1 00:00:00,020 --> 00:00:04,420 Josh: If I had a crown in my hands, I would place it on the heads of Google because they have done it again. 2 00:00:04,740 --> 00:00:09,820 Josh: They have the world's best AI model ever in history by a shockingly large margin. 3 00:00:09,960 --> 00:00:11,580 Josh: Gemini 3.0 just got released. 4 00:00:11,800 --> 00:00:15,420 Josh: It's available now to anybody in the world to go use it. And the benchmarks 5 00:00:15,420 --> 00:00:18,480 Josh: are kind of blowing everyone's expectations out of the water, myself included. 6 00:00:18,720 --> 00:00:22,940 Josh: And most importantly, it places another data point on the chart that shows we 7 00:00:22,940 --> 00:00:26,460 Josh: are continuing to ascend up this exponential curve towards AGI. 8 00:00:26,740 --> 00:00:29,860 Josh: And the roadmap is still intact and we are very quickly moving through it. 9 00:00:30,000 --> 00:00:32,920 Josh: This, EJs, I was just, I was going through the benchmarks before recording this 10 00:00:32,920 --> 00:00:37,860 Josh: and I, it's, it's shocking because we live in this world and yet somehow I'm 11 00:00:37,860 --> 00:00:41,760 Josh: still continually blown away by the progress that's made by these models. So let's get into it. 12 00:00:41,940 --> 00:00:45,520 Josh: Please walk everyone through, tell me, what did Gemini and the Google team just 13 00:00:45,520 --> 00:00:47,060 Josh: release with this 3.0 update? 14 00:00:47,580 --> 00:00:51,180 Ejaaz: People probably think we say the world's best model every week, 15 00:00:51,320 --> 00:00:53,320 Ejaaz: but this time we really, really mean it. 16 00:00:53,440 --> 00:00:57,700 Ejaaz: Like they have blown every single other model provider out the water. 17 00:00:58,220 --> 00:01:01,640 Ejaaz: The things that this thing can do. Well, how about how about I just show you? 18 00:01:01,840 --> 00:01:02,780 Ejaaz: How about I show you, Josh? 19 00:01:03,020 --> 00:01:04,060 Josh: Please, let's see some examples. 20 00:01:04,400 --> 00:01:09,220 Ejaaz: We have a thread here. And Sundar basically says, you can give Gemini 3 anything, 21 00:01:09,440 --> 00:01:13,100 Ejaaz: images, PDFs, scribbles on a napkin, and it'll create whatever you like. 22 00:01:13,180 --> 00:01:15,260 Ejaaz: For example, an image becomes a board game. 23 00:01:15,400 --> 00:01:17,980 Ejaaz: A napkin sketch transforms into a full website. 24 00:01:18,280 --> 00:01:22,820 Ejaaz: And a diagram could turn into an interactive lesson, right? So there's two examples 25 00:01:22,820 --> 00:01:25,080 Ejaaz: I want to show you, Josh. I want to get your opinion on this. 26 00:01:25,320 --> 00:01:30,820 Ejaaz: So number one, there's a short video of someone playing pickleball here and 27 00:01:30,820 --> 00:01:36,520 Ejaaz: he or she rather uploads it into Gemini and says, hey, can you tell me how well 28 00:01:36,520 --> 00:01:38,200 Ejaaz: I've done here and how I can improve my game? 29 00:01:38,480 --> 00:01:40,620 Ejaaz: And it analyzes the entire video. 30 00:01:40,960 --> 00:01:44,560 Ejaaz: It knows that she's wearing a knee brace. It analyzes her positions, 31 00:01:44,720 --> 00:01:47,700 Ejaaz: telling her where she can move to better position herself to score the point. 32 00:01:48,040 --> 00:01:50,520 Ejaaz: That's pretty nuts. But before I get your reaction to that, because Josh, 33 00:01:50,640 --> 00:01:53,160 Ejaaz: I know you're an athlete. I know you're very competitive when it comes to these 34 00:01:53,160 --> 00:01:54,640 Ejaaz: things. So this is a tool you could definitely use. 35 00:01:54,800 --> 00:01:58,900 Ejaaz: The second thing is probably applicable to a lot of listeners on this show. 36 00:01:59,600 --> 00:02:04,540 Ejaaz: They've embedded Gemini 3 into Google search and into new generative UI experiences. 37 00:02:05,200 --> 00:02:09,880 Ejaaz: The way I would summarize this is it basically is very intuitive, Josh. 38 00:02:10,200 --> 00:02:14,860 Ejaaz: It understands what you're asking for without you needing to really kind of explain yourself. 39 00:02:14,980 --> 00:02:19,960 Ejaaz: The example they're showing on the video here is, can you explain the three body problem to me? 40 00:02:20,120 --> 00:02:22,820 Ejaaz: And rather than just give you kind of like this simplistic text, 41 00:02:22,980 --> 00:02:27,320 Ejaaz: which explains the concept, it decides to create a video diagram from scratch 42 00:02:27,320 --> 00:02:30,040 Ejaaz: to show you a visual depiction of how this works. 43 00:02:30,540 --> 00:02:34,700 Ejaaz: Right. Give me your reaction in order from one to two. So starting with the sports. 44 00:02:35,200 --> 00:02:36,060 Josh: Okay, let's go. The first example. 45 00:02:36,480 --> 00:02:36,920 Ejaaz: Yes, sir. 46 00:02:37,080 --> 00:02:40,180 Josh: So this is really cool, the napkin example, where you can scribble something 47 00:02:40,180 --> 00:02:42,960 Josh: down on a piece of paper, it'll generate it in the real world. 48 00:02:43,100 --> 00:02:46,760 Josh: What all these examples are kind of showing me is what we always talk about 49 00:02:46,760 --> 00:02:50,920 Josh: with Google, where it has this awareness of physics, reality, 50 00:02:51,080 --> 00:02:53,160 Josh: and visuals and understanding what it's seeing. 51 00:02:53,300 --> 00:02:56,040 Josh: And all three of these examples are leaning into that. So it leads me to believe 52 00:02:56,040 --> 00:03:00,840 Josh: Gemini really is a multimodal first model, where it's meant to ingest, 53 00:03:00,980 --> 00:03:04,820 Josh: meant to understand the world around us. This example of the chessboard and 54 00:03:04,820 --> 00:03:07,840 Josh: the napkin is amazing because a lot of people oftentimes have sketches. 55 00:03:08,040 --> 00:03:10,700 Josh: You just draw it down on paper and it intuitively understands it. 56 00:03:10,800 --> 00:03:13,680 Josh: But the one that was most surprising to me is the video example. 57 00:03:13,820 --> 00:03:15,440 Josh: Because as far as I'm concerned. 58 00:03:16,150 --> 00:03:20,250 Josh: As far as I'm aware, there has never been a model that can ingest video and 59 00:03:20,250 --> 00:03:23,650 Josh: understand the video that it sees. And if it does exist, I've never tried it before. 