1 00:00:00,640 --> 00:00:03,800 Hi, I'm Eric Newcomer. And I am Madeline Rinbarger. 2 00:00:04,200 --> 00:00:07,040 And this is the newcomer podcast. 3 00:00:10,400 --> 00:00:14,080 Each week, Eric and I discuss the VC deals and the drama that 4 00:00:14,080 --> 00:00:15,800 went down. Let's do it. 5 00:00:16,079 --> 00:00:17,680 Here we go. All right. 6 00:00:18,040 --> 00:00:22,560 Let's dive in, jumping into the deals that defined last week. 7 00:00:22,720 --> 00:00:27,480 AI legend Fei Fei Lee, who's been the huge thought leader in 8 00:00:27,600 --> 00:00:30,120 AI, is doing her own company. Why wouldn't you? 9 00:00:30,120 --> 00:00:34,720 She raised $230 million to build large world models to interact 10 00:00:34,720 --> 00:00:38,920 with 3D space, came out of stealth and raised from a who's 11 00:00:38,920 --> 00:00:42,200 who of venture capitalists Andreessen Horowitz, NEA, 12 00:00:42,200 --> 00:00:45,240 Radical Ventures and includes, you know, big names like Marc 13 00:00:45,240 --> 00:00:48,080 Benioff, Jim Breyer, Ron Conway, Jeff Dean. 14 00:00:48,080 --> 00:00:49,200 Yeah, I'm not going to read them all. 15 00:00:49,200 --> 00:00:51,760 Lots of them. Fundamentally, this is what, 16 00:00:51,760 --> 00:00:54,760 robots plus AI, right, Maddie? I mean, this is a world you've 17 00:00:54,760 --> 00:00:56,480 written a ton about. Right. 18 00:00:56,480 --> 00:00:57,920 Yeah. I mean this is a really booming 19 00:00:57,920 --> 00:01:01,800 space within as a sub sector within the AI funding rounds 20 00:01:01,800 --> 00:01:05,840 that we've seen robotics and foundation models that can help 21 00:01:05,840 --> 00:01:09,480 translate into movement and physical space and understand 22 00:01:09,480 --> 00:01:11,600 the world beyond just self driving cars. 23 00:01:11,600 --> 00:01:14,840 It's something that's been AI has really provided a big 24 00:01:14,840 --> 00:01:17,400 breakthrough for. So it seems like there's lots of 25 00:01:17,400 --> 00:01:19,520 different companies playing in this space. 26 00:01:19,600 --> 00:01:22,480 I will say there's, you know, physical intelligence. 27 00:01:22,480 --> 00:01:26,120 Locky Grooms company alongside some other cofounders raised a 28 00:01:26,160 --> 00:01:29,200 sizable round earlier this year. And then there's several other 29 00:01:29,200 --> 00:01:33,600 smaller ones that are doing more kind of specified in factory 30 00:01:33,600 --> 00:01:36,760 robotics versus humanoid robotics companies that we're 31 00:01:36,760 --> 00:01:38,600 also seeing. And there is kind of a debate 32 00:01:38,600 --> 00:01:42,080 around whether humanoid is the way to go or we should build 33 00:01:42,320 --> 00:01:46,040 different kinds of robots to handle these models since robots 34 00:01:46,040 --> 00:01:48,160 are not humans. So maybe we should, you know, 35 00:01:48,200 --> 00:01:50,600 build things for their specific tasks rather than try to 36 00:01:50,600 --> 00:01:53,240 replicate a human assistant. Although the human assistance, 37 00:01:53,240 --> 00:01:55,600 you know, fulfill a sci-fi fantasy for some people. 38 00:01:55,600 --> 00:01:57,680 So that's cool. It's a huge round, lots of 39 00:01:57,680 --> 00:02:00,080 investors. All right, that was Fei Fei 40 00:02:00,080 --> 00:02:04,960 Lee's world Labs for our second notable deal defining this week. 41 00:02:05,120 --> 00:02:07,960 I think you're going to take us back to my old neck of the 42 00:02:07,960 --> 00:02:11,520 woods. I I use once covered delivery 43 00:02:11,520 --> 00:02:13,920 very fastidiously. Anyway, what's what's the? 44 00:02:13,920 --> 00:02:18,320 Deal Yes, we're jumping back into the heyday of delivery 45 00:02:18,320 --> 00:02:21,080 companies. If you remember in 2021, there 46 00:02:21,080 --> 00:02:24,160 were quite a few of these rapid delivery startups, one of which 47 00:02:24,160 --> 00:02:29,120 was Flank, a German company in the space, and they raised 150 48 00:02:29,120 --> 00:02:31,840 million in new funding from investors. 49 00:02:31,920 --> 00:02:35,840 Bond Mubadala, a Scandinavian supermarket giant, all jumping 50 00:02:35,840 --> 00:02:38,800 into this round. And I think what's most notable 51 00:02:38,800 --> 00:02:41,120 to me is how far the valuation has fallen. 52 00:02:41,120 --> 00:02:44,720 This one places it at just under a billion, but this was a 53 00:02:44,720 --> 00:02:48,560 company that was valued at 5 billion in 2021. 54 00:02:48,960 --> 00:02:50,960 And there have been a lot of competitors that have been 55 00:02:50,960 --> 00:02:54,320 acquired or just fallen out of the space, especially as you 56 00:02:54,320 --> 00:02:58,160 know, rapid delivery costs seem to be it's it's a tougher 57 00:02:58,160 --> 00:03:03,840 business than yes, this is go puff getter that space ultra 58 00:03:03,840 --> 00:03:07,240 fast delivery where they stock warehouses with specific items 59 00:03:07,240 --> 00:03:10,000 and then drive drivers get it to you as quickly as possible. 