Gemini 3 Pro topped the LM Arena<00:14:25.560> leaderboard,<00:14:26.440> which,<00:14:27.080> regardless<00:14:27.560> of 00:14:27.670 --> 00:14:27.680 align:start position:0% Arena leaderboard, which, regardless of 00:14:27.680 --> 00:14:29.350 align:start position:0% Arena leaderboard, which, regardless of how<00:14:27.840> much<00:14:28.080> weight<00:14:28.320> you<00:14:28.440> put<00:14:28.640> on<00:14:28.840> any<00:14:29.040> single 00:14:29.350 --> 00:14:29.360 align:start position:0% how much weight you put on any single 00:14:29.360 --> 00:14:30.829 align:start position:0% how much weight you put on any single leaderboard,<00:14:30.200> says<00:14:30.480> Google's 00:14:30.829 --> 00:14:30.839 align:start position:0% leaderboard, says Google's 00:14:30.839 --> 00:14:32.990 align:start position:0% leaderboard, says Google's infrastructure<00:14:31.560> and<00:14:31.720> DeepMind's<00:14:32.160> research 00:14:32.990 --> 00:14:33.000 align:start position:0% infrastructure and DeepMind's research 00:14:33.000 --> 00:14:35.470 align:start position:0% infrastructure and DeepMind's research are<00:14:33.120> producing<00:14:33.680> real,<00:14:34.160> top-tier<00:14:34.640> results, 00:14:35.470 --> 00:14:35.480 align:start position:0% are producing real, top-tier results, 00:14:35.480 --> 00:14:37.990 align:start position:0% are producing real, top-tier results, not<00:14:35.680> just<00:14:35.920> hype.<00:14:36.560> But,<00:14:37.280> and<00:14:37.400> this<00:14:37.520> matters<00:14:37.880> for 00:14:37.990 --> 00:14:38.000 align:start position:0% not just hype. But, and this matters for 00:14:38.000 --> 00:14:40.310 align:start position:0% not just hype. But, and this matters for credibility,<00:14:38.920> Google<00:14:39.240> is<00:14:39.400> genuinely<00:14:40.040> more 00:14:40.310 --> 00:14:40.320 align:start position:0% credibility, Google is genuinely more 00:14:40.320 --> 00:14:41.910 align:start position:0% credibility, Google is genuinely more conservative<00:14:40.920> about<00:14:41.200> rollout<00:14:41.600> than<00:14:41.720> some 00:14:41.910 --> 00:14:41.920 align:start position:0% conservative about rollout than some 00:14:41.920 --> 00:14:44.030 align:start position:0% conservative about rollout than some competitors.<00:14:42.839> Deep<00:14:43.120> Think<00:14:43.400> and<00:14:43.560> Spark<00:14:43.960> are 00:14:44.030 --> 00:14:44.040 align:start position:0% competitors. Deep Think and Spark are 00:14:44.040 --> 00:14:45.990 align:start position:0% competitors. Deep Think and Spark are still<00:14:44.320> gated<00:14:44.680> behind<00:14:45.040> ultra<00:14:45.320> subscriptions 00:14:45.990 --> 00:14:46.000 align:start position:0% still gated behind ultra subscriptions 00:14:46.000 --> 00:14:48.030 align:start position:0% still gated behind ultra subscriptions or<00:14:46.120> limited<00:14:46.480> testing,<00:14:47.240> while<00:14:47.400> some<00:14:47.640> rivals 00:14:48.030 --> 00:14:48.040 align:start position:0% or limited testing, while some rivals 00:14:48.040 --> 00:14:49.670 align:start position:0% or limited testing, while some rivals ship<00:14:48.280> new<00:14:48.440> capabilities<00:14:49.080> to<00:14:49.200> everyone<00:14:49.560> at 00:14:49.670 --> 00:14:49.680 align:start position:0% ship new capabilities to everyone at 00:14:49.680 --> 00:14:52.190 align:start position:0% ship new capabilities to everyone at once.