60 00:03:23,990 --> 00:03:26,810 Josh: So the idea that you can, I mean, I play baseball growing up. 61 00:03:26,910 --> 00:03:30,810 Josh: I, if I could take a video of myself swinging and get a corrective coach to 62 00:03:30,810 --> 00:03:32,190 Josh: walk me through exactly what was wrong. 63 00:03:32,330 --> 00:03:34,450 Josh: A lot of people play golf. I'm sure who are watching this. 64 00:03:34,610 --> 00:03:38,090 Josh: If you could have a phone recording of you playing golf and it can actually 65 00:03:38,090 --> 00:03:40,710 Josh: critique it and then critique me as if you are a tiger woods, 66 00:03:40,830 --> 00:03:43,170 Josh: critique me as if you are whoever else is good that plays golf. 67 00:03:43,270 --> 00:03:45,690 Josh: I don't know, Rory McIlroy, whatever they are, but like critique Critique me 68 00:03:45,690 --> 00:03:49,970 Josh: as if you are an expert who is really good at golf and can give me some feedback 69 00:03:49,970 --> 00:03:50,990 Josh: on how I can better my swing. 70 00:03:51,210 --> 00:03:54,770 Josh: And what this offers in this one just narrow case example is now you have this 71 00:03:54,770 --> 00:03:58,190 Josh: personal tutor that can do anything. If you're dancing, if you're doing anything 72 00:03:58,190 --> 00:04:01,110 Josh: physical, if you're whatever it is, it can evaluate things for you. 73 00:04:01,210 --> 00:04:04,870 Josh: Even if you have a video, a podcast, EJS, we uploaded to Gemini 3.0, 74 00:04:05,010 --> 00:04:07,510 Josh: it could critique us. What did we do well? What did we not? 75 00:04:07,730 --> 00:04:10,810 Josh: What did the visuals look like? How can we improve them? And that awareness 76 00:04:10,810 --> 00:04:12,650 Josh: of video is like really cool. 77 00:04:13,190 --> 00:04:18,390 Ejaaz: Yeah, I just want to say, I think the closest we got to this was with GPT, 78 00:04:18,570 --> 00:04:22,750 Ejaaz: where you can upload an image, like what's under my car bonnet and say, 79 00:04:22,830 --> 00:04:24,590 Ejaaz: hey, what's wrong? My car stopped working. 80 00:04:24,750 --> 00:04:28,350 Ejaaz: And it can kind of like identify the point that you need to kind of like change, 81 00:04:28,590 --> 00:04:29,510 Ejaaz: change the oil, blah, blah, blah. 82 00:04:29,710 --> 00:04:35,430 Ejaaz: But that's just a static image. To go from that to live video and for it to 83 00:04:35,430 --> 00:04:38,950 Ejaaz: analyze all the frames in that video and then give you a response on that is 84 00:04:38,950 --> 00:04:42,370 Ejaaz: a massive leap upwards. We just haven't seen that anyway. So yeah, you're right. 85 00:04:42,850 --> 00:04:45,910 Josh: It's amazing. And every example we go through, it kind of breaks the mold of 86 00:04:45,910 --> 00:04:47,070 Josh: what I believe should be possible. 87 00:04:47,210 --> 00:04:50,670 Josh: And I find that it's going to be difficult to use Gemini 3.0 because there are 88 00:04:50,670 --> 00:04:54,290 Josh: so many possibilities now that have not existed previously. You kind of need 89 00:04:54,290 --> 00:04:57,730 Josh: to relearn how to engage with AI because it's so capable. 90 00:04:57,930 --> 00:05:00,070 Josh: And there's a fourth example here that I just want to touch on briefly, 91 00:05:00,130 --> 00:05:03,250 Josh: which was also cool, is that it works just as well for the other things. 92 00:05:03,370 --> 00:05:07,150 Josh: The example is a trip planning one where it starts to plan a trip and a vacation. 93 00:05:07,150 --> 00:05:12,170 Josh: And it shows you a full list that is fully interactive of all the places broken up day by day. 94 00:05:12,430 --> 00:05:15,030 Josh: And there's an option that you could just choose visual layout. 95 00:05:15,170 --> 00:05:18,510 Josh: And you see on the screen here, it'll take every single day of your trip, 96 00:05:18,650 --> 00:05:21,970 Josh: break it into images, section it out into this really nice visual grid. 97 00:05:22,150 --> 00:05:25,550 Josh: So what I'm seeing the themes here are, okay, real world understanding, 98 00:05:25,870 --> 00:05:30,890 Josh: video first, and really nice presentation, which I think a lot of models sometimes struggle on. 99 00:05:31,450 --> 00:05:35,030 Josh: Demo's out of control. I'm excited to use it. Everyone else can use it now. 100 00:05:35,190 --> 00:05:39,150 Josh: It's live. Now I want to get to benchmarks, EJS, because this is where things 101 00:05:39,150 --> 00:05:43,310 Josh: get kind of crazy, where we could actually compare one to another and see exactly 102 00:05:43,310 --> 00:05:47,030 Josh: how impressive this is relative to everybody else. So please, we have the card here. 103 00:05:47,310 --> 00:05:50,530 Josh: Walk us through what we're seeing in this model card and all the specs that we need to know. 104 00:05:50,750 --> 00:05:54,650 Ejaaz: As you guys probably know by now, benchmarks is typically how we evaluate any 105 00:05:54,650 --> 00:05:57,010 Ejaaz: typical AI model against each other. 106 00:05:57,150 --> 00:05:59,190 Ejaaz: And they're measured against a range of different benchmarks. 107 00:05:59,330 --> 00:06:01,330 Ejaaz: A benchmark can be considered as sort of like a test. 108 00:06:01,650 --> 00:06:06,290 Ejaaz: Now, right at the top, you've got humanity's last exam. This is by default, 109 00:06:06,630 --> 00:06:09,610 Ejaaz: the hardest exam that an AI model is tested against. 110 00:06:10,090 --> 00:06:14,510 Ejaaz: And it's kind of like an academic reasoning test with no tools accessible to it. 111 00:06:14,590 --> 00:06:21,910 Ejaaz: It scored a very impressive 37.5%, which is more than I think is about a 15% 112 00:06:21,910 --> 00:06:23,630 Ejaaz: increase from its previous model. 113 00:06:23,870 --> 00:06:28,610 Ejaaz: Very, very impressive. But what really blew my mind was the second stat listed 114 00:06:28,610 --> 00:06:31,170 Ejaaz: here, which is the ARK AGI 2 benchmark. 