60 00:03:10,000 --> 00:03:13,640 And it's one of those things that just seems much tougher in 61 00:03:13,640 --> 00:03:15,440 execution when it comes down to it. 62 00:03:15,440 --> 00:03:19,800 But obviously Flink has held on and has still running a 63 00:03:19,960 --> 00:03:21,400 business. But it just seems like the 64 00:03:21,400 --> 00:03:24,720 margins don't look as frothy as they once did, you know, back in 65 00:03:24,720 --> 00:03:26,320 the day. It's amazing how much of the 66 00:03:26,320 --> 00:03:29,120 narrative just forgets about these companies. 67 00:03:29,120 --> 00:03:32,400 You know, I, I feel like they never got to the level of Uber. 68 00:03:32,480 --> 00:03:35,520 I mean, it's sort of like Lime in the scooter businesses where 69 00:03:35,800 --> 00:03:39,160 they seemed crazy. I mean, I, I would say 1 70 00:03:39,160 --> 00:03:42,720 similarity between the scooter businesses and these is they 71 00:03:42,800 --> 00:03:46,120 require actual spending. They're not like Uber was 72 00:03:46,120 --> 00:03:49,760 fundamentally a marketplace with sure, a lot of subsidies and 73 00:03:49,760 --> 00:03:52,480 other costs. It burned through billions, mini 74 00:03:52,480 --> 00:03:57,200 billions, 20 billion plus. But it's still like you could 75 00:03:57,200 --> 00:04:00,120 see how it could cut them all away and have a marketplace at 76 00:04:00,120 --> 00:04:01,960 the end of the day. Whereas if you have to build 77 00:04:01,960 --> 00:04:04,640 warehouses or you have to buy scooters, that's a tough 78 00:04:04,640 --> 00:04:06,920 business. And you know, we don't even know 79 00:04:07,360 --> 00:04:10,240 the terms of this round. It's like under a billion 80 00:04:10,240 --> 00:04:13,800 raising 150 million. I imagine that's it's a tough 81 00:04:13,800 --> 00:04:16,040 round. It'll be interesting to see if 82 00:04:16,040 --> 00:04:18,120 there are any soft landings in this space at all. 83 00:04:18,560 --> 00:04:21,320 It's a tough round for sure. And I do feel like there's been 84 00:04:21,320 --> 00:04:24,320 kind of entrance outside of the US that have gained more 85 00:04:24,320 --> 00:04:25,880 traction. Several of these companies are 86 00:04:25,880 --> 00:04:28,480 based in India as well, where they've seemed to have more 87 00:04:28,480 --> 00:04:30,560 success. So it could be just that it's 88 00:04:30,560 --> 00:04:33,880 not a market for US consumers. And if you take a more global 89 00:04:33,880 --> 00:04:35,440 approach, obviously Flink is from Germany. 90 00:04:35,440 --> 00:04:38,120 So they could have, you know, some staying power in a market 91 00:04:38,120 --> 00:04:41,560 outside of the US and that could be enough to keep it going. 92 00:04:41,560 --> 00:04:45,200 But it certainly is a fall from the the zerp arrow highs. 93 00:04:45,800 --> 00:04:47,960 I'm going to give one sense honorable mention, which is of 94 00:04:47,960 --> 00:04:52,600 course the open AI deal. This $6.5 billion funding round 95 00:04:52,600 --> 00:04:55,000 is coming together. I think the Information reported 96 00:04:55,000 --> 00:04:58,080 the other day that investors had a minimum 250 million 97 00:04:58,080 --> 00:05:00,720 investment. It's going to be bigger than 98 00:05:00,720 --> 00:05:03,960 like all the funding in the startup world, you know, not so 99 00:05:03,960 --> 00:05:07,880 long ago. Anyway, so he's open AIA startup 100 00:05:07,880 --> 00:05:11,040 in the traditional sense. I'm sure people debate, but 101 00:05:11,040 --> 00:05:14,080 anyway, it's it's an epic round and so we'll keep watching that 102 00:05:14,080 --> 00:05:16,200 one. Jumping into our top story this 103 00:05:16,200 --> 00:05:19,960 week, Y Combinators Demo Day is taking place Wednesday and 104 00:05:19,960 --> 00:05:22,560 Thursday. It's the, you know, semi annual 105 00:05:22,560 --> 00:05:24,680 event that's about to happen more times a year. 106 00:05:24,680 --> 00:05:28,720 We get four of them a year now. Yay, so many more zooms to 107 00:05:28,720 --> 00:05:31,160 watch. You really captured what's 108 00:05:31,160 --> 00:05:34,720 happening in YC Demo Day this week, so I won't steal all your 109 00:05:34,720 --> 00:05:39,240 punches, but obviously it remains artificial intelligence 110 00:05:39,240 --> 00:05:42,760 driven, lots of boring enterprise software, but you 111 00:05:43,000 --> 00:05:46,000 tried to drill into some of the areas where there's some new 112 00:05:46,000 --> 00:05:46,640 stuff. What? 113 00:05:46,640 --> 00:05:49,560 What are you seeing with the YC Demo Day right now? 114 00:05:49,760 --> 00:05:51,480 Yeah. So this was the first Demo Day 115 00:05:51,480 --> 00:05:54,600 after they released their latest call to startups that mentioned 116 00:05:54,600 --> 00:05:57,800 defense tech as an area or space tech an area that they were 117 00:05:57,800 --> 00:06:00,960 interested in missiles. I think I mean it's put missiles 118 00:06:00,960 --> 00:06:03,000 in the headline. We had to put missiles in the 119 00:06:03,000 --> 00:06:05,280 headline. It is Y CS first weapons company 120 00:06:05,280 --> 00:06:06,560 period. Not just missiles. 