1 00:00:00,020 --> 00:00:04,700 Ejaaz: Google might have just popped the AI bubble. They released a new algorithm yesterday, 2 00:00:05,240 --> 00:00:08,920 Ejaaz: which shrinks the requirement of memory needed to run an AI model, 3 00:00:09,080 --> 00:00:13,680 Ejaaz: which isn't great news for the top memory stocks, which had billions of dollars 4 00:00:13,680 --> 00:00:16,460 Ejaaz: of their market cap wiped out over the last 24 hours. 5 00:00:16,600 --> 00:00:19,780 Ejaaz: We're looking at SanDisk is down almost 7% today. 6 00:00:20,260 --> 00:00:24,460 Ejaaz: Micron is down 18% over the past five days, the majority of that yesterday. 7 00:00:24,900 --> 00:00:28,820 Ejaaz: And SK Hynix, another major memory provider is down 6% today. 8 00:00:29,040 --> 00:00:32,480 Ejaaz: It's absolutely decimated the market. We're going to get into this story and 9 00:00:32,480 --> 00:00:33,640 Ejaaz: so much more on this episode. 10 00:00:33,960 --> 00:00:37,040 Josh: Yeah. And some important context to know before we get into this is the reason 11 00:00:37,040 --> 00:00:40,960 Josh: why these memory stocks are so high is because memory has been the single largest 12 00:00:40,960 --> 00:00:42,280 Josh: constraint in the world of AI. 13 00:00:42,500 --> 00:00:47,920 Josh: If you have ever tried to build a PC or buy a PC in the last 12 to 18 months, 14 00:00:48,040 --> 00:00:53,200 Josh: chances are it's almost doubly as expensive as it was the previous 18 months. That's because AI. 15 00:00:53,460 --> 00:00:57,760 Josh: It's what happened when crypto stole all the GPUs. AI is taking all of the memory 16 00:00:57,760 --> 00:01:01,840 Josh: and memory is the limiting factor memory is what serves inference whenever you 17 00:01:01,840 --> 00:01:03,960 Josh: ask chat gpt or anthropica question, 18 00:01:04,410 --> 00:01:08,390 Josh: When it serves you the answer, that uses inference, that uses memory. 19 00:01:08,670 --> 00:01:11,610 Josh: So memory has been the single crux. It's the reason why companies haven't been 20 00:01:11,610 --> 00:01:14,970 Josh: able to scale as fast as they want, because the costs have been out of control. 21 00:01:15,190 --> 00:01:19,550 Josh: This algorithm changes things a little bit. This changes the math here just 22 00:01:19,550 --> 00:01:24,170 Josh: a smidge, but enough to make it scare the market entirely, right? 23 00:01:24,290 --> 00:01:29,010 Ejaaz: Yeah, it changes it in a huge way. And the wildest part is the Google research 24 00:01:29,010 --> 00:01:33,230 Ejaaz: team made this public and available for everyone to read the blueprint of. 25 00:01:33,230 --> 00:01:38,110 Ejaaz: So it's called TurboQuant, and they're describing it as a new compression algorithm 26 00:01:38,110 --> 00:01:45,090 Ejaaz: that reduces LLM key value cache memory by six times and delivers up to 8x speedup. 27 00:01:45,170 --> 00:01:47,130 Ejaaz: So let me translate what that means. 28 00:01:47,290 --> 00:01:52,790 Ejaaz: When you type or speak to an AI model, it has a short-term memory. 29 00:01:52,910 --> 00:01:55,530 Ejaaz: Think about like humans, they kind of hold things in their head for a bit, 30 00:01:55,630 --> 00:01:58,890 Ejaaz: and then they go to bed, and then they commit some of those memories for long-term 31 00:01:58,890 --> 00:02:02,390 Ejaaz: memory, and some of those they kind of discard. But LLMs work in a similar way, 32 00:02:02,470 --> 00:02:04,370 Ejaaz: and it's called the cache memory. 33 00:02:04,550 --> 00:02:09,230 Ejaaz: Now, the issue with LLMs are that cache memory, as you talk to an AI more and 34 00:02:09,230 --> 00:02:14,690 Ejaaz: more, gets really quite big and bloated, which means that AIs are very expensive to run. 35 00:02:14,830 --> 00:02:17,950 Ejaaz: That's why we spend so much money on GPUs. That's why these data centers cost 36 00:02:17,950 --> 00:02:19,430 Ejaaz: billions of dollars to build out. 37 00:02:19,650 --> 00:02:22,330 Ejaaz: It is part of the reason when you inference a model. 38 00:02:22,730 --> 00:02:25,510 Ejaaz: Now, what Google did was they released an algorithm which said, 39 00:02:25,970 --> 00:02:28,190 Ejaaz: hey, yeah, you know that thing that bloated up your AI model? 40 00:02:28,190 --> 00:02:31,170 Ejaaz: We can cut that down pretty massively by six times, actually, 41 00:02:31,330 --> 00:02:33,710 Ejaaz: and save you a ton of money. 42 00:02:33,830 --> 00:02:37,210 Ejaaz: So the world is going crazy with this new research, because if you could scale 43 00:02:37,210 --> 00:02:42,430 Ejaaz: this up to Claude or ChatGPT, you may not need as many GPUs as we originally 44 00:02:42,430 --> 00:02:45,530 Ejaaz: thought to serve Frontier AI models to anyone. 45 00:02:45,910 --> 00:02:49,090 Josh: Yeah, it's a really big deal. Like you mentioned, every time you message a chatbot, 46 00:02:49,430 --> 00:02:51,870 Josh: the model stores your entire conversation in that KV cache. 47 00:02:52,110 --> 00:02:55,890 Josh: And on a large model, like the 70 billion parameter model, if you run a long 48 00:02:55,890 --> 00:03:00,870 Josh: conversation, that cache will eat up to like 40 gigabytes of space on your GPU, 49 00:03:00,990 --> 00:03:03,150 Josh: which is more than the actual model itself. 50 00:03:03,410 --> 00:03:07,150 Josh: So that's why this is such a huge issue. And the big breakthrough here is that 51 00:03:07,150 --> 00:03:09,150 Josh: Google can do so without any loss functions. 52 00:03:09,330 --> 00:03:14,150 Josh: If anyone's a Silicon Valley fan, the TV show, the crown jewel of the entire 53 00:03:14,150 --> 00:03:18,510 Josh: show was this middle out compression algorithm that didn't have any loss function. 54 00:03:18,670 --> 00:03:21,210 Josh: You can just compress the data without losing any of the quality. 55 00:03:21,330 --> 00:03:23,730 Josh: And Google's announcing that they've actually done this. 56 00:03:23,850 --> 00:03:27,130 Josh: They've actually solved this problem. And that's why the memory market is freaking out. 57 00:03:27,330 --> 00:03:30,770 Josh: It's like, oh my God, wait, if we can get six times the compression, 58 00:03:31,190 --> 00:03:35,690 Josh: eight times faster on the same GPUs, surely there will be less demand for memory. 59 00:03:35,810 --> 00:03:41,030 Ejaaz: Well, despite these stocks being down massively, I think the market is kind of overreacting. 60 00:03:41,270 --> 00:03:45,690 Ejaaz: This is a good thing for AI in general, and it will lead to more consumption of AI. 61 00:03:45,810 --> 00:03:49,250 Ejaaz: It's great that the AI models become cheaper to inference and that it frees 62 00:03:49,250 --> 00:03:52,870 Ejaaz: up space for us to compute other AI prompts and queries. 63 00:03:53,390 --> 00:03:56,890 Ejaaz: Jevon's paradox is going to occur here in the same way that people were kind 64 00:03:56,890 --> 00:04:00,370 Ejaaz: of crapping on GPUs and saying these things are too expensive and aren't going to get cheaper. 65 00:04:00,470 --> 00:04:04,010 Ejaaz: The same thing is going to happen with memory here. It's not going to result in less demand. 66 00:04:04,170 --> 00:04:07,910 Ejaaz: It's going to result in even more demand. The demand for AI products right now 67 00:04:07,910 --> 00:04:11,050 Ejaaz: is insatiable, and I don't think this changes anytime soon. 68 00:04:11,210 --> 00:04:15,190 Ejaaz: The other part is, I think that what Google has created here 69 00:04:15,600 --> 00:04:20,740 Ejaaz: is something that's going to take time to scale out. This is not going to happen tomorrow. 70 00:04:20,920 --> 00:04:23,740 Ejaaz: This is a research paper, which you and I were actually joking about before 71 00:04:23,740 --> 00:04:26,220 Ejaaz: we started recording, was released 11 months ago. 72 00:04:26,320 --> 00:04:30,940 Ejaaz: But because the paper was so complex and had so many algorithms and mathematical 73 00:04:30,940 --> 00:04:33,700 Ejaaz: formulas, it didn't actually surface to the public until yesterday. 