<00:14:50.400> And<00:14:50.520> the<00:14:50.600> tier<00:14:50.800> structure<00:14:51.200> itself, 00:14:52.190 --> 00:14:52.200 align:start position:0% once. And the tier structure itself, 00:14:52.200 --> 00:14:56.550 align:start position:0% once. And the tier structure itself, free,<00:14:53.000> pro,<00:14:53.960> ultra,<00:14:54.880> API<00:14:55.400> pricing,<00:14:56.280> can<00:14:56.440> be 00:14:56.550 --> 00:14:56.560 align:start position:0% free, pro, ultra, API pricing, can be 00:14:56.560 --> 00:14:58.470 align:start position:0% free, pro, ultra, API pricing, can be genuinely<00:14:57.040> confusing<00:14:57.600> next<00:14:57.839> to<00:14:57.960> a<00:14:58.040> simpler 00:14:58.470 --> 00:14:58.480 align:start position:0% genuinely confusing next to a simpler 00:14:58.480 --> 00:15:00.590 align:start position:0% genuinely confusing next to a simpler flat<00:14:58.839> subscription<00:14:59.440> from<00:14:59.600> a<00:14:59.640> competitor.<00:15:00.480> If 00:15:00.590 --> 00:15:00.600 align:start position:0% flat subscription from a competitor. If 00:15:00.600 --> 00:15:02.110 align:start position:0% flat subscription from a competitor. If you've<00:15:00.720> ever<00:15:00.960> opened<00:15:01.240> the<00:15:01.320> Gemini<00:15:01.680> pricing 00:15:02.110 --> 00:15:02.120 align:start position:0% you've ever opened the Gemini pricing 00:15:02.120 --> 00:15:04.230 align:start position:0% you've ever opened the Gemini pricing page<00:15:02.440> and<00:15:02.560> closed<00:15:02.920> it<00:15:03.040> 5<00:15:03.320> minutes<00:15:03.560> later<00:15:03.880> still 00:15:04.230 --> 00:15:04.240 align:start position:0% page and closed it 5 minutes later still 00:15:04.240 --> 00:15:06.230 align:start position:0% page and closed it 5 minutes later still unsure<00:15:04.640> which<00:15:04.839> plan<00:15:05.160> you<00:15:05.240> need,<00:15:05.839> that's<00:15:06.080> not 00:15:06.230 --> 00:15:06.240 align:start position:0% unsure which plan you need, that's not 00:15:06.240 --> 00:15:08.510 align:start position:0% unsure which plan you need, that's not just<00:15:06.480> you.<00:15:07.040> Where<00:15:07.200> this<00:15:07.360> is<00:15:07.480> actually<00:15:07.839> headed, 00:15:08.510 --> 00:15:08.520 align:start position:0% just you. Where this is actually headed, 00:15:08.520 --> 00:15:10.270 align:start position:0% just you. Where this is actually headed, a<00:15:08.600> few<00:15:08.800> things<00:15:09.000> are<00:15:09.080> confirmed<00:15:09.839> and<00:15:09.920> a<00:15:10.000> few<00:15:10.200> are 00:15:10.270 --> 00:15:10.280 align:start position:0% a few things are confirmed and a few are 00:15:10.280 --> 00:15:11.990 align:start position:0% a few things are confirmed and a few are still<00:15:10.600> rumor,<00:15:11.280> and<00:15:11.400> it's<00:15:11.520> worth<00:15:11.720> keeping 00:15:11.990 --> 00:15:12.000 align:start position:0% still rumor, and it's worth keeping 00:15:12.000 --> 00:15:14.910 align:start position:0% still rumor, and it's worth keeping those<00:15:12.200> separate.<00:15:12.839> Confirmed,<00:15:13.720> Gemini<00:15:14.120> 3.5 00:15:14.910 --> 00:15:14.920 align:start position:0% those separate. Confirmed, Gemini 3.5 00:15:14.920 --> 00:15:16.750 align:start position:0% those separate. Confirmed, Gemini 3.5 Pro<00:15:15.240> is<00:15:15.400> currently<00:15:15.800> in<00:15:15.920> partner<00:15:16.240> testing<00:15:16.640> with 00:15:16.750 --> 00:15:16.760 align:start position:0% Pro is currently in partner testing with 00:15:16.760 --> 00:15:18.829 align:start position:0% Pro is currently in partner testing with a<00:15:16.839> public<00:15:17.200> release<00:15:17.520> expected<00:15:18.000> soon,<00:15:18.680> and 00:15:18.829 --> 00:15:18.839 align:start position:0% a public release expected soon, and 00:15:18.839 --> 00:15:20.230 align:start position:0% a public release expected soon, and