121 00:06:06,560 --> 00:06:09,760 Company came out of this batch. Aries Industries Y CS call for 122 00:06:09,760 --> 00:06:12,760 startups and their batches post that call usually are pretty 123 00:06:12,760 --> 00:06:17,040 reflective of trends are in Silicon Valley at large. 124 00:06:17,080 --> 00:06:20,400 They do lead in some ways on the cutting edge, but they also are 125 00:06:20,400 --> 00:06:24,560 reflective of where VC interest is a lot of the time, and it has 126 00:06:24,560 --> 00:06:26,760 shifted more to hard tech than it's been in recent years. 127 00:06:27,040 --> 00:06:30,280 I, I will say an investor texted me after our story came out. 128 00:06:30,480 --> 00:06:33,040 I think we were somewhat skeptical of Aries in the story 129 00:06:33,040 --> 00:06:36,160 too, but we were, yeah, this person said the company is so 130 00:06:36,160 --> 00:06:38,960 awful. And I was, I said in quality or 131 00:06:38,960 --> 00:06:40,920 intentions. You know, he's a missile 132 00:06:40,920 --> 00:06:43,320 company. I didn't, he said quality. 133 00:06:43,360 --> 00:06:46,880 So I, I feel like there's the investors, lots of Silicon 134 00:06:46,880 --> 00:06:49,720 Valley investors are ready to bet on a missile company, but 135 00:06:49,960 --> 00:06:52,080 we'll see if this is the one or not. 136 00:06:52,080 --> 00:06:54,000 Right. There's a lot will come down 137 00:06:54,000 --> 00:06:57,240 ultimately still to the team, especially with seed stage 138 00:06:57,240 --> 00:06:59,320 investing where the team is so important. 139 00:06:59,520 --> 00:07:02,000 I think also, I mean just in my conversations with some of the 140 00:07:02,000 --> 00:07:04,360 people over at Founders Fund, which have been really heavily 141 00:07:04,360 --> 00:07:08,400 involved in funding and founding Anderl especially Trey Stevens 142 00:07:08,400 --> 00:07:11,960 is a Co founder over there. That company has become sort of 143 00:07:11,960 --> 00:07:14,080 the banner success defense company. 144 00:07:14,080 --> 00:07:18,280 So obviously there is room for more than one, but how many more 145 00:07:18,400 --> 00:07:19,840 remains to be seen at this point. 146 00:07:19,840 --> 00:07:22,360 So they could, a lot of these companies could eventually 147 00:07:22,360 --> 00:07:25,840 become prime acquisition targets for like an and a roll or a 148 00:07:25,840 --> 00:07:27,120 Palantir or something like that or. 149 00:07:27,480 --> 00:07:30,960 Prime. Or Prime Exactly, pun intended. 150 00:07:31,360 --> 00:07:34,840 Another company among the defense tech plays in this batch 151 00:07:34,840 --> 00:07:37,840 was Theseus, which is building hardware and software for 152 00:07:37,840 --> 00:07:40,800 drones. So not necessarily weapons based 153 00:07:40,800 --> 00:07:43,640 drones that could be among their clientele. 154 00:07:43,720 --> 00:07:46,760 And that was a really hot round, highly contested first round 155 00:07:46,760 --> 00:07:49,320 capital ultimately ended up winning that round out from a 156 00:07:49,320 --> 00:07:52,400 couple other Tier 1 investors. So it'll be interesting to see 157 00:07:52,560 --> 00:07:56,320 how that company grows. And our story, besides looking 158 00:07:56,320 --> 00:07:59,600 at the companies, we also interrogated this sort of growth 159 00:07:59,680 --> 00:08:02,160 story of YC. I remember a few years ago I 160 00:08:02,160 --> 00:08:06,960 headlined the profile of YC in the Jeff Ralston era as YC 161 00:08:06,960 --> 00:08:09,080 equals growth. That was during the pandemic 162 00:08:09,080 --> 00:08:11,080 where they could do everything remote and they were scaling and 163 00:08:11,080 --> 00:08:12,640 scaling and scaling. They became a lot more 164 00:08:12,640 --> 00:08:14,440 international too during the pandemic. 165 00:08:14,440 --> 00:08:17,040 Like they accepted a lot of founders that were not based in 166 00:08:17,040 --> 00:08:18,840 San Francisco, based all around the world. 167 00:08:18,920 --> 00:08:22,040 And now that has sharply shifted under Gary Tan. 168 00:08:22,320 --> 00:08:24,560 YC is in person. You must relocate to San 169 00:08:24,560 --> 00:08:27,000 Francisco. It's been this way since 2022, 170 00:08:27,000 --> 00:08:30,240 but they've really honed in on that being the message that YC 171 00:08:30,240 --> 00:08:33,000 is also now very much based in San Francisco. 172 00:08:33,480 --> 00:08:39,320 So we we peaked at 749 companies a year in 2021, fell to 639 in 173 00:08:39,320 --> 00:08:43,360 2022 and now we are at 508 in 2024. 174 00:08:43,360 --> 00:08:47,880 That's part of the new batch strategy is keeping it at that 175 00:08:47,880 --> 00:08:51,000 same size of around 500 where they want to stick, which is 176 00:08:51,000 --> 00:08:52,760 still a lot bigger than what it was pre. 177 00:08:53,000 --> 00:08:59,800 Pandemic almost 5 * 2011 you know 2012 was 1512013 was 99. 178 00:08:59,880 --> 00:09:02,920 These batches that had like the iconic YC companies like Airbnb 179 00:09:02,960 --> 00:09:06,360 in them were much smaller when those took place. 