74 00:04:33,800 --> 00:04:35,720 Ejaaz: And everyone was like, wow. So it takes time. 75 00:04:36,080 --> 00:04:38,900 Ejaaz: Since those 11 months, we haven't scaled it out to major models just yet. 76 00:04:38,980 --> 00:04:40,660 Ejaaz: This only applies to small to medium models. 77 00:04:40,780 --> 00:04:43,920 Ejaaz: But I can imagine the scaling to frontier models at some point, 78 00:04:43,960 --> 00:04:45,500 Ejaaz: which will be a win-win for both sides. 79 00:04:45,600 --> 00:04:48,600 Josh: I mean yeah there's there's no reason why you wouldn't be focused solely 80 00:04:48,600 --> 00:04:51,540 Josh: on this right it's like if you can serve six times the inference on the 81 00:04:51,540 --> 00:04:54,600 Josh: same amount of gpus or tpus or whatever gpu accelerator you're 82 00:04:54,600 --> 00:04:57,420 Josh: using that's a huge breakthrough for any company and it also 83 00:04:57,420 --> 00:05:00,620 Josh: extends past the cloud it moves to local inference as 84 00:05:00,620 --> 00:05:03,540 Josh: well it's like if you're running a mac mini or you're even 85 00:05:03,540 --> 00:05:06,420 Josh: on your iphone the amount of intelligence you 86 00:05:06,420 --> 00:05:09,820 Josh: can run on a model that is six times more efficient 87 00:05:09,820 --> 00:05:12,880 Josh: eight times faster and has zero 88 00:05:12,880 --> 00:05:16,420 Josh: loss function is a really big deal and i think google 89 00:05:16,420 --> 00:05:19,320 Josh: may be moving towards this what i see this trend happening with 90 00:05:19,320 --> 00:05:22,140 Josh: google is i mean they have the tpu they're building their own stack 91 00:05:22,140 --> 00:05:25,840 Josh: they have the partnership with apple to get into all of the iphones they have 92 00:05:25,840 --> 00:05:29,100 Josh: their like kind of android division they're they're publishing their research 93 00:05:29,100 --> 00:05:33,400 Josh: publicly to disrupt the market as they go it's a really noteworthy strategy 94 00:05:33,400 --> 00:05:37,180 Josh: from google and it seems like as always they're leading on the thought leader 95 00:05:37,180 --> 00:05:40,280 Josh: front right they've always been the top tier of research we'll see how they 96 00:05:40,280 --> 00:05:41,660 Josh: can implement this into actual products 97 00:05:42,100 --> 00:05:46,120 Josh: But it seems like the market's wrong. I can't imagine a world in which there 98 00:05:46,120 --> 00:05:47,840 Josh: is less memory demand as a result of this. 99 00:05:48,180 --> 00:05:51,760 Josh: Like, as you said, there will only be more. If we get six times more intelligence, 100 00:05:51,760 --> 00:05:55,380 Josh: we will gobble that up so quickly. So market seems like it's wrong. 101 00:05:55,560 --> 00:05:59,820 Josh: Maybe it's just an excuse to dump because the stocks have gone up so much over the last 12 months. 102 00:06:00,020 --> 00:06:04,220 Josh: But I wouldn't be too worried about this if you are a owner of any sort of memory. 103 00:06:04,360 --> 00:06:07,660 Ejaaz: It's actually very bullish Google stock, in my opinion. I actually bought a 104 00:06:07,660 --> 00:06:10,720 Ejaaz: bunch more when this came out because I was like, oh, this is going to result 105 00:06:10,720 --> 00:06:12,620 Ejaaz: in massive efficiencies for AI. 106 00:06:12,980 --> 00:06:17,040 Ejaaz: Another thing is like, lest we forget, Google has been like the pioneer of a 107 00:06:17,040 --> 00:06:18,280 Ejaaz: lot of these different AI trends. 108 00:06:18,500 --> 00:06:24,420 Ejaaz: Like one of the earliest LLM formations was from a bunch of Google or ex-Google researchers. 109 00:06:25,020 --> 00:06:28,460 Ejaaz: TPUs they've been working on for over a decade now. So they saw the GPU trend 110 00:06:28,460 --> 00:06:31,520 Ejaaz: like way, way, way in advance. And now they're doing the same here. 111 00:06:31,680 --> 00:06:36,080 Ejaaz: My question to you is, why do you think they publicize this for everyone to see? 112 00:06:36,260 --> 00:06:39,560 Ejaaz: Surely this, like i saw a few comments about this as well like surely they should 113 00:06:39,560 --> 00:06:44,540 Ejaaz: have patented this because this is a pretty sick algorithm that now like any company can use 114 00:06:44,540 --> 00:06:47,540 Josh: Yeah i don't think the patenting thing really works anywhere 115 00:06:47,540 --> 00:06:50,320 Josh: at all because a lot of this stuff is just a couple lines of 116 00:06:50,320 --> 00:06:53,100 Josh: code it's like andre wrote auto research and it's 117 00:06:53,100 --> 00:06:56,080 Josh: 600 lines of code it's like someone will take it 118 00:06:56,080 --> 00:06:59,380 Josh: someone will use it the the race is far too powerful 119 00:06:59,380 --> 00:07:02,400 Josh: far too there's far too much of a monetary incentive to 120 00:07:02,400 --> 00:07:05,160 Josh: do this and you could think of it almost like google's deep 121 00:07:05,160 --> 00:07:08,020 Josh: seek moment where deep seek created this unbelievable 122 00:07:08,020 --> 00:07:10,800 Josh: research dropped the bomb they got all the publicity but 123 00:07:10,800 --> 00:07:14,160 Josh: then the rest of the market kind of rushed to tackle that this is google dropping 124 00:07:14,160 --> 00:07:19,000 Josh: their their deep seek moments and i think it's it's noteworthy it's really impressive 125 00:07:19,000 --> 00:07:22,520 Josh: for google google has been kind of at the forefront of all this stuff and they're 126 00:07:22,520 --> 00:07:26,500 Josh: proving that their research team really are super impressive and have the ability 127 00:07:26,500 --> 00:07:29,380 Josh: to continue to lead the frontier forward. 128 00:07:29,840 --> 00:07:34,740 Ejaaz: So speaking of companies that are at the frontier, one of those companies are 129 00:07:34,740 --> 00:07:36,780 Ejaaz: falling behind this week, Josh. 130 00:07:36,940 --> 00:07:41,000 Josh: We got to do a little victory lap, slam dunk. I'm not really sure how to frame this. 131 00:07:41,140 --> 00:07:45,260 Ejaaz: But this is about Sora. Yeah, let's start with the victory lap, I guess. 132 00:07:45,480 --> 00:07:51,660 Ejaaz: So Sora, for those of you who don't know, is one of OpenAI's winning consumer apps. 133 00:07:51,800 --> 00:07:55,820 Ejaaz: It is their AI video generation app that functions very similarly to TikTok. 134 00:07:56,260 --> 00:07:58,640 Ejaaz: You kind of scroll through Reels, but all the videos are air-generated. 135 00:07:58,740 --> 00:08:01,700 Ejaaz: But the coolest part is you can feature yourself in them. 136 00:08:01,840 --> 00:08:04,560 Ejaaz: You can feature your friends or any kind of IP or characters. 137 00:08:04,900 --> 00:08:09,480 Ejaaz: This week, they announced that they were sunsetting the entire project. 138 00:08:09,680 --> 00:08:13,140 Ejaaz: Sora, which six months ago wasn't a thing, is now dead. 139 00:08:13,460 --> 00:08:16,960 Ejaaz: Some fun stats from Sora's very short but mighty reign. 140 00:08:17,400 --> 00:08:22,300 Ejaaz: Within the first day, they achieved over 5 million downloads, which is just insane. 141 00:08:22,480 --> 00:08:25,420 Ejaaz: I don't think any other app has done that before. they also hit 142 00:08:25,420 --> 00:08:28,100 Ejaaz: number one and stayed at number one of the apple app store and i 143 00:08:28,100 --> 00:08:30,940 Ejaaz: think the android play app store as well for a number of different 144 00:08:30,940 --> 00:08:33,640 Ejaaz: weeks but it wasn't without any kind of 145 00:08:33,640 --> 00:08:39,780 Ejaaz: grimace the public was split hollywood is basically uh how do i say um dancing 146 00:08:39,780 --> 00:08:44,180 Ejaaz: on their graves uh open ai is shutting down its ai video slot making platform 147 00:08:44,180 --> 00:08:48,000 Ejaaz: sora and if you look at the comments from different people it's just like did 148 00:08:48,000 --> 00:08:52,180 Ejaaz: we just win uh people are like saying like this is such a big W. 149 00:08:52,420 --> 00:08:57,680 Ejaaz: So people are generally happy outside of their AI sphere. But I don't know, 150 00:08:57,720 --> 00:08:59,520 Ejaaz: Josh, how do you feel about this? Are you happy? 