115 00:06:31,930 --> 00:06:38,510 Ejaaz: Josh, when I say this 2x'd the previous state-of-the-art model, I absolutely mean it. 116 00:06:38,590 --> 00:06:41,590 Ejaaz: In fact, let me just show you this chart here. 117 00:06:41,950 --> 00:06:45,170 Ejaaz: Now, you may notice a couple of bustly specs here. 118 00:06:45,550 --> 00:06:51,910 Ejaaz: GPT-5 Pro, Grok 4 Thinking. And then can you see that outlier right at the top 119 00:06:51,910 --> 00:06:52,890 Ejaaz: right. Do you see that, Josh? 120 00:06:53,050 --> 00:06:55,190 Josh: That's insane. The two outliers. 121 00:06:55,870 --> 00:07:02,110 Ejaaz: The two outliers. So these are the Gemini 3 Pro and the Gemini 3 deep thinking model. 122 00:07:02,250 --> 00:07:04,630 Ejaaz: Deep thinking being like, you know, a large number that can like kind of give 123 00:07:04,630 --> 00:07:07,330 Ejaaz: you a more research response. They are a... 124 00:07:07,770 --> 00:07:11,850 Ejaaz: Stand out from every single other model. And the reason why this is so crazy, 125 00:07:12,050 --> 00:07:13,270 Ejaaz: well, there's a few reasons. 126 00:07:13,630 --> 00:07:18,350 Ejaaz: Number one, all the other model progressions, as you can see over time, 127 00:07:18,830 --> 00:07:23,590 Ejaaz: has been kind of impressive, but kind of small. 128 00:07:23,970 --> 00:07:28,130 Ejaaz: Like they've been a good jump. It's been impressive, but it hasn't been as impressive 129 00:07:28,130 --> 00:07:30,790 Ejaaz: to be like, oh, you know, another model provider couldn't catch up. 130 00:07:31,210 --> 00:07:36,410 Ejaaz: These results from Google literally put it miles ahead of every other model. 131 00:07:36,590 --> 00:07:40,770 Ejaaz: So when I sat at this chart, I think, wow, Google probably has the lead for 132 00:07:40,770 --> 00:07:45,930 Ejaaz: another six months and in six months time, they're going to have a more impressive model by then. 133 00:07:46,090 --> 00:07:49,650 Ejaaz: So at this point, I'm kind of thinking, can anyone catch up to Google? 134 00:07:49,770 --> 00:07:51,670 Ejaaz: Josh, do you have any reactions to this benchmark? 135 00:07:51,850 --> 00:07:55,270 Josh: This is the chart that like the first thing I said to myself when I saw this 136 00:07:55,270 --> 00:07:57,450 Josh: is like, oh, my God, there is no wall there. 137 00:07:57,610 --> 00:08:01,350 Josh: We are not going to stop scaling. The scaling will still apply because these 138 00:08:01,350 --> 00:08:03,830 Josh: two new data points that we have blow everything else out of the water. 139 00:08:03,970 --> 00:08:05,830 Josh: And this is how exponential growth happens. 140 00:08:05,990 --> 00:08:08,570 Josh: It seems like a really small cluster down there in the bottom, 141 00:08:08,690 --> 00:08:12,530 Josh: but the reality is that was the top just a couple hours ago. 142 00:08:12,690 --> 00:08:17,390 Josh: And Gemini kind of refactored this entire chart to make it seem like it's so 143 00:08:17,390 --> 00:08:18,770 Josh: small because the progress is so high. 144 00:08:18,910 --> 00:08:23,310 Josh: And although Gemini 3.0 thinking is seemingly the most impressive, 145 00:08:23,530 --> 00:08:27,870 Josh: the really anomaly chart is the Gemini 3.0 Pro, which is basically a vertical 146 00:08:27,870 --> 00:08:32,010 Josh: line up from these other models, where the score is higher, but the cost is 147 00:08:32,010 --> 00:08:33,030 Josh: actually slightly lower. 148 00:08:33,230 --> 00:08:37,490 Josh: And if you connect a dot between these averages, you start to see literal vertical 149 00:08:37,490 --> 00:08:40,410 Josh: line in terms of improvement and acceleration in these models. 150 00:08:40,610 --> 00:08:44,970 Josh: And that to me shows that there is no scaling wall that we're hitting. 151 00:08:45,270 --> 00:08:48,810 Josh: Like we can continue to scale resources, energy, compute, and we could continue 152 00:08:48,810 --> 00:08:51,730 Josh: along this path towards AGI in a world where some people were saying, 153 00:08:51,870 --> 00:08:55,630 Josh: we don't know if it continues. The answer to me is very clearly, it continues. 154 00:08:55,870 --> 00:08:57,890 Josh: This is a step much closer to AGI. 155 00:08:58,150 --> 00:09:03,850 Josh: And again, that real world understanding makes it feel much closer to AGI than it ever has before. 156 00:09:04,550 --> 00:09:08,470 Josh: Because now it really like intuitively understands the world through video, 157 00:09:08,790 --> 00:09:11,950 Josh: through photo, through audio, through basically every sensory input we have 158 00:09:11,950 --> 00:09:13,290 Josh: outside of what taste and feel. 159 00:09:13,930 --> 00:09:17,270 Josh: So this to me, I saw this chart. I was like, oh, my God, Gemini, 160 00:09:17,490 --> 00:09:18,710 Josh: you really outdid yourselves. 161 00:09:18,910 --> 00:09:22,630 Ejaaz: I'm just going to be honest. I think over the last couple of months, 162 00:09:22,630 --> 00:09:27,810 Ejaaz: I've been getting a little bored with the models that have been released by other model providers. 163 00:09:27,810 --> 00:09:32,790 Ejaaz: And it led me to think that we're not going to make many breakthroughs until 164 00:09:32,790 --> 00:09:37,090 Ejaaz: some model provider figures out a new, unique way to train their model. 165 00:09:37,890 --> 00:09:41,090 Ejaaz: Gemini or Google has convinced me otherwise with this release. 166 00:09:41,230 --> 00:09:45,530 Ejaaz: But I know you guys are probably like fed up with listening to us hop on about benchmarks. 167 00:09:45,570 --> 00:09:50,150 Ejaaz: So how about I materialize that for you in a much more easy to understand way, right? 168 00:09:50,410 --> 00:09:55,390 Ejaaz: So here are the four big takeaways that you need to learn about Gemini 3. 169 00:09:55,930 --> 00:10:00,110 Ejaaz: Number one, for its intelligence, for the intelligence that you're getting, 170 00:10:00,310 --> 00:10:02,530 Ejaaz: it is not that super expensive. 171 00:10:03,110 --> 00:10:07,710 Ejaaz: Google trained this from scratch, as this tweet says, using their own GPU infrastructure. 172 00:10:07,910 --> 00:10:11,250 Ejaaz: And it used this kind of like layout called a mixture of experts, 173 00:10:11,530 --> 00:10:15,930 Ejaaz: which basically means that whenever you prompt the model, it's not going to use the entire model. 