Google<00:15:19.080> has<00:15:19.240> already<00:15:19.480> started<00:15:19.800> training<00:15:20.120> on 00:15:20.230 --> 00:15:20.240 align:start position:0% Google has already started training on 00:15:20.240 --> 00:15:23.430 align:start position:0% Google has already started training on Gemini<00:15:20.600> 4,<00:15:21.440> according<00:15:21.839> to<00:15:21.920> its<00:15:22.160> own<00:15:22.360> July<00:15:22.720> 2026 00:15:23.430 --> 00:15:23.440 align:start position:0% Gemini 4, according to its own July 2026 00:15:23.440 --> 00:15:24.990 align:start position:0% Gemini 4, according to its own July 2026 announcement,<00:15:24.200> though<00:15:24.320> there's<00:15:24.520> no<00:15:24.680> public 00:15:24.990 --> 00:15:25.000 align:start position:0% announcement, though there's no public 00:15:25.000 --> 00:15:27.150 align:start position:0% announcement, though there's no public timeline<00:15:25.440> for<00:15:25.560> that<00:15:25.800> yet.<00:15:26.280> Workspace<00:15:26.880> AI 00:15:27.150 --> 00:15:27.160 align:start position:0% timeline for that yet. Workspace AI 00:15:27.160 --> 00:15:29.590 align:start position:0% timeline for that yet. Workspace AI rollout<00:15:27.600> continues<00:15:28.080> expanding,<00:15:29.080> and<00:15:29.240> Gemini 00:15:29.590 --> 00:15:29.600 align:start position:0% rollout continues expanding, and Gemini 00:15:29.600 --> 00:15:31.270 align:start position:0% rollout continues expanding, and Gemini Live's<00:15:29.960> regional<00:15:30.320> language<00:15:30.680> support<00:15:31.040> keeps 00:15:31.270 --> 00:15:31.280 align:start position:0% Live's regional language support keeps 00:15:31.280 --> 00:15:33.910 align:start position:0% Live's regional language support keeps growing.<00:15:31.960> Speculative<00:15:33.080> and<00:15:33.240> worth<00:15:33.480> labeling 00:15:33.910 --> 00:15:33.920 align:start position:0% growing. Speculative and worth labeling 00:15:33.920 --> 00:15:35.710 align:start position:0% growing. Speculative and worth labeling clearly<00:15:34.320> as<00:15:34.480> such,<00:15:35.120> there's<00:15:35.320> talk<00:15:35.560> of<00:15:35.680> a 00:15:35.710 --> 00:15:35.720 align:start position:0% clearly as such, there's talk of a 00:15:35.720 --> 00:15:38.070 align:start position:0% clearly as such, there's talk of a dedicated<00:15:36.360> on-device<00:15:37.000> AI<00:15:37.280> chip<00:15:37.600> for<00:15:37.720> future 00:15:38.070 --> 00:15:38.080 align:start position:0% dedicated on-device AI chip for future 00:15:38.080 --> 00:15:40.030 align:start position:0% dedicated on-device AI chip for future Pixel<00:15:38.400> phones,<00:15:39.120> and<00:15:39.280> some<00:15:39.400> experimental 00:15:40.030 --> 00:15:40.040 align:start position:0% Pixel phones, and some experimental 00:15:40.040 --> 00:15:42.470 align:start position:0% Pixel phones, and some experimental DeepMind<00:15:40.520> research<00:15:40.920> around<00:15:41.240> 3D<00:15:41.720> avatars<00:15:42.320> and 00:15:42.470 --> 00:15:42.480 align:start position:0% DeepMind research around 3D avatars and 00:15:42.480 --> 00:15:44.590 align:start position:0% DeepMind research around 3D avatars and world<00:15:42.760> simulation<00:15:43.760> that<00:15:43.920> hasn't<00:15:44.240> shipped<00:15:44.480> as 00:15:44.590 --> 00:15:44.600 align:start position:0% world simulation that hasn't shipped as 00:15:44.600 --> 00:15:46.430 align:start position:0% world simulation that hasn't shipped as a<00:15:44.640> product.<00:15:45.400> Treat<00:15:45.640> both<00:15:45.840> of<00:15:45.920> those<00:15:46.200> as 00:15:46.430 --> 00:15:46.440 align:start position:0% a product. Treat both of those as 00:15:46.440 --> 00:15:49.270 align:start position:0% a product. Treat both of those as possible,<00:15:47.360> not<00:15:47.680> coming.