151 00:08:59,960 --> 00:09:03,060 Josh: Yeah, well, this was just a failed experiment in the way that Meta did. 152 00:09:03,200 --> 00:09:06,200 Josh: I forget. I don't even remember Meta's. What was it? 153 00:09:06,380 --> 00:09:06,740 Ejaaz: Vibes? 154 00:09:07,220 --> 00:09:10,020 Josh: Vibes perhaps vibes where they tried to do the same thing it 155 00:09:10,020 --> 00:09:13,200 Josh: was tiktok but for ai and open ai tried tiktok but 156 00:09:13,200 --> 00:09:16,140 Josh: for ai again but through the sora platform and i 157 00:09:16,140 --> 00:09:18,800 Josh: am i mean we said this was stupid i haven't opened the 158 00:09:18,800 --> 00:09:21,800 Josh: app since the week of i think there's there's a 159 00:09:21,800 --> 00:09:24,860 Josh: harsh reality that open ai is learning that they perhaps don't understand 160 00:09:24,860 --> 00:09:27,920 Josh: just due to the nature of them being a new company sam really 161 00:09:27,920 --> 00:09:30,820 Josh: this is his first large company that he's run but i mean 162 00:09:30,820 --> 00:09:33,820 Josh: it seems like a very predictable outcome when you just look 163 00:09:33,820 --> 00:09:36,840 Josh: at what standard consumer behavior is and you kind of compare to other platforms 164 00:09:36,840 --> 00:09:39,660 Josh: because when you think about a platform like 165 00:09:39,660 --> 00:09:43,220 Josh: meta right it has two billion users or something like that and particularly 166 00:09:43,220 --> 00:09:46,860 Josh: on instagram how many people of those 10 billion actually want to create content 167 00:09:46,860 --> 00:09:51,920 Josh: maybe 10 million maybe 0.5 percent of the people would actually be serious creators 168 00:09:51,920 --> 00:09:55,560 Josh: and the reality is is that almost every single human just wants to scroll and 169 00:09:55,560 --> 00:09:59,440 Josh: zone out instead of spending energy like conceptualizing and editing these complicated things, 170 00:10:00,210 --> 00:10:05,630 Josh: The novelty factor of it was very fun, but people don't want this and it's not that good. 171 00:10:05,790 --> 00:10:09,390 Josh: And I think that novelty wears off. It's like we created those videos that show 172 00:10:09,390 --> 00:10:12,350 Josh: us and you in them, and it was really cool for like two hours, 173 00:10:12,350 --> 00:10:15,770 Josh: but creating a dedicated application where you have to go and download it and 174 00:10:15,770 --> 00:10:20,130 Josh: you have to use it outside of the ChatGPT OpenAI universe is very annoying. 175 00:10:20,590 --> 00:10:23,270 Josh: Had they rolled this up into the same app that everyone's using every day with 176 00:10:23,270 --> 00:10:26,990 Josh: ChatGPT, maybe that's a different story, but it's, I don't know, 177 00:10:27,110 --> 00:10:30,810 Josh: it's tough to kind of reason why they would have gone through with this in the 178 00:10:30,810 --> 00:10:31,690 Josh: first place the way they did. 179 00:10:31,790 --> 00:10:36,190 Ejaaz: Let me try and argue the other side. I think what Sora did really well was they 180 00:10:36,190 --> 00:10:40,570 Ejaaz: made it really easy if you were super lazy, but you kind of had an idea for 181 00:10:40,570 --> 00:10:42,810 Ejaaz: a piece of content to make that content. 182 00:10:43,150 --> 00:10:46,550 Ejaaz: Just write it in words, takes a couple seconds, maybe max a minute. 183 00:10:46,670 --> 00:10:49,530 Ejaaz: And then a couple of minutes later, you have a fully fledged video. 184 00:10:49,670 --> 00:10:52,450 Ejaaz: Now, obviously, Sora v1 wasn't that great. 185 00:10:52,590 --> 00:10:55,930 Ejaaz: Sora v2 started to improve and become much better. Sora v3 was way better, 186 00:10:56,070 --> 00:11:00,350 Ejaaz: but there was still some AI cringiness around this. I do think this improves 187 00:11:00,350 --> 00:11:03,770 Ejaaz: with a bunch of now other AI models that still exist to do this. 188 00:11:04,110 --> 00:11:08,190 Ejaaz: CapCut actually yesterday released their own version just as Sora was sunsetted 189 00:11:08,190 --> 00:11:11,110 Ejaaz: that uses Seed Dance 2, which is a Chinese AI model. 190 00:11:11,290 --> 00:11:15,670 Ejaaz: And my God, the content that creators can now create just through words looks insane there. 191 00:11:15,850 --> 00:11:19,570 Ejaaz: And then you have Grok Imagine on the ex-AI side under Elon Musk, 192 00:11:19,750 --> 00:11:23,890 Ejaaz: and they're piling so many millions and millions of dollars to try and improve that model. 193 00:11:24,010 --> 00:11:29,350 Ejaaz: So there is something there. I just think that a lot of this reason behind OpenAI doing this is 194 00:11:29,720 --> 00:11:34,100 Ejaaz: One, they're constrained on cash. They've spent a lot of money on GPUs, 195 00:11:34,100 --> 00:11:35,340 Ejaaz: Stargate. They had to sunset that. 196 00:11:35,700 --> 00:11:38,980 Ejaaz: And now they're figuring out that they need to focus all their cash and resources 197 00:11:38,980 --> 00:11:43,700 Ejaaz: on building the best LLM to beat Anthropic and the best coding AI to beat Anthropic, 198 00:11:43,860 --> 00:11:46,380 Ejaaz: who is, quite frankly, eating their lunch right now. 199 00:11:46,520 --> 00:11:49,460 Ejaaz: But there is another bit of news, which I didn't realize immediately. 200 00:11:49,720 --> 00:11:51,980 Ejaaz: But then I noticed Disney. 201 00:11:52,440 --> 00:11:58,240 Ejaaz: OpenAI had signed a $1 billion deal with Disney through Sora. 202 00:11:58,240 --> 00:12:02,280 Ejaaz: And the idea here was Disney would invest $1 billion in OpenAI, 203 00:12:02,560 --> 00:12:07,860 Ejaaz: and OpenAI would also get access to Disney's IP and characters and make them 204 00:12:07,860 --> 00:12:12,440 Ejaaz: super famous on Sora via the AI-generated type of video medium. 205 00:12:12,620 --> 00:12:14,660 Ejaaz: It sounds like Disney didn't even know. 206 00:12:14,940 --> 00:12:18,880 Ejaaz: I've seen multiple tweets like this, where on Friday, they were signing a deal 207 00:12:18,880 --> 00:12:23,760 Ejaaz: with OpenAI, and then on Wednesday of this week, when they decided to announce 208 00:12:23,760 --> 00:12:26,160 Ejaaz: the news that they were shutting down Sora, the Disney team had no idea. 209 00:12:26,320 --> 00:12:31,500 Josh: It's tough. I think this is a good lesson. there is this quote from Johnny Ive 210 00:12:31,500 --> 00:12:35,320 Josh: that I love that I think would be nice for Sam Altman to hear when he was asked 211 00:12:35,320 --> 00:12:37,340 Josh: about what Steve Jobs taught him about focus. 212 00:12:37,680 --> 00:12:41,180 Josh: And he says, this sounds really simplistic, but it shocks me how few people 213 00:12:41,180 --> 00:12:42,020 Josh: actually practice this. 214 00:12:42,260 --> 00:12:44,940 Josh: One of the things that Steve would say to me was because he was worried that 215 00:12:44,940 --> 00:12:47,000 Josh: I wouldn't focus. He would say, how many things have you said no to? 216 00:12:47,180 --> 00:12:49,340 Josh: And I would tell him, I said no to this and I said no to that. 217 00:12:49,460 --> 00:12:51,900 Josh: And he would create these sacrificial things that he was saying no to. 218 00:12:52,140 --> 00:12:56,020 Josh: But focus means saying no to something with every bone in your body you think is a phenomenal idea. 219 00:12:56,300 --> 00:12:59,220 Josh: And you wake up thinking about it, But because you are focusing on something 220 00:12:59,220 --> 00:13:00,420 Josh: else, you can't work on that one thing. 221 00:13:00,720 --> 00:13:03,160 Josh: And OpenAI does not have that focus. 222 00:13:03,540 --> 00:13:06,880 Josh: OpenAI is creating these sacrificial things. They are not deeply focused on 223 00:13:06,880 --> 00:13:08,040 Josh: one thing. And that's why there's 224 00:13:08,040 --> 00:13:10,240 Josh: four different apps you have to download to use all of their tools. 