174 00:10:16,030 --> 00:10:19,070 Ejaaz: So it actually ends up being cheaper than what it could eventually become. 175 00:10:19,310 --> 00:10:24,130 Ejaaz: One million token context input, 64K token output. We'll get to the costs in 176 00:10:24,130 --> 00:10:27,250 Ejaaz: a bit of a second, But the point that I'm making here is that it's not as expensive 177 00:10:27,250 --> 00:10:29,670 Ejaaz: as you would expect for the intelligence that you're getting. 178 00:10:29,770 --> 00:10:36,350 Ejaaz: Now, if you compared Gemini 3 to GPT 5.1 from OpenAI, on a relative basis, 179 00:10:36,350 --> 00:10:39,250 Ejaaz: it is more expensive. But for the jump in intelligence that you're getting, 180 00:10:39,490 --> 00:10:42,030 Ejaaz: it's way better. So it's, in my opinion, worth it. 181 00:10:42,430 --> 00:10:47,810 Ejaaz: Number two, when it comes to computer use, so that means letting the AI model 182 00:10:47,810 --> 00:10:51,870 Ejaaz: control your computer and do tasks for you whilst you go do something else. 183 00:10:52,540 --> 00:10:56,200 Ejaaz: It is state of the art. It is the best here. They measured it against a benchmark 184 00:10:56,200 --> 00:11:00,900 Ejaaz: called ScreenSpot Pro, which is a benchmark which kind of like analyzes its 185 00:11:00,900 --> 00:11:06,380 Ejaaz: ability to understand images and visuals on a desktop. It just absolutely crushes it. 186 00:11:06,700 --> 00:11:09,800 Ejaaz: Number three, it is the best AI for math by far. 187 00:11:09,960 --> 00:11:13,680 Ejaaz: So again, the point I'm making here or the theme that we're seeing here is it's 188 00:11:13,680 --> 00:11:17,300 Ejaaz: not just good at one thing, it's good at many things, which makes it the best 189 00:11:17,300 --> 00:11:20,820 Ejaaz: generalist AI model in the world right now, by far. 190 00:11:21,260 --> 00:11:24,780 Ejaaz: And the final thing, Josh, and this is where it might slip up. 191 00:11:24,860 --> 00:11:26,200 Ejaaz: I'm curious to get your take on this. 192 00:11:26,420 --> 00:11:32,680 Ejaaz: It is insanely good at coding, but we don't quite know if it is the best at coding yet. 193 00:11:32,800 --> 00:11:37,100 Ejaaz: What I mean by that is it completely crushed everyone else on one coding benchmark, 194 00:11:37,300 --> 00:11:40,620 Ejaaz: but the coding benchmark that matters, which is the software engineering, 195 00:11:40,800 --> 00:11:45,360 Ejaaz: SWE, it didn't do as well as its competitor, Claude 4.5 from Anthropic. 196 00:11:45,700 --> 00:11:47,620 Ejaaz: So those are the four main takeaways. 197 00:11:47,620 --> 00:11:51,020 Josh: I would much prefer a model that understands the world than understands how to code. 198 00:11:51,220 --> 00:11:55,380 Josh: And I think we're starting to see these subset niches where if Anthropic has 199 00:11:55,380 --> 00:11:56,540 Josh: the best coding model, that's great. 200 00:11:56,840 --> 00:11:59,980 Josh: Let them focus on code. Let them narrowly make that the best model. 201 00:12:00,120 --> 00:12:03,000 Josh: Let Google handle everything else. And I think that's what Gemini is focusing 202 00:12:03,000 --> 00:12:06,880 Josh: on. So the code thing doesn't really bother me because I don't care to use Gemini for code. 203 00:12:07,000 --> 00:12:11,060 Josh: I'm happy to be in Claude Camp for code and then use Gemini for everything else. 204 00:12:11,200 --> 00:12:13,700 Josh: And then one of the points earlier that you mentioned on the pricing is. 205 00:12:14,250 --> 00:12:18,390 Josh: I find it a little interesting because it's a little bit more than just a little bit more expensive. 206 00:12:18,730 --> 00:12:22,630 Josh: The pricing, I was looking through it and it's for over 200,000 tokens. 207 00:12:22,630 --> 00:12:27,270 Josh: They're charging $4 for inputs and $18 for outputs. 208 00:12:27,410 --> 00:12:32,310 Josh: Now, relative to GPT 5.1, which just got released, they're charging for a million 209 00:12:32,310 --> 00:12:35,050 Josh: tokens, $1.25 in, $10 out. 210 00:12:35,230 --> 00:12:41,830 Josh: So you're talking about, what is that? that's about $20 versus $1.25 on inputs. 211 00:12:42,030 --> 00:12:45,410 Josh: And that is a fairly significant margin that you're paying for this quality. 212 00:12:45,550 --> 00:12:48,690 Josh: So we're starting to see the trade-offs happening on that Pareto curve that 213 00:12:48,690 --> 00:12:52,150 Josh: we talked about in a few episodes earlier, where there are trade-offs coming 214 00:12:52,150 --> 00:12:53,450 Josh: in terms of cost and quality. 215 00:12:53,570 --> 00:12:57,030 Josh: And it's clear that while OpenAI may have optimized for cost, 216 00:12:57,550 --> 00:13:02,550 Josh: Google is kind of optimizing for a little further up the cost curve in exchange for super high quality. 217 00:13:02,690 --> 00:13:06,230 Josh: And it seems like this is kind of a balanced data point for now because unless 218 00:13:06,230 --> 00:13:09,510 Josh: you are using this via API and you're requiring a ton of tokens, 219 00:13:09,850 --> 00:13:14,430 Josh: a $20 a month Google membership will get you all of the use that you need. And that is just fine. 220 00:13:14,590 --> 00:13:17,610 Josh: So in terms of a usability perspective, I think that's okay. 221 00:13:18,050 --> 00:13:21,230 Josh: But it's just an interesting thing to know is that this is a better model. 222 00:13:21,390 --> 00:13:22,270 Josh: It is also more expensive. 223 00:13:22,470 --> 00:13:25,690 Josh: And that is a trade-off that was made. And in the case that OpenAI decides to 224 00:13:25,690 --> 00:13:29,750 Josh: make this trade-off with GPT-6 or Grok decides to make this with Grok-5, 225 00:13:30,260 --> 00:13:34,020 Josh: or Grok 6. I'm losing track of all these models now. I think we're going to 226 00:13:34,020 --> 00:13:37,480 Josh: start to see the dynamic shift in terms of that Pareto curve and what model 227 00:13:37,480 --> 00:13:39,700 Josh: architects decide to remove and add. 228 00:13:39,860 --> 00:13:42,260 Josh: And in this case, it looks like Google added quality, but they also did add 229 00:13:42,260 --> 00:13:43,860 Josh: quite a significant cost increase. 