180 00:09:06,360 --> 00:09:10,160 So this is in some ways a way to maintain sort of the point that 181 00:09:10,160 --> 00:09:13,280 they're at now with scaling back a bit from the pandemic, but 182 00:09:13,280 --> 00:09:15,800 then also offer smaller batch sizes. 183 00:09:15,800 --> 00:09:18,520 So hypothetically more personalized time with group 184 00:09:18,520 --> 00:09:22,600 partners, kind of how how to scale while remaining small is 185 00:09:22,600 --> 00:09:25,240 sort of the puzzle that they're figuring out right now. 186 00:09:25,240 --> 00:09:27,840 And investors tell me that. I mean, the batch quality has 187 00:09:27,840 --> 00:09:30,200 been great these last few. It seems like it's paying off, 188 00:09:30,200 --> 00:09:32,680 but we'll see once they add in some more batches. 189 00:09:33,000 --> 00:09:35,520 So there also has been this debate, you know, you look at 190 00:09:35,520 --> 00:09:38,960 why C's winners on their, you know, top companies page within 191 00:09:38,960 --> 00:09:41,640 their directory. And a lot of the biggest 192 00:09:41,640 --> 00:09:44,080 companies that have come out of why C really did come from sort 193 00:09:44,080 --> 00:09:47,560 of this the golden era of the mid twenty 10s and. 194 00:09:47,640 --> 00:09:50,720 Put it another way, when when Gary Tan wanted to talk about 195 00:09:50,720 --> 00:09:53,160 Founders mode, he was still talking about Brian Chesky. 196 00:09:53,160 --> 00:09:57,800 You know, Airbnb from the class of 2009, which still there's 197 00:09:57,920 --> 00:10:00,880 dining out on. Certainly Coinbase is another 198 00:10:00,880 --> 00:10:04,360 top win. DoorDash, Amplitude, Dropbox. 199 00:10:04,760 --> 00:10:08,480 So a lot of sort of old names. I think there's some more 200 00:10:08,480 --> 00:10:10,840 promising companies it up in recent classes. 201 00:10:10,840 --> 00:10:14,920 I think Cruz was AYC company. So I mean, there have been some 202 00:10:15,360 --> 00:10:17,320 bigger wins. That's not that recent to me. 203 00:10:17,840 --> 00:10:20,560 Yeah, but more recent than 2009, let's say. 204 00:10:21,000 --> 00:10:24,640 But that that's a good point that there is still much to 205 00:10:24,640 --> 00:10:28,280 prove in this new cohort. And part of this is also the 206 00:10:28,280 --> 00:10:30,840 world of seed investing. It takes a long time to see Bush 207 00:10:30,840 --> 00:10:32,960 vets take off and these are pre seed companies. 208 00:10:32,960 --> 00:10:35,920 So we should be thinking on, you know, 10 year time frames, but 209 00:10:36,160 --> 00:10:39,640 the size of the batches and potentially the dilution of 210 00:10:39,640 --> 00:10:42,320 quality of the companies that came about during this massive 211 00:10:42,320 --> 00:10:44,640 time of expansion is something that's clearly on the top of 212 00:10:44,640 --> 00:10:48,400 mind of Gary Tan and the YC team going forward. 213 00:10:48,560 --> 00:10:52,040 They want to retain their banner Ivy League brand of startup 214 00:10:52,040 --> 00:10:54,040 accelerators. And for the most part, investors 215 00:10:54,040 --> 00:10:56,600 still view them this way as from my conversations with people who 216 00:10:56,600 --> 00:10:58,560 back YC companies, they do view them there. 217 00:10:58,560 --> 00:11:02,160 But my Combinator is also facing a little more competition from 218 00:11:02,200 --> 00:11:04,720 accelerators that it didn't have in the past. 219 00:11:04,720 --> 00:11:07,520 While Techstars has kind of fallen off, the Neo accelerator 220 00:11:07,840 --> 00:11:11,160 has grown quite a bit and many investors speak to the quality 221 00:11:11,160 --> 00:11:12,480 of the founders that come out of that. 222 00:11:12,480 --> 00:11:16,920 Also, there's the HF 0 hacker house or hacker residency, if 223 00:11:16,920 --> 00:11:18,640 you will. I know that they're sensitive 224 00:11:18,640 --> 00:11:20,680 about that. It's not a hacker house, it's a 225 00:11:20,680 --> 00:11:23,600 real residency accelerator. But those are deep pockets at 226 00:11:23,600 --> 00:11:27,200 that accelerator to fund these, you know, AI forward companies. 227 00:11:27,200 --> 00:11:30,280 So Y Combinator is not the only game in town anymore. 228 00:11:30,280 --> 00:11:34,160 So quality and making sure that the experience is as 229 00:11:34,400 --> 00:11:37,160 personalized in depth and what founders want out of it is 230 00:11:37,160 --> 00:11:39,240 definitely something that's on the top of mind for them. 231 00:11:40,000 --> 00:11:43,280 Our second story of the week Microsoft taps 3 Mile Island 232 00:11:43,280 --> 00:11:46,920 nuclear plant to power AI Crazy. I mean just the future of. 233 00:11:46,960 --> 00:11:53,840 Nuclear It's the 20th century. A sign of how much power 234 00:11:53,960 --> 00:11:57,960 insiders and AI really think it's going to take to get to 235 00:11:58,280 --> 00:12:02,560 what AGI or to get to the, you know, $10 billion compute large 236 00:12:02,560 --> 00:12:05,640 language model. So if Microsoft is saying, oh, 237 00:12:05,640 --> 00:12:09,280 we need to find nuclear power to make