225 00:13:10,540 --> 00:13:14,940 Josh: There's just all of these spread attempts at creating virality, 226 00:13:15,100 --> 00:13:19,680 Josh: at manufacturing virality through TikTokification of it, through this like image 227 00:13:19,680 --> 00:13:21,340 Josh: generation of the Studio Ghibli stuff. 228 00:13:21,620 --> 00:13:25,720 Josh: None of it is just building their core thing. And I think when you look at Anthropica, 229 00:13:25,780 --> 00:13:29,300 Josh: when you look at OpenAI, and the difference between the two is that one has 230 00:13:29,300 --> 00:13:31,800 Josh: focus and one does not. When you look at OpenAI... 231 00:13:31,970 --> 00:13:35,510 Josh: You see five different product SKUs. They're all kind of all over the place 232 00:13:35,510 --> 00:13:36,190 Josh: with different intentions. 233 00:13:36,390 --> 00:13:39,750 Josh: When you look at Anthropic, they're the best coding model in the world. And that's it. 234 00:13:39,930 --> 00:13:43,050 Josh: Everyone can converge on that one point. And therefore their velocity is so 235 00:13:43,050 --> 00:13:45,050 Josh: fast. That's why we just filmed an episode yesterday. 236 00:13:45,230 --> 00:13:48,430 Josh: Everyone should go watch about the eight updates they published in eight weeks 237 00:13:48,430 --> 00:13:50,150 Josh: that basically entirely replaced OpenClaw. 238 00:13:50,350 --> 00:13:54,810 Josh: They have the focus. And I'm hoping that OpenAI can now kind of shift this focus 239 00:13:54,810 --> 00:13:58,530 Josh: into this singular place and start to move towards this new organization that 240 00:13:58,530 --> 00:14:00,290 Josh: they are calling AGI Deployment. 241 00:14:00,370 --> 00:14:02,850 Josh: This is crazy. They're preparing for AGI already. 242 00:14:03,230 --> 00:14:08,370 Ejaaz: Yeah, so the team is basically getting reshuffled under a product organization 243 00:14:08,370 --> 00:14:10,610 Ejaaz: within OpenAI called AGI Deployment. 244 00:14:10,730 --> 00:14:14,750 Ejaaz: And this kind of harkens to a wider strategy that OpenAI has assumed for the 245 00:14:14,750 --> 00:14:18,230 Ejaaz: last couple of months now, which is to build the best LLM, build the best coding 246 00:14:18,230 --> 00:14:22,890 Ejaaz: AI model, and also the best world model, which is actually where the Sora team 247 00:14:22,890 --> 00:14:24,230 Ejaaz: researchers are moving to. 248 00:14:24,350 --> 00:14:28,130 Ejaaz: They're going to focus on building a world model, which basically helps AI models 249 00:14:28,130 --> 00:14:31,330 Ejaaz: see the world and understand the physics of the world. the advantage of this 250 00:14:31,330 --> 00:14:35,610 Ejaaz: is you get to better understand how humans perceive and interact with the world, 251 00:14:35,730 --> 00:14:37,250 Ejaaz: which is something that LLMs can't do. 252 00:14:37,370 --> 00:14:41,070 Ejaaz: Think of LLMs as an AI model that sits in the library and reads all the books, 253 00:14:41,190 --> 00:14:42,950 Ejaaz: but doesn't actually experience the world for himself. 254 00:14:43,210 --> 00:14:45,830 Ejaaz: World models actually help you understand the world. That's what the Sora team 255 00:14:45,830 --> 00:14:50,070 Ejaaz: is going to focus on right now. And that's super exciting for one major reason, which is robotics. 256 00:14:50,390 --> 00:14:51,830 Ejaaz: Robotics is going to be a huge thing. 257 00:14:52,210 --> 00:14:56,170 Ejaaz: OpenAI teased robotics in one of their major announcements a couple of weeks 258 00:14:56,170 --> 00:14:58,650 Ejaaz: ago. So we know that they're going to focus heavily on that. 259 00:14:58,750 --> 00:14:59,970 Ejaaz: And I do think that's a smart move. 260 00:15:00,110 --> 00:15:04,290 Ejaaz: I'm bullish OpenAI after all of this. But just to recount, this isn't the only 261 00:15:04,290 --> 00:15:06,790 Ejaaz: major pivot they've made over the last week. 262 00:15:06,970 --> 00:15:08,530 Ejaaz: Over the last couple of weeks, they've done quite a few things. 263 00:15:08,650 --> 00:15:09,610 Ejaaz: So they've sunsetted Sora. 264 00:15:09,870 --> 00:15:12,530 Ejaaz: They've announced that their Stargate project, which was like their major project 265 00:15:12,530 --> 00:15:17,470 Ejaaz: to scale GPUs across the globe, is now canceled one year later due to funding issues. 266 00:15:18,110 --> 00:15:19,770 Ejaaz: And Altman said that like, you know, 267 00:15:19,870 --> 00:15:22,870 Ejaaz: I'm never going to release ads on the platform. He ended up doing that. 268 00:15:23,310 --> 00:15:26,990 Ejaaz: In-app shopping completely failed and flopped. And now they've got to redesign 269 00:15:26,990 --> 00:15:29,550 Ejaaz: that entire thing but it's been put on pause for now until they focus on all 270 00:15:29,550 --> 00:15:31,270 Ejaaz: this other stuff all of this to say 271 00:15:31,680 --> 00:15:34,800 Ejaaz: I think OpenAI is going to IPO soon, Josh. And I don't know what you think the 272 00:15:34,800 --> 00:15:37,720 Ejaaz: odds are for this, but we should probably like bet on that. 273 00:15:37,880 --> 00:15:43,020 Josh: Well, if only we had a marketplace for it. And thanks to our friends at Polymarket, when will OpenAI IPO by? 274 00:15:43,200 --> 00:15:48,900 Josh: And the answer is kind of surprising where there's only a 36% chance that they'll even IPO this year. 275 00:15:49,060 --> 00:15:52,020 Josh: And I think earlier in the year, this would have been a little bit different, 276 00:15:52,160 --> 00:15:55,020 Josh: but it appears as if there's been some trouble in paradise for Anthropic. 277 00:15:55,240 --> 00:15:57,360 Josh: Maybe that is moving over to OpenAI. 278 00:15:57,620 --> 00:16:01,540 Josh: The reality that we're starting to see now is SpaceX might be the big one, and that might be it. 279 00:16:01,680 --> 00:16:04,300 Josh: Um which is a little disappointing because i was hoping this would be 280 00:16:04,300 --> 00:16:07,860 Josh: the year of all of the ipos but according to polymarket 36 chance 281 00:16:07,860 --> 00:16:11,260 Josh: and if you think it's happening this june there's four percent chance so that's 282 00:16:11,260 --> 00:16:15,000 Josh: absolutely not happening spacex is going to be the one that will be taking that 283 00:16:15,000 --> 00:16:17,900 Josh: crown uh polymarket thank you very much for sponsoring this part of the episode 284 00:16:17,900 --> 00:16:19,060 Josh: and maybe we should just go right 285 00:16:19,060 --> 00:16:24,440 Josh: into spacex and their ipo which now is rumored to be filing next week, 286 00:16:25,020 --> 00:16:27,860 Josh: confidentially this will not be public but they will do their confidential s1 287 00:16:27,860 --> 00:16:32,820 Josh: filing later this week, early next week, which would point them to a public listing, 288 00:16:33,510 --> 00:16:38,210 Josh: around the first couple of weeks of June. This is a really big deal for the 289 00:16:38,210 --> 00:16:43,590 Josh: world because, I mean, this is likely to be now over a $2 trillion IPO. 290 00:16:44,410 --> 00:16:48,990 Josh: That's the biggest IPO by far. This is going to be the largest thing to enter the market. 291 00:16:49,150 --> 00:16:53,070 Josh: It's going to be larger than the GDP of some countries, and it will just go 292 00:16:53,070 --> 00:16:57,270 Josh: live on the New York Stock Exchange or the NASDAQ sometime in a few months from 293 00:16:57,270 --> 00:16:58,030 Josh: now. We're getting close. 294 00:16:58,230 --> 00:17:02,570 Ejaaz: What a sequence of events to get us to this point. we had the rumors of SpaceX 295 00:17:02,570 --> 00:17:05,330 Ejaaz: and XAI emerging and then it happened. 296 00:17:05,610 --> 00:17:10,050 Ejaaz: Now we have rumors of SpaceX and Tesla merging. But first, I think we're going to get the SpaceX IPO. 297 00:17:10,430 --> 00:17:15,110 Ejaaz: Rumors also say that they're going to be raising $75 billion in this listing, 298 00:17:15,290 --> 00:17:19,070 Ejaaz: which is just absolutely insane and probably the largest raise that's ever been 299 00:17:19,070 --> 00:17:21,290 Ejaaz: done on an American public stock market. 