<00:15:48.600> Nothing<00:15:48.920> official 00:15:49.270 --> 00:15:49.280 align:start position:0% possible, not coming. Nothing official 00:15:49.280 --> 00:15:51.590 align:start position:0% possible, not coming. Nothing official has<00:15:49.440> confirmed<00:15:50.040> either<00:15:50.160> one.<00:15:50.800> The<00:15:50.920> verdict. 00:15:51.590 --> 00:15:51.600 align:start position:0% has confirmed either one. The verdict. 00:15:51.600 --> 00:15:53.630 align:start position:0% has confirmed either one. The verdict. So,<00:15:51.760> where<00:15:51.920> does<00:15:52.040> that<00:15:52.200> leave<00:15:52.440> things?<00:15:53.200> Gemini 00:15:53.630 --> 00:15:53.640 align:start position:0% So, where does that leave things? Gemini 00:15:53.640 --> 00:15:56.150 align:start position:0% So, where does that leave things? Gemini in<00:15:53.760> 2026<00:15:54.680> isn't<00:15:54.960> a<00:15:55.000> chatbot<00:15:55.520> you<00:15:55.640> occasionally 00:15:56.150 --> 00:15:56.160 align:start position:0% in 2026 isn't a chatbot you occasionally 00:15:56.160 --> 00:15:58.430 align:start position:0% in 2026 isn't a chatbot you occasionally open.<00:15:56.880> It's<00:15:57.080> an<00:15:57.280> AI<00:15:57.600> layer<00:15:57.920> Google<00:15:58.280> has 00:15:58.430 --> 00:15:58.440 align:start position:0% open. It's an AI layer Google has 00:15:58.440 --> 00:16:00.830 align:start position:0% open. It's an AI layer Google has threaded<00:15:58.840> through<00:15:59.040> search,<00:15:59.840> Gmail,<00:16:00.680> your 00:16:00.830 --> 00:16:00.840 align:start position:0% threaded through search, Gmail, your 00:16:00.840 --> 00:16:03.390 align:start position:0% threaded through search, Gmail, your documents,<00:16:01.880> and<00:16:02.000> increasingly<00:16:02.960> your<00:16:03.120> phone 00:16:03.390 --> 00:16:03.400 align:start position:0% documents, and increasingly your phone 00:16:03.400 --> 00:16:05.870 align:start position:0% documents, and increasingly your phone itself.<00:16:04.160> The<00:16:04.280> models<00:16:04.640> handle<00:16:04.920> the<00:16:05.040> thinking, 00:16:05.870 --> 00:16:05.880 align:start position:0% itself. The models handle the thinking, 00:16:05.880 --> 00:16:08.190 align:start position:0% itself. The models handle the thinking, the<00:16:06.000> modes<00:16:06.320> handle<00:16:06.640> how<00:16:06.760> you<00:16:06.960> ask,<00:16:07.640> and<00:16:07.840> agents 00:16:08.190 --> 00:16:08.200 align:start position:0% the modes handle how you ask, and agents 00:16:08.200 --> 00:16:10.030 align:start position:0% the modes handle how you ask, and agents like<00:16:08.400> Deep<00:16:08.640> Research<00:16:09.080> are<00:16:09.200> the<00:16:09.280> clearest<00:16:09.720> sign 00:16:10.030 --> 00:16:10.040 align:start position:0% like Deep Research are the clearest sign 00:16:10.040 --> 00:16:11.950 align:start position:0% like Deep Research are the clearest sign of<00:16:10.160> where<00:16:10.360> all<00:16:10.560> of<00:16:10.680> it<00:16:10.760> is<00:16:10.880> actually<00:16:11.320> heading. 00:16:11.950 --> 00:16:11.960 align:start position:0% of where all of it is actually heading. 