230 00:13:44,040 --> 00:13:46,260 Ejaaz: I personally don't think it matters. I think it's a nothing burger. 231 00:13:46,460 --> 00:13:49,980 Ejaaz: I think that if Google wanted to make it affordable for everyone, 232 00:13:50,060 --> 00:13:52,880 Ejaaz: including the developers that want to get API access, that think it might be 233 00:13:52,880 --> 00:13:54,700 Ejaaz: too expensive, they could subsidize it. 234 00:13:54,860 --> 00:13:57,480 Ejaaz: They are a cash flow giant. They have enough money to do that. 235 00:13:57,600 --> 00:14:00,120 Ejaaz: OpenAI has been doing that for so long now that it doesn't even matter. 236 00:14:00,340 --> 00:14:02,540 Ejaaz: I don't see any reason why Google couldn't do that. 237 00:14:02,780 --> 00:14:08,240 Ejaaz: The other reason is Google just released their latest TPU, which is their GPU 238 00:14:08,240 --> 00:14:10,760 Ejaaz: that they use to train their models and inference their models. 239 00:14:11,040 --> 00:14:15,180 Ejaaz: And typically with every generation, we get a much cheaper cost of inference. 240 00:14:15,280 --> 00:14:18,520 Ejaaz: I think by the time that they release their next generation model, 241 00:14:18,700 --> 00:14:24,740 Ejaaz: which might be, you know, Gemini 3.1, we're going to see a considerable reduction 242 00:14:24,740 --> 00:14:29,460 Ejaaz: in the cost for using Gemini 3 Pro and Gemini Pro deep research. 243 00:14:29,680 --> 00:14:32,540 Ejaaz: So I'm not too worried about that. I think it's kind of like a short-term problem 244 00:14:32,540 --> 00:14:33,880 Ejaaz: and not a long-term problem. 245 00:14:34,040 --> 00:14:39,100 Ejaaz: But kind of speaking of TPUs, I just want to take a moment to really kind of 246 00:14:39,100 --> 00:14:44,000 Ejaaz: belay the point that using their own TPUs to train a state-of-the-art model 247 00:14:44,000 --> 00:14:48,860 Ejaaz: that is 2x better than the previous state-of-the-art model and probably puts 248 00:14:48,860 --> 00:14:53,100 Ejaaz: them in a six-month lead after Google started off on the back foot, 249 00:14:53,260 --> 00:14:56,340 Ejaaz: creating probably the worst model I've ever seen and 250 00:14:56,340 --> 00:14:59,680 Ejaaz: changing that all around in what's it under two years is 251 00:14:59,680 --> 00:15:02,880 Ejaaz: nothing short of insanity tpus is 252 00:15:02,880 --> 00:15:05,980 Ejaaz: uh google's kind of version of the gpu gpu is 253 00:15:05,980 --> 00:15:09,800 Ejaaz: kind of like what nvidia controls the monopoly over this is the hardware that 254 00:15:09,800 --> 00:15:13,980 Ejaaz: you use to train your ai and inference your ai the uh unique part here is that 255 00:15:13,980 --> 00:15:19,580 Ejaaz: google's never used an nvidia gpu uh or in any considerable way to train their 256 00:15:19,580 --> 00:15:23,040 Ejaaz: models they've always trained it in-house and that's such a difficult and tricky 257 00:15:23,040 --> 00:15:25,360 Ejaaz: thing to do because designing and building these 258 00:15:25,840 --> 00:15:29,700 Ejaaz: TPUs at scale, these GPUs at scale is a super hard and complex thing. 259 00:15:29,760 --> 00:15:33,260 Ejaaz: You need so much talent, you need so much expertise and insight to be able to do that. 260 00:15:33,480 --> 00:15:36,840 Ejaaz: The unique thing about Google's TPUs, well, there's two main takeaways. 261 00:15:36,980 --> 00:15:41,660 Ejaaz: Number one, it's cheaper to train the same amount of intelligence that an NVIDIA 262 00:15:41,660 --> 00:15:43,280 Ejaaz: GPU is. So it's more cost efficient. 263 00:15:43,440 --> 00:15:47,680 Ejaaz: And the second way is, and this is their secret sauce, you can stack those TPUs 264 00:15:47,680 --> 00:15:53,920 Ejaaz: on top of each other in a really scalable way that you can start training really, really large models. 265 00:15:53,920 --> 00:15:57,120 Ejaaz: If you wanted to train the same size model with nvidia 266 00:15:57,120 --> 00:16:00,180 Ejaaz: gpus it would cost way more and it would take way 267 00:16:00,180 --> 00:16:05,580 Ejaaz: longer so google made a really risky and big bet about a decade ago saying we're 268 00:16:05,580 --> 00:16:07,800 Ejaaz: going to build our infrastructure in-house and we're not going to rely on nvidia 269 00:16:07,800 --> 00:16:11,860 Ejaaz: and we're going to benefit from the full stack experience and this model is 270 00:16:11,860 --> 00:16:15,400 Ejaaz: a prime example of that bet paying off so i just want to call them out like 271 00:16:15,400 --> 00:16:19,760 Ejaaz: it's not like google has gotten lucky here they've been planning it for a while now The 272 00:16:19,760 --> 00:16:24,260 Josh: Interesting thing to me is that this is the first number one model in the world 273 00:16:24,260 --> 00:16:26,880 Josh: built on something other than an NVIDIA GPU. 274 00:16:27,160 --> 00:16:30,980 Josh: And that's fairly significant because every company in the world is trying, 275 00:16:31,140 --> 00:16:32,580 Josh: but this is proof that it's actually possible. 276 00:16:32,980 --> 00:16:36,820 Josh: And I think when we talk about Tesla and AI5 and the XAI team, 277 00:16:37,080 --> 00:16:41,120 Josh: when we talk about OpenAI working with whoever they're working with to build 278 00:16:41,120 --> 00:16:45,220 Josh: their own in-house GPUs, I think this sets a precedent that it is possible. 279 00:16:45,340 --> 00:16:49,060 Josh: And I suspect that will result in more companies putting their foot on the gas. 280 00:16:49,280 --> 00:16:53,380 Josh: When it comes to kind of destructing part of NVIDIA's monopoly that it holds 281 00:16:53,380 --> 00:16:56,100 Josh: over GPUs. So that to me is the interesting takeaway of this. 282 00:16:56,220 --> 00:17:00,300 Josh: And hearing that it was fully done, trained on these TPUs, that's very high 283 00:17:00,300 --> 00:17:03,160 Josh: signal to me saying, okay, there is an architecture chips happening. 284 00:17:03,360 --> 00:17:06,440 Josh: There is a real benefit to vertical integration if you could figure out manufacturing 285 00:17:06,440 --> 00:17:08,440 Josh: these compute units at scale. 