300 00:17:21,570 --> 00:17:25,050 Ejaaz: The other thing is this comes off the back of Elon announcing some pretty ambitious 301 00:17:25,050 --> 00:17:26,670 Ejaaz: projects for SpaceX itself. 302 00:17:27,090 --> 00:17:29,510 Ejaaz: They're going to be launching a TerraFab. We made an episode about this. 303 00:17:29,590 --> 00:17:33,590 Ejaaz: Definitely go check that out, which is a gigantic AI chip factory. 304 00:17:33,750 --> 00:17:36,950 Ejaaz: They're going to be launching 80% of those chips into space via SpaceX. 305 00:17:37,050 --> 00:17:40,250 Ejaaz: So it's going to become this whole monopoly of like AI in space, 306 00:17:40,390 --> 00:17:42,530 Ejaaz: which I can't wait to see. Mass drivers on the moon. 307 00:17:42,950 --> 00:17:45,550 Ejaaz: Definitely go check out that episode earlier this week. But yeah, 308 00:17:45,670 --> 00:17:48,090 Ejaaz: it's going to be one of the biggest IPOs, Josh. Are you buying it? Absolutely. 309 00:17:48,370 --> 00:17:52,850 Josh: I cannot miss it. I currently own SpaceX. It's the only private company I own. I'm ready to buy more. 310 00:17:53,030 --> 00:17:56,650 Josh: I think it's going to be the most valuable company in the world. rolled in with tesla and 311 00:17:56,650 --> 00:18:00,610 Josh: i could not be more bullish more optimistic more excited for the future with 312 00:18:00,610 --> 00:18:03,770 Josh: these companies if elon's predictions are right we'll be seeing nvidia at 10 313 00:18:03,770 --> 00:18:07,330 Josh: trillion and spacex and tesla will be right there with them um the companies 314 00:18:07,330 --> 00:18:10,330 Josh: are going to grow quick i think the spacex ipo is very much the starting gun 315 00:18:10,330 --> 00:18:13,770 Josh: now there is another gun that's been fired and it's been a not not a good one 316 00:18:13,770 --> 00:18:15,770 Josh: from apple not a good because they have been, 317 00:18:16,380 --> 00:18:19,260 Josh: banging out all the applications of everyone who 318 00:18:19,260 --> 00:18:22,520 Josh: wants to get in the app store now there's been a recent issue with the app store 319 00:18:22,520 --> 00:18:25,420 Josh: is now the fact that anybody can make apps means that a lot more people are 320 00:18:25,420 --> 00:18:29,300 Josh: submitting apps so for the larger companies and the smaller companies like if 321 00:18:29,300 --> 00:18:32,420 Josh: you have an app in the app store it's taken a lot longer to get your app approved 322 00:18:32,420 --> 00:18:37,020 Josh: whether it be for just a normal update or to release an actual app and now it 323 00:18:37,020 --> 00:18:38,700 Josh: appears that apple has stopped, 324 00:18:39,600 --> 00:18:42,780 Josh: allowing updates for popular Vibe-coded applications? 325 00:18:43,160 --> 00:18:45,660 Ejaaz: Yeah, they've basically put a 326 00:18:45,660 --> 00:18:51,060 Ejaaz: not official, but unofficial halt on anyone launching AI Vibe-coded apps. 327 00:18:51,280 --> 00:18:54,060 Ejaaz: And I'm confused about this for a few different reasons. 328 00:18:54,360 --> 00:18:59,800 Ejaaz: Number one, if you look at a chart of app launches on the App Store over the 329 00:18:59,800 --> 00:19:04,300 Ejaaz: last year, before Vibe-coding became a thing, it had completely plateaued. 330 00:19:04,560 --> 00:19:09,220 Ejaaz: No one was making apps anymore. And that was becoming an issue for the Apple 331 00:19:09,220 --> 00:19:10,720 Ejaaz: app developer ecosystem. They're 332 00:19:10,720 --> 00:19:13,060 Ejaaz: actually kind of trying to encourage developers to build more things. 333 00:19:13,220 --> 00:19:17,520 Ejaaz: Then AI vibe coding came along and suddenly they had a flourish of different 334 00:19:17,520 --> 00:19:20,160 Ejaaz: vibe coded apps. Now granted, 90% of them were crap, 335 00:19:20,760 --> 00:19:23,300 Ejaaz: 10% of them were pretty good. We got the replets of the world, 336 00:19:23,440 --> 00:19:28,220 Ejaaz: which helped other people kind of build apps just by typing words or talking to an AI on your phone. 337 00:19:28,580 --> 00:19:32,460 Ejaaz: And replet as a company is worth, I think, $9 billion. As of last week, 338 00:19:32,540 --> 00:19:34,080 Ejaaz: they did another massive raise. 339 00:19:34,220 --> 00:19:37,280 Ejaaz: So we aren't talking about small fish here. These are pretty big fish. 340 00:19:37,600 --> 00:19:42,980 Ejaaz: Apple just cut them at the head this week or last week over the last seven days 341 00:19:42,980 --> 00:19:46,560 Ejaaz: by saying that they are not going to encourage AI vibe coding anymore. 342 00:19:46,560 --> 00:19:51,260 Ejaaz: And it just signals that monopolies like Apple, which have a chokehold on the 343 00:19:51,260 --> 00:19:54,880 Ejaaz: app ecosystem like this, that take a massive 30% cut off of fees. 344 00:19:55,040 --> 00:19:58,080 Ejaaz: We saw this happen in the crypto world as well, where people wanted to allow 345 00:19:58,080 --> 00:20:01,060 Ejaaz: transactions. Apple said, we'll allow it for a bit and then we don't want this. 346 00:20:01,280 --> 00:20:04,660 Ejaaz: They're now doing it with Vibe coding. I don't get the issue for them unless 347 00:20:04,660 --> 00:20:07,860 Ejaaz: there's like major security exploits. I just don't think this is a good move in general. 348 00:20:08,020 --> 00:20:12,420 Josh: This topic actually makes me pretty sad as an Apple fanboy, not because of what 349 00:20:12,420 --> 00:20:16,140 Josh: they're doing to the developers, which is messed up. 350 00:20:16,140 --> 00:20:19,180 Josh: And i don't like but the idea that apple is going 351 00:20:19,180 --> 00:20:21,920 Josh: to exist as the king as it stands now 352 00:20:21,920 --> 00:20:24,980 Josh: is just an impossibility in this new world when you 353 00:20:24,980 --> 00:20:27,680 Josh: think about how easy it is to generate one of these apps and how easy it is 354 00:20:27,680 --> 00:20:30,560 Josh: just to sideload them onto your phone like anyone with a 355 00:20:30,560 --> 00:20:34,840 Josh: test flight account can vibe code an app of whatever you've ever wanted an iphone 356 00:20:34,840 --> 00:20:38,660 Josh: app to look like you can send it direct to your iphone you can't put on the 357 00:20:38,660 --> 00:20:41,380 Josh: app store but you can send it directly to your iphone and you could have your 358 00:20:41,380 --> 00:20:44,240 Josh: own version of your dream app and it can do whatever functionality you want 359 00:20:44,240 --> 00:20:48,000 Josh: and you're one prompt away from updating it and editing it and changing it to do whatever you want. 360 00:20:48,220 --> 00:20:51,500 Josh: And the fact that it's so accessible now, and Apple is... 361 00:20:51,660 --> 00:20:55,300 Josh: Clearly not leaning into adopting this new paradigm shift, there's this clash 362 00:20:55,300 --> 00:20:57,800 Josh: that's happening. And we're seeing it here with this vibe coded apps. 363 00:20:57,940 --> 00:20:59,120 Josh: They're not the good guy anymore. 364 00:20:59,300 --> 00:21:02,700 Josh: They're not the person who is helping developers do what they want, 365 00:21:02,860 --> 00:21:05,360 Josh: empowering them for this new paradigm of engineering. 366 00:21:05,580 --> 00:21:10,540 Josh: They are the bottleneck. They are the like hammer who is stopping people from innovating. 367 00:21:10,780 --> 00:21:14,040 Josh: And that to me makes me really sad. And that has always throughout history been 368 00:21:14,040 --> 00:21:16,480 Josh: a losing formula. So I hope they turn it around. 369 00:21:16,660 --> 00:21:21,060 Josh: We just got a WWDC announcement, which is happening in the next month, 370 00:21:21,240 --> 00:21:22,080 Josh: I think. I forgot the dates. 371 00:21:22,200 --> 00:21:25,840 Josh: But that's going to be their time where they're going to showcase all of the new AI stuff. 372 00:21:26,200 --> 00:21:28,840 Josh: That's when they initially announced Apple Intelligence. It was the biggest 373 00:21:28,840 --> 00:21:30,820 Josh: flop of all time. They're going to try to do it again. 