this work, I think that's 238 00:12:09,280 --> 00:12:12,440 as much evidence as you need of how consumptive I think these 239 00:12:12,440 --> 00:12:14,160 models are going to be and how much they're going to keep 240 00:12:14,560 --> 00:12:18,080 investing enormous sums. I mean, I went to a tech 241 00:12:18,080 --> 00:12:21,680 conference a couple months ago now, and certainly power was one 242 00:12:21,680 --> 00:12:24,800 of their top concerns. Not to, you know, plug, you 243 00:12:24,800 --> 00:12:27,480 know, Super Valley again, but power was a big conversation 244 00:12:27,480 --> 00:12:29,840 topic at our Super Valley New York event and one of the 245 00:12:29,840 --> 00:12:32,800 discussion groups how we willpower all of the compute 246 00:12:32,800 --> 00:12:36,680 needed for AI to become as ubiquitous as we expect it to be 247 00:12:36,680 --> 00:12:40,160 and need it to be in the future. So I think there's also, you 248 00:12:40,160 --> 00:12:44,520 know, the conversation around clean energy and a lot of the 249 00:12:44,520 --> 00:12:47,000 tech world had been moving climate tech had been has is 250 00:12:47,000 --> 00:12:50,200 still, you know, an investing space that some people are 251 00:12:50,520 --> 00:12:53,280 putting lots of money into because there's, you know, an 252 00:12:53,280 --> 00:12:55,960 economical side to it. We can't have the AI revolution 253 00:12:55,960 --> 00:12:57,760 if we're burning Earth resources. 254 00:12:57,760 --> 00:12:59,720 You know where nuclear falls into. 255 00:12:59,720 --> 00:13:02,720 The green energy debate is also something that has gone back and 256 00:13:02,720 --> 00:13:05,600 forth over the years, but it's certainly not fossil fuels, and 257 00:13:05,720 --> 00:13:08,560 it can be another way to meet the power demands without overly 258 00:13:08,560 --> 00:13:11,960 relying on fossil fuels for AI. Yeah, if this is an excuse to 259 00:13:11,960 --> 00:13:15,440 invest in nuclear or like godspeed to the core idea of 260 00:13:15,440 --> 00:13:19,120 whether AI will require this, I think there are two themes sort 261 00:13:19,120 --> 00:13:21,240 of at once, right? It's like, OK, models keep 262 00:13:21,400 --> 00:13:23,560 needing to get bigger and bigger to get smarter. 263 00:13:23,560 --> 00:13:26,080 And so if you chart that out, you're like, man, we're going to 264 00:13:26,080 --> 00:13:27,960 need a lot of power, let's invest in it. 265 00:13:28,560 --> 00:13:31,960 On the other hand, you're seeing people use sort of like cheap 266 00:13:31,960 --> 00:13:35,960 models and open source models and small models to get a lot of 267 00:13:35,960 --> 00:13:38,320 stuff done. And certainly a player like 268 00:13:38,320 --> 00:13:42,880 Microsoft has sort of a vested interest in believing like, our 269 00:13:42,880 --> 00:13:44,400 big models are going to dominate. 270 00:13:44,400 --> 00:13:46,120 So we're going to need a ton of power. 271 00:13:46,320 --> 00:13:49,160 But there's certainly a world where it's like, what's tech 272 00:13:49,160 --> 00:13:52,560 good at doing? It's like doing things cheaply 273 00:13:52,560 --> 00:13:55,840 and sort of lightly. And like, you know, the returns 274 00:13:55,840 --> 00:13:58,120 from all these foundation models erode so quickly. 275 00:13:58,560 --> 00:14:02,080 So it just leads me to believe, like, are we really on this 276 00:14:02,080 --> 00:14:05,320 escalating sort of power consumption curve or are people 277 00:14:05,320 --> 00:14:08,160 going to find ways to do most of the stuff we need to do fairly 278 00:14:08,240 --> 00:14:11,080 cheaply? As chip competition maybe comes 279 00:14:11,080 --> 00:14:13,800 up more too, when people there's, you know, the shortage 280 00:14:13,800 --> 00:14:16,920 that is slightly easing, but still demand for Nvidia's chips 281 00:14:16,920 --> 00:14:19,280 is really high. But there's all these chip 282 00:14:19,280 --> 00:14:22,080 companies that are pitching themselves to run these models 283 00:14:22,080 --> 00:14:24,440 at a much more efficient pace using a lot less energy. 284 00:14:24,680 --> 00:14:27,600 So I think competition in the market could also just bring 285 00:14:27,600 --> 00:14:30,360 that on the hardware, the demand for energy down a bit too. 286 00:14:30,800 --> 00:14:31,920 Maddie, you want to talk about runway? 287 00:14:32,400 --> 00:14:35,400 Sure, Runway reached an agreement with Lionsgate to 288 00:14:35,400 --> 00:14:38,880 train its models on the studios library of movies and TV shows. 289 00:14:39,280 --> 00:14:43,120 This is the first of its kind of these deals that a lot of these 290 00:14:43,120 --> 00:14:46,320 model provider companies have been making with publications 291 00:14:46,320 --> 00:14:48,520 for the most part. But this is, you know, a video 292 00:14:48,520 --> 00:14:51,680 model that got a licensing deal with a production studio. 