286 00:17:08,600 --> 00:17:11,900 Josh: And now the race is on for everyone to do this. 287 00:17:12,020 --> 00:17:14,820 Josh: Because again, using the Apple example, the M series chips, unbelievable, 288 00:17:15,080 --> 00:17:16,720 Josh: and they unlocked the best computers in the world. 289 00:17:16,880 --> 00:17:20,040 Josh: And if companies can really start to refine this vertical integration of their 290 00:17:20,040 --> 00:17:24,200 Josh: own chips, you're going to see that exponential curve go vertical times 10. 291 00:17:24,420 --> 00:17:27,880 Josh: Like it is going to, I suspect that is very... 292 00:17:28,170 --> 00:17:32,170 Josh: Obviously now how we reach AGI faster than people previously thought, 293 00:17:32,210 --> 00:17:34,850 Josh: because the efficiency improvements from those vertical integrations, 294 00:17:35,010 --> 00:17:39,010 Josh: once they're able to manufacture these at scale, are going to be unbelievable. 295 00:17:39,190 --> 00:17:43,390 Josh: And I'm so excited for that to happen in the near future. Google has a big head 296 00:17:43,390 --> 00:17:46,090 Josh: start, but let me tell you, the other companies are not far behind. 297 00:17:46,920 --> 00:17:51,900 Ejaaz: Well, let me introduce you to another big advantage of being the big dog, Google. 298 00:17:52,140 --> 00:17:55,620 Ejaaz: You thought you were going to come on to this episode and listen to us hopping 299 00:17:55,620 --> 00:17:57,360 Ejaaz: on about a generalized model? No. 300 00:17:57,660 --> 00:18:00,900 Ejaaz: You're forgetting Google has many other products in their arsenal, 301 00:18:00,900 --> 00:18:05,260 Ejaaz: and you're forgetting that they can plug in their new state-of-the-art model into all of them. 302 00:18:05,500 --> 00:18:10,280 Ejaaz: So Google not only today announced Gemini 3, but they also announced a different product. 303 00:18:10,420 --> 00:18:15,540 Ejaaz: It's called Google Anti-Gravity, which is basically a new software environment 304 00:18:15,540 --> 00:18:20,180 Ejaaz: for you to code up AI agents, except this time these AI agents are going to 305 00:18:20,180 --> 00:18:24,000 Ejaaz: be super, super smart because they get plugged in with Gemini 3. 306 00:18:24,260 --> 00:18:27,240 Ejaaz: Now, if you remember earlier, I mentioned that one of the cool benchmarks that 307 00:18:27,240 --> 00:18:31,640 Ejaaz: this new model sets is in computer use, which means that it can control your 308 00:18:31,640 --> 00:18:33,980 Ejaaz: computer, it can do things autonomously for you. 309 00:18:34,420 --> 00:18:36,960 Ejaaz: Now, typically, the reason why we haven't really spoken about that on the show 310 00:18:36,960 --> 00:18:38,520 Ejaaz: is that they've been kind of lame. 311 00:18:38,740 --> 00:18:41,980 Ejaaz: Like they can book you a dinner reservation and do different kinds of stuff. 312 00:18:42,180 --> 00:18:46,480 Ejaaz: With this model, it's way more intuitive. It can do way more intelligent tasks 313 00:18:46,480 --> 00:18:50,740 Ejaaz: and it can take a lot more complex work off of your hands such that the value 314 00:18:50,740 --> 00:18:54,380 Ejaaz: that it produces to you over like the eight hours that you take to sleep overnight 315 00:18:54,380 --> 00:18:58,780 Ejaaz: would be considerable for you to actually be serious to use in your enterprise, 316 00:18:58,800 --> 00:19:01,400 Ejaaz: in your business, or just at your at-home lifestyle, right? 317 00:19:01,640 --> 00:19:07,620 Ejaaz: So the point I want to make around here is Google's moat is not just its intelligence 318 00:19:07,620 --> 00:19:10,600 Ejaaz: or ability to create new models. It's not its TPUs. 319 00:19:10,760 --> 00:19:15,680 Ejaaz: It's its distribution. It's the entire product suite that it has that regular 320 00:19:15,680 --> 00:19:19,960 Ejaaz: users like you and I that use Gmail, that use Google Suite can now kind of benefit 321 00:19:19,960 --> 00:19:21,940 Ejaaz: from simply by plugging in that model. 322 00:19:22,160 --> 00:19:25,260 Ejaaz: And I think like products like this, anti-gravity, I bet you, 323 00:19:25,320 --> 00:19:28,920 Ejaaz: Josh, we're going to see a slew of new Google product releases over the next 324 00:19:28,920 --> 00:19:31,700 Ejaaz: couple of weeks simply because they created this model. 325 00:19:32,570 --> 00:19:36,010 Josh: I hope so. I guess the contrarian take is like, okay, how many people are actually 326 00:19:36,010 --> 00:19:36,750 Josh: going to want to use them? 327 00:19:37,870 --> 00:19:42,070 Josh: We just spoke about how Claude is the superior code model. Everyone loves Cursor. 328 00:19:42,550 --> 00:19:46,290 Josh: No one really uses the mobile applications of these. A lot of people are engaging 329 00:19:46,290 --> 00:19:49,270 Josh: with AI on their phone. So maybe it works for the right type of person. 330 00:19:49,470 --> 00:19:54,090 Josh: But Google still does have that product problem where they kind of have a tough time. 331 00:19:54,250 --> 00:19:56,670 Josh: They have the amazing intelligence. They just have a tough time using it. 332 00:19:56,750 --> 00:20:00,210 Josh: I mean, I don't have the Gemini app on my phone. I mostly use Grok and ChatGPT. 333 00:20:00,470 --> 00:20:01,970 Josh: And there is this bar that they 334 00:20:01,970 --> 00:20:04,850 Josh: still need to cross that I think they're trying with Google AI Studio. 335 00:20:05,050 --> 00:20:08,550 Josh: And we had Logan Kilpatrick on, who was the head of that, to talk about it when Nano Banana came out. 336 00:20:08,670 --> 00:20:13,930 Josh: But there is still a bit of a long shot for them to get good at products to actually develop this. 337 00:20:14,030 --> 00:20:17,670 Josh: But what we saw this week is that there was a resounding, overwhelming amount 338 00:20:17,670 --> 00:20:20,350 Josh: of support, to your point, you guys, where the market just believes in Google. 