00:16:11.960 --> 00:16:13.390 align:start position:0% of where all of it is actually heading. If<00:16:12.080> there's<00:16:12.280> one<00:16:12.480> thing<00:16:12.720> worth<00:16:12.920> trying<00:16:13.200> this 00:16:13.390 --> 00:16:13.400 align:start position:0% If there's one thing worth trying this 00:16:13.400 --> 00:16:15.510 align:start position:0% If there's one thing worth trying this week,<00:16:14.080> it's<00:16:14.320> Deep<00:16:14.600> Research<00:16:15.040> on<00:16:15.160> something 00:16:15.510 --> 00:16:15.520 align:start position:0% week, it's Deep Research on something 00:16:15.520 --> 00:16:17.030 align:start position:0% week, it's Deep Research on something you'd<00:16:15.680> normally<00:16:16.080> spend<00:16:16.360> an<00:16:16.480> evening<00:16:16.760> looking 00:16:17.030 --> 00:16:17.040 align:start position:0% you'd normally spend an evening looking 00:16:17.040 --> 00:16:19.150 align:start position:0% you'd normally spend an evening looking into<00:16:17.280> yourself,<00:16:18.120> and<00:16:18.280> actually<00:16:18.640> watching<00:16:19.040> it 00:16:19.150 --> 00:16:19.160 align:start position:0% into yourself, and actually watching it 00:16:19.160 --> 00:16:20.750 align:start position:0% into yourself, and actually watching it work<00:16:19.480> instead<00:16:19.760> of<00:16:19.840> just<00:16:20.080> reading<00:16:20.360> the<00:16:20.440> final 00:16:20.750 --> 00:16:20.760 align:start position:0% work instead of just reading the final 00:16:20.760 --> 00:16:22.710 align:start position:0% work instead of just reading the final report.<00:16:21.480> Drop<00:16:21.720> a<00:16:21.760> comment<00:16:22.160> with<00:16:22.320> which<00:16:22.520> piece 00:16:22.710 --> 00:16:22.720 align:start position:0% report. Drop a comment with which piece 00:16:22.720 --> 00:16:24.630 align:start position:0% report. Drop a comment with which piece of<00:16:22.800> this<00:16:22.960> surprised<00:16:23.440> you<00:16:23.520> most,<00:16:24.280> the<00:16:24.360> model 00:16:24.630 --> 00:16:24.640 align:start position:0% of this surprised you most, the model 00:16:24.640 --> 00:16:27.310 align:start position:0% of this surprised you most, the model lineup,<00:16:25.360> the<00:16:25.520> agent<00:16:25.839> side,<00:16:26.680> or<00:16:26.800> just<00:16:27.040> how<00:16:27.120> much 00:16:27.310 --> 00:16:27.320 align:start position:0% lineup, the agent side, or just how much 00:16:27.320 --> 00:16:28.750 align:start position:0% lineup, the agent side, or just how much of<00:16:27.400> this<00:16:27.600> you<00:16:27.680> were<00:16:27.800> already<00:16:28.160> using<00:16:28.480> without 00:16:28.750 --> 00:16:28.760 align:start position:0% of this you were already using without 00:16:28.760 --> 00:16:30.470 align:start position:0% of this you were already using without realizing<00:16:29.240> it.<00:16:29.600> I'll<00:16:29.720> be<00:16:29.839> back<00:16:30.080> soon<00:16:30.320> with<00:16:30.440> a 00:16:30.470 --> 00:16:30.480 align:start position:0% realizing it. I'll be back soon with a 00:16:30.480 --> 00:16:32.390 align:start position:0% realizing it. I'll be back soon with a deeper<00:16:30.800> breakdown<00:16:31.320> on<00:16:31.440> how<00:16:31.600> Deep<00:16:31.920> Research 00:16:32.390 --> 00:16:32.400 align:start position:0% deeper breakdown on how Deep Research 00:16:32.400 --> 00:16:33.990 align:start position:0% deeper breakdown on how Deep Research actually<00:16:32.800> performs<00:16:33.320> against<00:16:33.640> a<00:16:33.680> real 00:16:33.990 --> 00:16:34.000 align:start position:0% actually performs against a real 00:16:34.000 --> 00:16:36.070 align:start position:0% actually performs against a real research<00:16:34.400> task.<00:16:35.120> Thanks<00:16:35.320> for<00:16:35.440> watching,<00:16:35.960> and 00:16:36.070 --> 00:16:36.080 align:start position:0% research task. Thanks for watching, and 00:16:36.080 --> 00:16:39.200 align:start position:0% research task. Thanks for watching, and I'll<00:16:36.200> see<00:16:36.360> you<00:16:36.480> in<00:16:36.560> the<00:16:36.640> next<00:16:36.880> one.