293 00:14:51,680 --> 00:14:55,000 So I mean there's some debate over whether or not these 294 00:14:55,000 --> 00:14:57,080 studios are kind of signing their own death certificate with 295 00:14:57,080 --> 00:15:00,040 these deals, similar with news media companies breaking in 296 00:15:00,040 --> 00:15:02,840 these deals with open AI. But I don't know, Eric, if you 297 00:15:02,840 --> 00:15:04,920 have thoughts, I. Mean it's a classic collective 298 00:15:04,920 --> 00:15:08,280 action problem where it's like if I were the monopolist who 299 00:15:08,280 --> 00:15:11,160 totally controlled the movie industry certainly I wouldn't 300 00:15:11,160 --> 00:15:16,400 train my potential disruptor to create movies with models. 301 00:15:16,760 --> 00:15:20,120 But if I'm just one studio among many that has earnings pressure 302 00:15:20,120 --> 00:15:23,520 and wants to make money today, and if not me, you know, my 303 00:15:23,520 --> 00:15:27,160 biggest enemy, then sure, I'll strike a deal with Runway to let 304 00:15:27,160 --> 00:15:29,680 them train and make the money now and hope that I'll figure it 305 00:15:29,680 --> 00:15:32,520 out down the road. So this is just sort of, yeah, 306 00:15:32,560 --> 00:15:35,040 it's business incentives. Same thing we're seeing in 307 00:15:35,040 --> 00:15:36,840 media. And fundamentally, I think it's 308 00:15:36,840 --> 00:15:40,040 good that Runway can take a crack at building models. 309 00:15:40,200 --> 00:15:42,880 And do so, you know, licensing with the licensing agreement to 310 00:15:42,880 --> 00:15:45,480 where it's not scraping these kind of movies to train on, you 311 00:15:45,480 --> 00:15:46,280 know? Right. 312 00:15:46,680 --> 00:15:49,040 As long as the argument isn't, well, we'll seal it if you don't 313 00:15:49,040 --> 00:15:51,440 give it to us. Certainly, right, That that's 314 00:15:51,440 --> 00:15:53,760 always the other tension with this, Like they're going to be 315 00:15:53,760 --> 00:15:56,920 rogue, actors are going to train and then you'll see them in a 316 00:15:56,920 --> 00:15:58,760 decade, you know? Absolutely. 317 00:15:59,160 --> 00:16:01,360 I mean, the other funny thing is, you know, just runways. 318 00:16:01,360 --> 00:16:06,080 CTO Anastasis spoke at Cerebral Valley New York in June and he 319 00:16:06,080 --> 00:16:08,120 was the one who was sort of saying, you know, we're going to 320 00:16:08,120 --> 00:16:12,120 have full AI movies on runway like imminently what, like two 321 00:16:12,120 --> 00:16:15,400 years or something. And our very AI friendly 322 00:16:15,400 --> 00:16:18,320 audience was like laughing at him fundamentally. 323 00:16:18,320 --> 00:16:21,720 So I love Runway, I really like the team there. 324 00:16:21,720 --> 00:16:26,000 But they certainly are projecting this sort of 325 00:16:26,120 --> 00:16:29,480 revolution in how movies are created that even some of their 326 00:16:29,480 --> 00:16:31,680 fellow technologists seem skeptical of. 327 00:16:31,680 --> 00:16:34,120 So we'll we'll see what they can make of this. 328 00:16:34,520 --> 00:16:37,200 Well, I've obviously they've been used in production with a 329 00:16:37,200 --> 00:16:40,720 24 movies before with Everything Everywhere All at Once, Yeah. 330 00:16:40,880 --> 00:16:43,800 Yeah, but they're not like, you know, they're sort of promising, 331 00:16:43,800 --> 00:16:45,560 like the whole movie, which is narrative. 332 00:16:45,560 --> 00:16:48,240 It's like, I don't think anyone doubts like you're gonna do 333 00:16:48,240 --> 00:16:52,000 special effects stuff with AI, but I do think if it's just 334 00:16:52,000 --> 00:16:55,320 special effects, my wife is a documentary and filmmaker. 335 00:16:55,320 --> 00:16:57,400 I don't know. I'm pretty bullish on Adobe, a 336 00:16:57,400 --> 00:17:00,520 company that really understands creators and can plug into 337 00:17:00,800 --> 00:17:02,360 creative tools they're already using. 338 00:17:02,360 --> 00:17:04,520 So you can see why there's pressure on a company like 339 00:17:04,520 --> 00:17:08,480 Runway to offer sort of the full answer, because that would be 340 00:17:08,480 --> 00:17:10,960 truly disruptive and not something maybe that Adobe is 341 00:17:10,960 --> 00:17:13,119 prepared to do. Yeah, that's a good point. 342 00:17:13,280 --> 00:17:18,200 So it it seems like with runway there is also the pressures on 343 00:17:18,200 --> 00:17:24,160 to be able to deliver a quality full AI made film, whether short 344 00:17:24,160 --> 00:17:27,240 or feature film that people will enjoy watching. 345 00:17:27,240 --> 00:17:29,640 And so training on, you know, narratives that people have 346 00:17:29,640 --> 00:17:31,200 already enjoyed certainly will help with that. 347 00:17:31,200 --> 00:17:34,040 But it remains to be seen if that really is within, you know, 348 00:17:34,040 --> 00:17:36,240 the next two years or four. It's going to take a little bit 349 00:17:36,240 --> 00:17:39,280 longer to get there. Our last story of the week, more 350 00:17:39,280 --> 00:17:43,160 lighthearted or or sort of sad, I don't know, depending how you 351 00:17:43,160 --> 00:17:46,120 look at it. Comedian John Mulaney, who I 352 00:17:46,120 --> 00:17:51,720 love, was I think the closeout speaker at Dreamforce or sort of 353 00:17:51,720 --> 00:17:53,720 helping wrap them up with entertainment. 