339 00:20:20,570 --> 00:20:23,930 Josh: And in a week where all of the stocks, all of the Mag7 was down, 340 00:20:24,370 --> 00:20:26,490 Josh: Google was the one anomaly. Google was up this week. 341 00:20:26,650 --> 00:20:29,770 Josh: And I think it's because the market is starting to realize, one, 342 00:20:29,930 --> 00:20:32,850 Josh: vertical integration through these TPUs is a huge deal. 343 00:20:33,130 --> 00:20:36,490 Josh: Two, Google has an existing business that is not reliant on AI. 344 00:20:36,890 --> 00:20:41,090 Josh: And sure, AI places a huge hand on that scale, but it is not everything. 345 00:20:41,270 --> 00:20:45,170 Josh: And they are cashflow positive in the absence of AI. So all of this innovation 346 00:20:45,170 --> 00:20:48,550 Josh: that they're doing is really just applying later fluid on top of an already 347 00:20:48,550 --> 00:20:51,130 Josh: great business. And the market's starting to evaluate that properly. 348 00:20:51,290 --> 00:20:53,810 Josh: So Google is positioned very strongly. 349 00:20:53,990 --> 00:20:56,270 Josh: They have very high intelligence. Gemini 3 rocks. 350 00:20:56,770 --> 00:20:59,370 Josh: And I mean, again, we continue on the bull train for Google. 351 00:20:59,610 --> 00:21:01,150 Josh: I am a believer. I am a supporter. 352 00:21:01,370 --> 00:21:04,330 Josh: I am stoked that they have the crown. I assumed it was only a matter of time. 353 00:21:04,410 --> 00:21:06,010 Josh: And now the question is, who's next? 354 00:21:06,310 --> 00:21:10,650 Josh: Who is the next competitor? Who's going to set the next plot on that chart and 355 00:21:10,650 --> 00:21:13,270 Josh: set the vertical trajectory on the exponential curve we're on? 356 00:21:14,000 --> 00:21:15,840 Josh: Do you have any guesses who you think it's going to be? 357 00:21:16,640 --> 00:21:20,700 Ejaaz: Yeah, well, I don't because I don't think it's going to be anyone for a while. 358 00:21:20,880 --> 00:21:23,940 Ejaaz: I said this earlier in the show and I'm going to say it again. I think there's going 359 00:21:23,940 --> 00:21:29,720 Ejaaz: to be a six month period now where either the other model providers don't release 360 00:21:29,720 --> 00:21:33,980 Ejaaz: the model because it's not as good as Google's or they just kind of release 361 00:21:33,980 --> 00:21:38,660 Ejaaz: these kind of mediocre kind of like consumer products that kind of maybe benefit 362 00:21:38,660 --> 00:21:41,620 Ejaaz: certain consumers in one way or another, 363 00:21:41,780 --> 00:21:44,240 Ejaaz: but doesn't really kind of break the generalized model 364 00:21:44,900 --> 00:21:46,820 Ejaaz: standard that Google has just set. 365 00:21:47,220 --> 00:21:51,480 Ejaaz: Just a last point on the kind of Google bull case thesis, they may not just 366 00:21:51,480 --> 00:21:53,260 Ejaaz: play in the same ring as cursor does. 367 00:21:53,700 --> 00:21:57,700 Ejaaz: Like I was critiquing Microsoft on another episode, Josh, do you remember? 368 00:21:57,900 --> 00:22:02,780 Ejaaz: And then I got off that episode and I was just like, Microsoft like dominates 369 00:22:02,780 --> 00:22:03,660 Ejaaz: the enterprise environment. 370 00:22:04,040 --> 00:22:09,280 Ejaaz: All the boomer companies and institutions love Microsoft and they have all their data and memory. 371 00:22:09,500 --> 00:22:11,940 Ejaaz: And just because you and I don't use it or just I'll speak for myself, 372 00:22:12,080 --> 00:22:15,240 Ejaaz: just because I don't use it and I think it's doesn't mean that they're not absolutely crushing. 373 00:22:15,380 --> 00:22:19,260 Ejaaz: Google just came off a hundred billion dollar quarter of revenue. 374 00:22:19,420 --> 00:22:23,840 Ejaaz: That's like the highest they've ever had. So I don't want to be too hasty to 375 00:22:23,840 --> 00:22:26,820 Ejaaz: say that like Google's not going to make it because they can't make a sick consumer 376 00:22:26,820 --> 00:22:28,240 Ejaaz: product like OpenAI can maybe. 377 00:22:28,400 --> 00:22:30,600 Ejaaz: I just think they're maybe playing in different fields. 378 00:22:31,160 --> 00:22:34,100 Ejaaz: But to the point around like I don't think anyone else is going to catch up. 379 00:22:34,240 --> 00:22:38,080 Ejaaz: Look at these comments. I want to show you two comments all right one is from sam altman 380 00:22:38,760 --> 00:22:41,740 Ejaaz: He goes, congrats to Google and Gemini 3. This looks like a great model. 381 00:22:42,080 --> 00:22:47,020 Ejaaz: The other is from the almighty being, Elon Musk, saying, I can't wait to try this out. 382 00:22:47,120 --> 00:22:49,320 Ejaaz: And this is just one of a series of tweets that he's been putting out this week 383 00:22:49,320 --> 00:22:53,060 Ejaaz: saying, can you guys just drop Gemini 3? Because I need to see how good this thing is. 384 00:22:53,180 --> 00:22:57,960 Ejaaz: And the reason why I bring up these two people is both Sam Altman and Elon Musk 385 00:22:57,960 --> 00:23:02,080 Ejaaz: have released new versions of their models, GPT and Grok respectively, 386 00:23:02,340 --> 00:23:04,240 Ejaaz: but it's been the 0.1 upgrade. 387 00:23:04,420 --> 00:23:09,900 Ejaaz: It's GPT 5.1. It is Grok 4.1. and they are almost identical updates. 388 00:23:10,040 --> 00:23:12,600 Ejaaz: You want to know what the biggest and coolest thing about their model updates were? 389 00:23:13,480 --> 00:23:16,620 Ejaaz: Personality traits, which don't get me wrong, is cool. Like I would like my 390 00:23:16,620 --> 00:23:19,180 Ejaaz: model to kind of respond in a very intuitive manner and get me, 391 00:23:19,440 --> 00:23:23,320 Ejaaz: but it's nowhere near the state of the art standard that we've just seen broken by Gemini. 392 00:23:23,400 --> 00:23:28,360 Ejaaz: So the point I'm making is, I think these two companies might have run out of fuel for the near term. 393 00:23:28,820 --> 00:23:32,500 Josh: Grok is going to be next. You think? They're the next one. By the end of Q1, 394 00:23:32,600 --> 00:23:33,340 Josh: Grok will have the crown. 395 00:23:33,500 --> 00:23:38,960 Josh: Why? And I assume by a fairly large margin. But I assume it will be a different type of crown. 396 00:23:39,100 --> 00:23:41,820 Josh: And this is where I'm really excited to see how these models progress. 397 00:23:42,120 --> 00:23:45,340 Josh: We spoke a little bit earlier about how Cursor is kind of the coding model. 