374 00:21:31,200 --> 00:21:33,380 Josh: Hopefully, they'll come out with some new policies to address this. 375 00:21:33,600 --> 00:21:35,360 Josh: And hopefully, they'll have a chance to turn the ship around. 376 00:21:35,460 --> 00:21:36,780 Josh: But that is their last chance. 377 00:21:37,320 --> 00:21:40,180 Ejaaz: So something tells me, Josh, correct me if you think I'm wrong, 378 00:21:40,460 --> 00:21:45,000 Ejaaz: that this may not be an outlandish move by Apple. 379 00:21:45,160 --> 00:21:49,440 Ejaaz: It might be a strategic one because we also got some other news that broke this 380 00:21:49,440 --> 00:21:51,460 Ejaaz: week, that their deal with Google, 381 00:21:51,680 --> 00:21:56,820 Ejaaz: the deal with Google, where they pay Google $1 billion and get access to Google's 382 00:21:56,820 --> 00:22:01,000 Ejaaz: Gemini AI models, what we originally thought was going to be some kind of licensing, right, 383 00:22:01,520 --> 00:22:06,600 Ejaaz: actually is Apple getting full access to Google's Gemini model weights, 384 00:22:06,800 --> 00:22:10,840 Ejaaz: which means that they can fine tune and build the model in any way that they 385 00:22:10,840 --> 00:22:14,600 Ejaaz: want, they get full access to a model that Google spent hundreds of millions 386 00:22:14,600 --> 00:22:17,640 Ejaaz: of dollars training years training just for a billion dollars a year. 387 00:22:17,780 --> 00:22:20,400 Ejaaz: I don't know what's going on here. And it just means that Apple's got an absolute 388 00:22:20,400 --> 00:22:24,620 Ejaaz: steal of a model, a Frontier LLM from Google, and they can now run with it and 389 00:22:24,620 --> 00:22:27,880 Ejaaz: build their own AI apps, which is presumably what I think they're going to do. 390 00:22:28,080 --> 00:22:31,580 Josh: This is a huge win for Apple. And we have to ask the question, 391 00:22:31,740 --> 00:22:34,680 Josh: why is Google doing this? And I 392 00:22:34,850 --> 00:22:38,610 Josh: I think it kind of pairs to that first topic that we spoke about, 393 00:22:38,950 --> 00:22:43,230 Josh: which was that research paper where they're kind of democratizing intelligence 394 00:22:43,230 --> 00:22:44,590 Josh: as best they can with TurboQuant, right? 395 00:22:44,650 --> 00:22:47,930 Josh: It's like they're increasing the efficiency, they have their own custom TPUs 396 00:22:47,930 --> 00:22:49,070 Josh: to make the price per token down. 397 00:22:49,190 --> 00:22:53,590 Josh: And when we consider what the most existential threat in the world of AI is, 398 00:22:53,910 --> 00:22:57,830 Josh: it's that edge inference and local inference gets really good so that you don't 399 00:22:57,830 --> 00:23:00,370 Josh: need to generate tokens from these megacloud providers. 400 00:23:00,370 --> 00:23:05,610 Josh: And I think the reality is that it's becoming more and more true because we 401 00:23:05,610 --> 00:23:09,810 Josh: have now these turbo quant models that are going to be six times more efficient. 402 00:23:10,090 --> 00:23:14,030 Josh: We have Apple who is taking Gemini models which are leading edge and distilling 403 00:23:14,030 --> 00:23:17,650 Josh: them down into something that runs on your iPhone and the majority of the people 404 00:23:17,650 --> 00:23:19,790 Josh: don't need world-class intelligence. 405 00:23:19,990 --> 00:23:23,470 Josh: They don't need the bleeding edge stuff that is solving new physics and math. 406 00:23:23,590 --> 00:23:27,090 Josh: They just want it to solve the day-to-day stuff and this 407 00:23:27,090 --> 00:23:30,330 Josh: is possibly a hedge against that for Google is they're 408 00:23:30,330 --> 00:23:33,310 Josh: creating the problem by throwing out papers like turbo quant 409 00:23:33,310 --> 00:23:36,410 Josh: but then they also have access to the solution which is apple i 410 00:23:36,410 --> 00:23:39,450 Josh: mean when we think about where the most local inference is held it's just 411 00:23:39,450 --> 00:23:43,410 Josh: on the entire apple network on your laptops on your iphones you can run these 412 00:23:43,410 --> 00:23:48,170 Josh: unbelievable models and now apple has the ability to do that with gemini and 413 00:23:48,170 --> 00:23:51,070 Josh: i think it's strategic it makes sense they're collecting a paycheck they have 414 00:23:51,070 --> 00:23:53,650 Josh: a great relationship with apple they've been doing it with the browser forever 415 00:23:53,650 --> 00:23:56,470 Josh: and this is just a natural extension of that apple. 416 00:23:56,470 --> 00:23:57,490 Ejaaz: Has all the distribution. 417 00:23:58,100 --> 00:24:03,700 Ejaaz: They have 2.5 to 3 billion live Apple devices active in the world right now. 418 00:24:03,900 --> 00:24:07,340 Josh: The Mag 7 stocks too. When you look at the Mag 7 stocks who've been like suffering 419 00:24:07,340 --> 00:24:09,160 Josh: the most, Apple suffered the least. 420 00:24:09,320 --> 00:24:12,720 Josh: It's just out of the race. It's right there unaffected by all of it. 421 00:24:12,980 --> 00:24:18,300 Ejaaz: Yeah, they stayed number two market cap in the entire world for a company without 422 00:24:18,300 --> 00:24:22,600 Ejaaz: touching a single GPU expenditure, without a single AI CapEx expenditure. 423 00:24:22,800 --> 00:24:24,520 Ejaaz: Just insane strategy from them. 424 00:24:24,860 --> 00:24:28,820 Ejaaz: They hold a very rich and important moat. But looking at Google as well, 425 00:24:28,920 --> 00:24:33,720 Ejaaz: they're smart because they know that Apple won't win any of this AI stuff unless 426 00:24:33,720 --> 00:24:36,540 Ejaaz: they get access to a major frontier LLM. 427 00:24:36,660 --> 00:24:42,040 Ejaaz: Google's bet here is compute plus data equals the best model in the world. 428 00:24:42,420 --> 00:24:45,740 Ejaaz: Apple does not have the compute. They do not have the model. 429 00:24:45,940 --> 00:24:48,840 Ejaaz: Google can supply that for them. And something tells me that whereas Google 430 00:24:48,840 --> 00:24:52,460 Ejaaz: might not be getting paid as much by Apple, they're going to exchange or broker 431 00:24:52,460 --> 00:24:55,780 Ejaaz: a deal where they you get access to some of Apple's user data, 432 00:24:55,820 --> 00:24:59,500 Ejaaz: which they can then use to train a better model, and it becomes this synonymous thing. 433 00:25:00,100 --> 00:25:05,280 Ejaaz: But now, if the future is actually locally run edge compute models that run 434 00:25:05,280 --> 00:25:08,920 Ejaaz: on your laptop that are frontier, then Google's shot themselves in the foot, 435 00:25:09,080 --> 00:25:14,100 Ejaaz: maybe shot themselves in both feet at this point. But it's a bet they're making. 436 00:25:14,280 --> 00:25:17,460 Ejaaz: And I think, I don't know, I think Google might have this one, if I'm being honest. 437 00:25:17,740 --> 00:25:21,080 Josh: Yeah, I'm rooting for them. I'm also rooting for Meta now, who has this like 438 00:25:21,080 --> 00:25:24,040 Josh: unbelievable pay package that I just saw. 439 00:25:24,900 --> 00:25:29,560 Josh: $9 trillion is the valuation they need to exercise all these pay packages. 440 00:25:29,560 --> 00:25:32,500 Josh: Like, okay, Elon, okay, Tesla board. I've seen this before. 441 00:25:33,560 --> 00:25:36,700 Josh: What's the deal with this milestone package here? 442 00:25:36,780 --> 00:25:41,960 Josh: Because a $9 trillion valuation for Meta, that seems like a lot of money. 443 00:25:42,620 --> 00:25:49,600 Ejaaz: Okay. So two things have happened over the last week, which are at extreme odds at each other in Meta. 444 00:25:49,820 --> 00:25:54,020 Ejaaz: Number one, they are closing down their metaverse division and rumored to be 445 00:25:54,020 --> 00:25:59,460 Ejaaz: laying off up to 20% of the company. Now, these employees range from low-level 446 00:25:59,460 --> 00:26:03,020 Ejaaz: product employees, engineering employees, all the way up to senior director roles. 