354 00:17:54,200 --> 00:17:58,120 And he just sort of roasted the audience, pulled a couple of the 355 00:17:58,120 --> 00:18:00,560 quotes. Imagine me as John Mulaney. 356 00:18:00,560 --> 00:18:02,840 I'm not. I'm not a person impersonator 357 00:18:02,840 --> 00:18:05,040 here. You look like a group who looked 358 00:18:05,040 --> 00:18:08,080 at the self checkout counters at CVS and thought, this is the 359 00:18:08,080 --> 00:18:12,320 future, Mulaney said. If AI is truly smarter than us 360 00:18:12,320 --> 00:18:15,920 and tells us that humans should die, then I think we should die, 361 00:18:15,920 --> 00:18:18,840 he said, looking out to the crowd from center stage. 362 00:18:19,080 --> 00:18:21,640 So many of you feel imminently replaceable. 363 00:18:22,360 --> 00:18:24,560 Can AI sit there in a fleece vest? 364 00:18:24,560 --> 00:18:27,600 Can AI not go to events and spend all day at the bar? 365 00:18:29,320 --> 00:18:30,200 I don't know. What do you think? 366 00:18:31,000 --> 00:18:34,960 I mean, well, it has flashbacks of when TJ Miller hosted the 367 00:18:34,960 --> 00:18:37,000 Crunchies and insulted everyone, for sure. 368 00:18:37,640 --> 00:18:39,640 I think it's a little more light hearted than that, to be fair. 369 00:18:39,640 --> 00:18:42,600 And I think John Mulaney as a comedian tends to land a little 370 00:18:42,600 --> 00:18:45,600 bit more softer. But it was not exactly a 371 00:18:45,600 --> 00:18:49,080 receptive audience perhaps at at in the same way. 372 00:18:49,080 --> 00:18:50,960 Although, yeah. What do you think, Eric? 373 00:18:50,960 --> 00:18:52,760 I mean, I'm also a huge fan of John Mulaney. 374 00:18:53,160 --> 00:18:55,160 Honestly, I do find these bits kind of funny. 375 00:18:55,320 --> 00:18:59,880 He's got much darker. I saw that TJ TJ Miller Sketcher 376 00:18:59,880 --> 00:19:02,760 referencing which I'll touch on. Oh, right, you were there. 377 00:19:02,960 --> 00:19:07,120 I mean, it's some way Salesforce is an event of saying what you 378 00:19:07,120 --> 00:19:10,720 do matters to all these people who do fairly boring stuff. 379 00:19:10,720 --> 00:19:12,800 You know, it's like, I'm Marc Benioff, I'm going to donate a 380 00:19:12,800 --> 00:19:14,480 lot. We're creating a culture here. 381 00:19:14,480 --> 00:19:18,320 We have all this sort of entertainment and it's party and 382 00:19:18,320 --> 00:19:21,280 we connect with each other. And this is more, you know, in 383 00:19:21,280 --> 00:19:25,200 some ways it's a whole event oriented around justifying 384 00:19:25,240 --> 00:19:29,000 people who are doing important parts of capitalism, but not 385 00:19:29,000 --> 00:19:31,840 necessarily the most, like, glorious and celebration. 386 00:19:31,920 --> 00:19:34,120 And. So I get there, you know, they 387 00:19:34,120 --> 00:19:37,120 get paid for it and like the money is reward that the whole 388 00:19:37,120 --> 00:19:40,400 like Mad Men clip. But I I feel a little bad for 389 00:19:40,400 --> 00:19:43,640 them that come on, this is their their event to sort of celebrate 390 00:19:43,640 --> 00:19:45,840 themselves. I get you're hiring John 391 00:19:45,840 --> 00:19:48,040 Mulaney, but I think a lot of people think of, I don't know, 392 00:19:48,040 --> 00:19:51,240 the prehab John Mulaney, who is a little more upbeat. 393 00:19:51,280 --> 00:19:52,200 Yeah. So I don't know. 394 00:19:52,200 --> 00:19:56,400 I'm I I feel on both ends. So the TJ Miller 1, you know, 395 00:19:56,400 --> 00:19:59,720 this was at TechCrunch Crunchies, which were these 396 00:19:59,720 --> 00:20:02,480 weird awards where, you know, they'd give out trophies to 397 00:20:02,480 --> 00:20:05,760 companies that the real award was like going public and making 398 00:20:05,760 --> 00:20:07,600 a bunch of money. But it'd be like best startup 399 00:20:07,640 --> 00:20:11,520 and Stewart Butterfield would go up and collect for Slack or, or 400 00:20:11,520 --> 00:20:13,240 Travis Kalanick would go for Uber. 401 00:20:13,240 --> 00:20:17,000 So it was super, super weird. It was to me one of the last, 402 00:20:17,000 --> 00:20:20,000 like, Silicon Valley is really small where it's like, oh, we 403 00:20:20,000 --> 00:20:22,640 can have this. You know, and it was so on 404 00:20:22,640 --> 00:20:25,680 Hollywood and really uncool. And, you know, VCs are the 405 00:20:25,680 --> 00:20:28,280 celebrities. And so they bring in TJ Miller 406 00:20:28,280 --> 00:20:32,480 and like, he's like, really. Like, I mean, it felt like, is 407 00:20:32,480 --> 00:20:34,840 this guy drunk? And he was just shitting on 408 00:20:35,280 --> 00:20:37,880 Silicon Valley? I think he also called Travis's 409 00:20:37,880 --> 00:20:41,560 girlfriend at the time a bitch, which obviously soured a more 410 00:20:41,560 --> 00:20:44,160 progressive audience. So there are lots of layers to 411 00:20:44,200 --> 00:20:46,240 it. I mean, I, in some ways I 412 00:20:46,240 --> 