398 00:23:46,200 --> 00:23:50,100 Josh: Google has a very deep understanding of the real world and physics and video 399 00:23:50,100 --> 00:23:51,200 Josh: and understanding how that works. 400 00:23:51,880 --> 00:23:55,860 Josh: Grok and the XAI team are very focused on the pursuit of truth and information. 401 00:23:56,300 --> 00:23:59,160 Josh: And I think that's kind of the alley that we'll see them going down. 402 00:23:59,260 --> 00:24:02,980 Josh: So they have the real-time data with X. They have the pursuit of truth. 403 00:24:03,180 --> 00:24:06,700 Josh: And where Google and OpenAI and all these other companies are trained on an 404 00:24:06,700 --> 00:24:10,780 Josh: existing data set, the XAI team and the Grok team are developing an entirely 405 00:24:10,780 --> 00:24:15,140 Josh: new synthetic data set that is maximally truth-seeking. 406 00:24:15,200 --> 00:24:19,020 Josh: And we saw that early version with Grokopedia that should provide the most accurate 407 00:24:19,020 --> 00:24:23,000 Josh: and I guess thoughtful information. It should be the best of thinking because 408 00:24:23,000 --> 00:24:25,700 Josh: it's the closest to source truth. So while I think, 409 00:24:26,410 --> 00:24:30,390 Josh: Gemini will probably be better at physics and video and understanding the real 410 00:24:30,390 --> 00:24:31,490 Josh: world for quite some time. 411 00:24:31,710 --> 00:24:36,450 Josh: I suspect Grok will be really good at just communicating via text. 412 00:24:36,690 --> 00:24:40,230 Josh: If text is a modality in which we interface from, Grok should be really good. 413 00:24:40,350 --> 00:24:42,750 Josh: And again, the rate of acceleration, Grok has been around for the least amount 414 00:24:42,750 --> 00:24:44,350 Josh: of time. They're accelerating the fastest. 415 00:24:44,550 --> 00:24:49,530 Josh: And I'm very, very, very excited for a Grok 6, Grok 5, whatever we're at announcement, 416 00:24:49,710 --> 00:24:51,870 Josh: hopefully early next year. So that's the predictions. 417 00:24:52,150 --> 00:24:56,230 Josh: That's the episode. That's Gemini 3.1. It is an unbelievable new model. 418 00:24:56,390 --> 00:24:58,150 Josh: Everyone could try it out. So here's how you try it out. 419 00:24:58,890 --> 00:25:04,130 Josh: I believe you need to be a Google premium plus subscriber, whatever it's called. It's $20 a month. 420 00:25:04,250 --> 00:25:07,410 Josh: You can go on the Gemini website and it's just a text box and you can play around 421 00:25:07,410 --> 00:25:08,710 Josh: with it. They also have a mobile application. 422 00:25:09,050 --> 00:25:11,630 Josh: It's very easy to download on your phone, play around with it. 423 00:25:12,010 --> 00:25:15,150 Josh: I'd love to see examples of cool things because I think one of the problems 424 00:25:15,150 --> 00:25:17,890 Josh: for me and one of the things I'd love help with from anyone who's listening 425 00:25:17,890 --> 00:25:20,170 Josh: is how do you use this thing to test it? 426 00:25:20,350 --> 00:25:25,210 Josh: What do I ask it? And how are you interfacing with it to get the maximum amount of results from it? 427 00:25:25,450 --> 00:25:28,250 Josh: Because intuitively, I would never think to record myself and ask for feedback, 428 00:25:28,250 --> 00:25:29,690 Josh: but that's a new possibility. 429 00:25:29,890 --> 00:25:33,170 Josh: So I guess the challenge to anyone who's listening is figuring out how to get 430 00:25:33,170 --> 00:25:36,130 Josh: the most out of these new models as these new features get released. 431 00:25:36,510 --> 00:25:40,470 Josh: And Gemini 3 has just opened up the gates to a gazillion new use cases. 432 00:25:41,870 --> 00:25:45,790 Ejaaz: Yeah, I mean, this is a super cool release for Google. 433 00:25:46,050 --> 00:25:49,530 Ejaaz: And weirdly enough, it's not the only release over the last week. 434 00:25:49,670 --> 00:25:53,470 Ejaaz: I mean, I've got a list pulled up here. They've released new Android iOS updates. 435 00:25:53,750 --> 00:25:57,130 Ejaaz: They've got a new search AI mode. They've released anti-gravity that we mentioned earlier. 436 00:25:57,330 --> 00:26:00,690 Ejaaz: We've got CIMA2 research, which we demoed on a previous episode. 437 00:26:00,810 --> 00:26:01,810 Ejaaz: You should definitely go check that out. 438 00:26:02,090 --> 00:26:06,550 Ejaaz: I mean, they are just not stopping and they're a force to reckon with. 439 00:26:07,010 --> 00:26:11,110 Ejaaz: And kind of similar to them, Josh, just to kind of round this episode out and 440 00:26:11,110 --> 00:26:12,790 Ejaaz: thank you guys for listening. 441 00:26:13,330 --> 00:26:15,570 Ejaaz: We are here in Argentina, in 442 00:26:15,570 --> 00:26:18,590 Ejaaz: Buenos Aires. We are kind of meeting some of the fans that are out here. 443 00:26:18,650 --> 00:26:23,290 Ejaaz: And we spoke to one just this afternoon, Josh. And you know what he said to me? Have a guess. 444 00:26:23,410 --> 00:26:23,730 Josh: What's that? 445 00:26:24,550 --> 00:26:30,090 Ejaaz: He said, your podcast Limitless is like the state-of-the-art AI podcast. 446 00:26:30,390 --> 00:26:34,910 Ejaaz: In fact, it is 2X better than any other AI podcast that I've ever heard. And you know what? 447 00:26:35,130 --> 00:26:36,990 Ejaaz: That sounds very similar to 448 00:26:36,990 --> 00:26:40,830 Ejaaz: the Gemini 3. So you could potentially call us the Gemini 3 of AI podcast. 449 00:26:41,230 --> 00:26:46,570 Ejaaz: And so if you're a listener to this, if you are a non-subscriber on our YouTube, 450 00:26:46,870 --> 00:26:48,450 Ejaaz: you should probably click that subscriber button. 451 00:26:48,530 --> 00:26:50,990 Ejaaz: You should probably click that notification button because guess what? 452 00:26:51,110 --> 00:26:56,170 Ejaaz: We've got more episodes coming this week. And guess what? the five star ratings help us out massively. 453 00:26:56,390 --> 00:26:59,290 Ejaaz: So if you enjoyed this episode and if you want to hear more episodes of this 454 00:26:59,290 --> 00:27:04,430 Ejaaz: nature and of cutting edge news in AI, you should give us a follow and we will see