447 00:26:03,180 --> 00:26:07,860 Ejaaz: But this week, news also broke that they are going to be rewarding or they've 448 00:26:07,860 --> 00:26:13,440 Ejaaz: offered very attractive comp packages to their top executives to the tune of 449 00:26:13,440 --> 00:26:17,820 Ejaaz: $800 million for one person, their CFO. 450 00:26:17,820 --> 00:26:22,600 Ejaaz: But it's under one premise and condition, which is over the next five years, 451 00:26:23,020 --> 00:26:25,380 Ejaaz: Meta needs to hit a $9 trillion. 452 00:26:25,840 --> 00:26:31,060 Ejaaz: That's with a T, dollar market cap. Just for context, no company in the history 453 00:26:31,060 --> 00:26:32,600 Ejaaz: of the world has ever hit that market cap. 454 00:26:32,860 --> 00:26:35,200 Ejaaz: In fact, the biggest and largest and most important company in the world right 455 00:26:35,200 --> 00:26:37,460 Ejaaz: now, NVIDIA, is currently sitting at $4.5 trillion. 456 00:26:37,920 --> 00:26:40,940 Ejaaz: So we're talking about a 2x for that. And for Meta, who is currently sitting 457 00:26:40,940 --> 00:26:44,420 Ejaaz: in, I don't know where, are they still in the Mag 7 at this point? 458 00:26:45,080 --> 00:26:49,360 Ejaaz: If so, I think they're like six or seven. They've got a long way to go, 459 00:26:49,520 --> 00:26:55,160 Ejaaz: but it's interesting seeing the incentive design that Zuck is setting up. He spent 460 00:26:55,640 --> 00:27:01,860 Ejaaz: I think upwards of $30 billion hiring 200 people this year and hasn't released a single AI model. 461 00:27:02,360 --> 00:27:06,180 Ejaaz: Llama, their open source model is dead in the water. All the apps that they've 462 00:27:06,180 --> 00:27:07,620 Ejaaz: released have been crap. 463 00:27:07,920 --> 00:27:12,420 Ejaaz: So Meta is kind of making a big bet here. And I don't know if they're incentivizing 464 00:27:12,420 --> 00:27:13,760 Ejaaz: the right people, if I'm being honest. 465 00:27:14,260 --> 00:27:17,200 Josh: I'm rooting for them. I hope it works. I mean, in five years, 466 00:27:17,280 --> 00:27:19,960 Josh: there will be a $9 trillion company, possibly multiple of them. 467 00:27:20,140 --> 00:27:24,700 Josh: So there is a world in which Meta can achieve this. The problem is that they 468 00:27:24,700 --> 00:27:27,360 Josh: haven't actually done anything to make progress towards that recently. 469 00:27:27,720 --> 00:27:29,900 Josh: Like they've spent all this money, they've hired all this talent, 470 00:27:30,060 --> 00:27:31,440 Josh: they haven't released anything compelling. 471 00:27:31,800 --> 00:27:35,200 Josh: They haven't actually proven that the money that they're spending is working. 472 00:27:35,300 --> 00:27:39,420 Josh: In fact, the counter argument to that is true, where Meta, the company, 473 00:27:39,820 --> 00:27:43,500 Josh: just crushed the division that was responsible for renaming the company Meta. 474 00:27:43,700 --> 00:27:47,140 Josh: So clearly there has been a series of big swings that haven't worked, 475 00:27:47,300 --> 00:27:51,000 Josh: and we've yet to actually see a big swing that has worked, everything's failed. 476 00:27:51,220 --> 00:27:54,520 Josh: Like you said, the open source models have failed. The pivot to the metaverse has failed. 477 00:27:54,820 --> 00:27:57,780 Josh: Everything besides the core product of Facebook, which is their, 478 00:27:58,630 --> 00:28:03,410 Josh: algorithm and their home feed and their social media platforms hasn't worked 479 00:28:03,410 --> 00:28:04,930 Josh: out. Even the hardware sucks. 480 00:28:05,090 --> 00:28:08,330 Josh: So they have a lot to figure out if they want to make this work. 481 00:28:08,450 --> 00:28:09,230 Josh: But I'm rooting for them. 482 00:28:09,650 --> 00:28:12,950 Josh: Zuck's an amazing CEO. I am like very hopeful that he can figure this out. 483 00:28:13,330 --> 00:28:16,750 Josh: And this incentive structure is the right way to do it. You want to incentivize 484 00:28:16,750 --> 00:28:19,550 Josh: people to bring the value. I hope they can bring the value. 485 00:28:19,730 --> 00:28:25,290 Ejaaz: Now, for the last story of this episode, I want to introduce the listeners and 486 00:28:25,290 --> 00:28:30,170 Ejaaz: watchers of this show to a secret, a secret within Limitless that none of you 487 00:28:30,170 --> 00:28:32,730 Ejaaz: have ever known or has been publicized before. 488 00:28:33,090 --> 00:28:38,670 Ejaaz: And this secret is about our co-host, Josh. Josh is a part-time music producer, 489 00:28:38,670 --> 00:28:42,910 Ejaaz: and he releases absolutely banging tunes for the world to hear. 490 00:28:43,250 --> 00:28:46,870 Josh: Yes. And this part-time skill actually originated about two weeks ago with the 491 00:28:46,870 --> 00:28:48,610 Josh: advent of Google Lear A3. 492 00:28:48,830 --> 00:28:51,950 Josh: And the fact that now I can make music with a single prompt, 493 00:28:52,110 --> 00:28:54,690 Josh: maybe a sentence or two. So that's what I did today. 494 00:28:55,250 --> 00:28:58,710 Josh: I said, generate me a three minute song about the Limitless podcast and how 495 00:28:58,710 --> 00:29:02,030 Josh: it's the absolute best number one podcast in the world. There's nobody better. 496 00:29:02,530 --> 00:29:05,710 Josh: The AI is scared of our show because it's so good. This just sounded like an 497 00:29:05,710 --> 00:29:09,790 Josh: R&B hip hop type vibe. And we've spoken about Lyria 3 before of the show. 498 00:29:10,210 --> 00:29:13,690 Josh: Lyria 3 is the music generation software from Google where you can prompt it, 499 00:29:14,030 --> 00:29:18,310 Josh: you say what you want the song to be about, you say if you want lyrics or not, 500 00:29:18,470 --> 00:29:22,610 Josh: what type of lyrics you want to sound, or you just let it run wild and choose everything for you. 501 00:29:22,610 --> 00:29:25,430 Josh: So that was all the instruction i gave and it generated me 502 00:29:25,430 --> 00:29:28,670 Josh: a song named no logic for the soul 503 00:29:28,670 --> 00:29:33,190 Josh: which is very elegant i love i love the artwork it has like this cloud and these 504 00:29:33,190 --> 00:29:36,610 Josh: roots and it's it's very good but the song is good so i want to play you a little 505 00:29:36,610 --> 00:29:41,110 Josh: clip here of this song i'm not sure if you've had a chance to really enjoy the 506 00:29:41,110 --> 00:29:43,690 Josh: full thing neither of us have so we're going to experience it for the first 507 00:29:43,690 --> 00:29:47,390 Josh: time together but this is our new original no logic for the soul enjoy okay. 508 00:30:30,000 --> 00:30:32,520 Ejaaz: That has to be a new outro song, 509 00:30:32,780 --> 00:30:36,240 Josh: Josh. That is a banger. It's unbelievable how good this music is. So good. 510 00:30:36,880 --> 00:30:41,020 Josh: The lyrics, not only do they make sense, but they rhyme. They have the right timing and cadence. 511 00:30:41,360 --> 00:30:44,380 Josh: The chorus is great. They had a full horn section in there. 512 00:30:44,700 --> 00:30:49,620 Josh: I mean, it's amazing, right? Because you think about the previous generation 513 00:30:49,620 --> 00:30:50,600 Josh: and how they enjoyed music. 514 00:30:50,800 --> 00:30:54,520 Josh: We had the Rolling Stones, the Beatles, the Grateful Dead. 515 00:30:54,980 --> 00:30:58,540 Josh: These were generation-defining artists and musicians. 516 00:30:58,540 --> 00:31:01,620 Josh: And the reason they were so powerful is because there was convergence around 517 00:31:01,620 --> 00:31:04,400 Josh: them you go on the radio it's playing their music you talk 518 00:31:04,400 --> 00:31:07,640 Josh: to your friends they have their cds and their vinyl records you go 519 00:31:07,640 --> 00:31:10,880 Josh: on the street this is what's being played because this is what's available and 520 00:31:10,880 --> 00:31:14,440 Josh: it was great but it was really the best that was available and everyone could 521 00:31:14,440 --> 00:31:18,060 Josh: kind of converge on that fact the idea that you can now generate good music 522 00:31:18,060 --> 00:31:21,680 Josh: not great music but good music and good enough for people who don't really care 523 00:31:21,680 --> 00:31:27,180 Josh: to seek out great music implies the fact that we might never have anything like that again. 