00:20:48,640 thought it was funny because it was like, so hostile to the 413 00:20:48,640 --> 00:20:51,680 crowd that people were like, ready to leave that as a bit 414 00:20:52,000 --> 00:20:55,360 that he was just sort of at this awards show turning on the 415 00:20:55,360 --> 00:20:58,000 audience. Like, I was laughing a lot, but 416 00:20:58,000 --> 00:21:00,000 the crowd was not laughing very much. 417 00:21:00,120 --> 00:21:02,200 But I can see why. You know, it's like part of 418 00:21:02,200 --> 00:21:05,440 hosting the Oscars is finding the right balance between how 419 00:21:05,440 --> 00:21:08,920 much you shit on Hollywood and how much you don't. 420 00:21:08,920 --> 00:21:12,040 And some of these comedians feel like they don't appreciate tech 421 00:21:12,120 --> 00:21:14,440 at all. You know, at least if you're in 422 00:21:14,440 --> 00:21:17,880 Hollywood cheating on Hollywood, there's a level of like, I'm in 423 00:21:17,880 --> 00:21:21,200 the industry, obviously I have some fundamental respect of what 424 00:21:21,200 --> 00:21:23,800 you guys do. Whereas I feel like for tech, it 425 00:21:23,800 --> 00:21:26,840 can be pretty disconcerting when people come in and it feels like 426 00:21:27,240 --> 00:21:29,640 they have no love for this industry and they're just taking 427 00:21:29,640 --> 00:21:33,000 the check and going to sort of hate them to their faces. 428 00:21:33,000 --> 00:21:35,120 And yeah, it's it's somewhat distressing. 429 00:21:35,600 --> 00:21:38,720 It's not the same as, you know, as sort of the the inside we're 430 00:21:38,720 --> 00:21:41,760 all laughing at ourselves kind of way that it takes place, you 431 00:21:41,760 --> 00:21:43,560 know, at these at the Hollywood awards shows. 432 00:21:43,560 --> 00:21:45,680 Like it, it does feel a little bit more like people coming in 433 00:21:45,680 --> 00:21:49,520 from the outside to say like really just lean in and roasting 434 00:21:49,520 --> 00:21:53,120 and insulting people to some extent, despite some those 435 00:21:53,120 --> 00:21:56,200 insults being funny, to be fair. So, yeah, I think potentially 436 00:21:56,200 --> 00:21:59,240 not the right crowd, but John Mulaney's still a good get. 437 00:21:59,240 --> 00:22:02,480 I'll give them that. I mean, like, can AI sit there 438 00:22:02,480 --> 00:22:04,440 in a fleece vest? To some degree? 439 00:22:04,440 --> 00:22:07,320 It is sort of like you're a loser, you're not cool. 440 00:22:07,320 --> 00:22:10,560 Like, I don't know, it does read as a little like bully to me. 441 00:22:10,600 --> 00:22:12,480 You know, it's the classic, like, are you punching down or 442 00:22:12,480 --> 00:22:15,080 punching up? And I think Hollywood often has 443 00:22:15,080 --> 00:22:18,280 this sense that you're always punching up a tech because 444 00:22:18,280 --> 00:22:22,120 they're the real rich people. But then Hollywood is sort of 445 00:22:22,120 --> 00:22:24,360 all things cool and tech is all things not. 446 00:22:24,360 --> 00:22:26,920 And so that if you're like, you're a loser wearing a fleece 447 00:22:26,920 --> 00:22:29,440 vest, like are you punching up or punching down? 448 00:22:29,440 --> 00:22:32,920 And I think part of this is that the constant who's more 449 00:22:32,920 --> 00:22:36,080 powerful, Hollywood or tech? On the one hand, tech has all 450 00:22:36,080 --> 00:22:39,120 the money and they're the ones sort of, hey, hey, monkey, come 451 00:22:39,120 --> 00:22:41,520 dance on our stage. And then the monkey sort of like 452 00:22:41,520 --> 00:22:43,400 mean. So it it's they're, they're 453 00:22:43,400 --> 00:22:46,040 weird power dynamics. And I think that's part of what 454 00:22:46,040 --> 00:22:48,160 we're seeing here. Well, looking forward to the 455 00:22:48,160 --> 00:22:51,240 next week, I mean, Demo day is happening Wednesday and 456 00:22:51,240 --> 00:22:53,160 Thursday. I will be tuning in among 457 00:22:53,240 --> 00:22:56,600 hundreds of seed investors. So excited to be in the Zoom 458 00:22:56,600 --> 00:22:58,480 audience for that. But what other stories are you 459 00:22:58,480 --> 00:23:01,840 looking forward towards, Eric? I'm interested to see this open 460 00:23:01,840 --> 00:23:04,440 AI round come together. Who's willing to fork up the 461 00:23:04,440 --> 00:23:08,400 money, who gets in and certainly the how, you know, the continued 462 00:23:08,400 --> 00:23:12,280 evaluations of Strawberry remain interesting to me and sort of 463 00:23:12,640 --> 00:23:17,000 where where we are on improvements in open AIS tech. 464 00:23:17,080 --> 00:23:20,400 I have some interesting off the record events this week, so 465 00:23:20,400 --> 00:23:22,960 we'll be inhaling information from investors. 466 00:23:22,960 --> 00:23:24,960 We will not be discussing those, but I know. 467 00:23:24,960 --> 00:23:27,560 Sometimes my schedule is not fun to preview because I'm like, oh, 468 00:23:27,640 --> 00:23:29,080 that's secret. So. 469 00:23:30,080 --> 00:23:32,640 But that's our show. See you next week. 470 00:23:33,120 --> 00:23:33,520 Bye.