524 00:31:27,420 --> 00:31:30,700 Josh: There might never be another Grateful Dead, another Rolling Stones, 525 00:31:30,820 --> 00:31:34,240 Josh: another generational defining artist because it's so accessible. 526 00:31:35,010 --> 00:31:39,050 Josh: If you look at the top charts on Spotify, it's the top trending tracks on TikTok. 527 00:31:39,370 --> 00:31:42,810 Josh: And there's this direct correlation between the two. And it's stripped out a 528 00:31:42,810 --> 00:31:47,850 Josh: lot of the humanness of artistry. It creates a whole different paradigm. 529 00:31:48,150 --> 00:31:52,930 Josh: And listening to this really hits hard. I'm like, okay, yeah, it's actually over. 530 00:31:53,110 --> 00:31:57,090 Josh: The fact that I can do this with a two-sentence prompt. I did this in 30 seconds. 531 00:31:57,210 --> 00:31:59,030 Ejaaz: How long is the track, Josh, as well? 532 00:31:59,150 --> 00:32:01,690 Josh: And the track is three minutes long. It's two minutes and 57 seconds. 533 00:32:01,690 --> 00:32:06,210 Ejaaz: That's a regular song. That's a song that could go viral and like take over the Spotify charts. 534 00:32:06,430 --> 00:32:09,190 Josh: And like if we did some prompt engineering, if we really refined the prompt 535 00:32:09,190 --> 00:32:12,550 Josh: from more than two sentences to like a proper setup where there's many, 536 00:32:12,610 --> 00:32:15,870 Josh: many paragraphs, many, many details, you kind of outline the lyrics and you 537 00:32:15,870 --> 00:32:17,030 Josh: outline the cadence and what you want. 538 00:32:17,210 --> 00:32:21,890 Josh: You can get unbelievably detailed with this and it sounds good. That's a good song. 539 00:32:22,290 --> 00:32:25,710 Josh: So it's tough. I mean, this is one of those bittersweet moments where you're 540 00:32:25,710 --> 00:32:29,310 Josh: like, holy shit, this is amazing, but holy shit, this is pretty powerful stuff. 541 00:32:29,310 --> 00:32:32,590 Ejaaz: Now, I know there are a bunch of audiophiles listening to this thinking, 542 00:32:32,830 --> 00:32:34,550 Ejaaz: God, Josh and Ijaz have no taste. 543 00:32:34,810 --> 00:32:38,590 Ejaaz: May I just remind you, this is the worst this model will ever be. 544 00:32:38,790 --> 00:32:42,590 Ejaaz: It's the worst. It's only going to get better. And there's already been a bunch 545 00:32:42,590 --> 00:32:45,310 Ejaaz: of AI soundtracks that have gone completely viral. 546 00:32:45,490 --> 00:32:49,050 Ejaaz: In fact, last week, Spotify had to shut down an AI music band, 547 00:32:49,150 --> 00:32:52,930 Ejaaz: which they didn't realize was an AI music band, because they just didn't like that it was AI. 548 00:32:53,090 --> 00:32:56,590 Ejaaz: But it was absolutely plugging through a bunch of streams. I think it gathered 549 00:32:56,590 --> 00:32:59,110 Ejaaz: like 300,000 streams over the last month. 550 00:32:59,310 --> 00:33:03,310 Ejaaz: Just insane. So these things aren't going away. Another thing I wanna point out is 551 00:33:03,810 --> 00:33:09,910 Ejaaz: OpenAI just sunsetted Sora, which was the video generation of AI video generation version of TikTok. 552 00:33:10,350 --> 00:33:13,930 Ejaaz: And now we have Google Liria, which is kind of doing a similar thing for music. 553 00:33:13,950 --> 00:33:18,790 Ejaaz: So I don't think artistry and that type of medium is going away. 554 00:33:18,950 --> 00:33:20,590 Ejaaz: There is an insatiable amount of demand. 555 00:33:20,790 --> 00:33:26,430 Ejaaz: It helps music producers themselves kind of express the type of sound or song 556 00:33:26,430 --> 00:33:29,590 Ejaaz: that they have in their head. So I think as a medium of translation, 557 00:33:29,830 --> 00:33:31,090 Ejaaz: these AI tools are really valuable. 558 00:33:31,970 --> 00:33:35,270 Ejaaz: But to your point, Josh, we are going to end up in a world where maybe everything 559 00:33:35,270 --> 00:33:38,530 Ejaaz: sounds kind of beige and consistent. 560 00:33:38,830 --> 00:33:42,550 Ejaaz: But I'm going to take the optimistic side of that, which is I think it's going 561 00:33:42,550 --> 00:33:46,710 Ejaaz: to force humans and AI alike to be even more creative and come up with better 562 00:33:46,710 --> 00:33:47,850 Ejaaz: soundtracks. I'm excited. 563 00:33:48,170 --> 00:33:52,010 Ejaaz: And to your point around the Grateful Dead, you can now, I don't know, 564 00:33:52,150 --> 00:33:56,290 Ejaaz: technically license their IP and create a never-ending album of their music. 565 00:33:56,290 --> 00:33:57,970 Ejaaz: So maybe that's an angle as well. 566 00:33:58,470 --> 00:34:03,510 Josh: Deadheads are not liking this at all. Oh man, this is going to disturb a lot of people. 567 00:34:04,490 --> 00:34:05,530 Ejaaz: But it's here. 568 00:34:05,690 --> 00:34:09,150 Josh: And like you said, it's the worst it will ever be. It's up only from here. 569 00:34:09,250 --> 00:34:10,150 Josh: The quality will get better. 570 00:34:10,290 --> 00:34:14,970 Josh: Of all of these tools, as Meta goes to $9 trillion, as Google disrupts entire 571 00:34:14,970 --> 00:34:17,770 Josh: industries, it's all happening so fast. 572 00:34:17,950 --> 00:34:20,830 Josh: That is another week fully covered in the books. If you've made it this far 573 00:34:20,830 --> 00:34:24,210 Josh: in this episode and you've listened to the three previous episodes from this 574 00:34:24,210 --> 00:34:25,690 Josh: week, you are now fully caught up. 575 00:34:25,690 --> 00:34:28,630 Josh: There's nothing you need to know that you don't already know you can 576 00:34:28,630 --> 00:34:31,930 Josh: go enjoy your weekend go touch some grass go hang out with a friend with um 577 00:34:31,930 --> 00:34:34,630 Josh: do some analog stuff maybe that's kind of cool because i'm sure 578 00:34:34,630 --> 00:34:37,270 Josh: everyone's just been in the trenches on um on their 579 00:34:37,270 --> 00:34:40,090 Josh: devices watching the war zone all week but 580 00:34:40,090 --> 00:34:43,650 Josh: yeah thank you guys so much for watching as always if you enjoy this episode 581 00:34:43,650 --> 00:34:47,310 Josh: don't forget to share it with your friends who might also enjoy it uh subscribe 582 00:34:47,310 --> 00:34:50,430 Josh: to our newsletter which is awesome it comes out twice a week we publish on x 583 00:34:50,430 --> 00:34:53,690 Josh: all the time pretty prolifically now we're getting lots of action here's our 584 00:34:53,690 --> 00:34:56,770 Josh: handles that you can find on screen right now Each has any final party thoughts 585 00:34:56,770 --> 00:34:57,830 Josh: before we head off for the weekend? 586 00:34:58,440 --> 00:35:03,980 Ejaaz: Um yeah i want a few of you to generate uh both josh and i individually and 587 00:35:03,980 --> 00:35:07,620 Ejaaz: together our own songs and jingle let's see if you can come up with the best 588 00:35:07,620 --> 00:35:10,120 Ejaaz: limitless intro outro song and maybe we'll feature it on a 589 00:35:10,120 --> 00:35:13,200 Josh: Future that would actually be cool and send it to us on on x 590 00:35:13,200 --> 00:35:15,820 Josh: because you can embed it in the post and we could just save it 591 00:35:15,820 --> 00:35:18,560 Josh: from there so if you generate something tag you guys and 592 00:35:18,560 --> 00:35:22,280 Josh: i on x and we will listen to it and maybe find a new intro outro track that 593 00:35:22,280 --> 00:35:25,540 Josh: would be kind of that would be or a theme song i love a theme song we'd be like 594 00:35:25,540 --> 00:35:28,740 Josh: superheroes we could have a theme song before we get on on camera that'd be 595 00:35:28,740 --> 00:35:31,600 Josh: sick well anyways i know this was a long one so if you made it through i mean 596 00:35:31,600 --> 00:35:35,780 Josh: 40 minutes of this show thank you you're a real one much appreciated as always 597 00:35:35,780 --> 00:35:37,600 Josh: have an amazing weekend and we'll see you guys soon.