1 00:00:00,640 --> 00:00:03,720 Hi, I'm Eric Newcomer. And I am Madeline Rinbarger. 2 00:00:04,120 --> 00:00:10,880 And this is the newcomer podcast each week. 3 00:00:11,040 --> 00:00:14,640 Eric and I discuss the VC deals and the drama that went down. 4 00:00:15,240 --> 00:00:16,800 Let's do it. Here we go. 5 00:00:17,800 --> 00:00:20,840 There are weeks where it feels like the world has changed. 6 00:00:21,200 --> 00:00:24,560 And sometimes, you know, world history does transpire in a week 7 00:00:24,800 --> 00:00:27,400 and other times feels like we can get very excited about 8 00:00:27,400 --> 00:00:31,280 something and then it's it's less exciting than it seems. 9 00:00:31,280 --> 00:00:34,960 And I think that is the story of Deep Seek, the Chinese AI 10 00:00:34,960 --> 00:00:36,720 company that many of us hadn't heard of. 11 00:00:37,080 --> 00:00:39,600 You know, I'd probably heard on tech Twitter for a little bit, 12 00:00:39,600 --> 00:00:44,080 but certainly wasn't front of mind story until the Chinese 13 00:00:44,080 --> 00:00:46,800 tech company devastated the NASDAQ and devastated Nvidia's 14 00:00:46,800 --> 00:00:51,160 stock price in particular. NVIDIA shares plunged 17% on 15 00:00:51,160 --> 00:00:53,600 Monday. It was the largest market cap 16 00:00:53,600 --> 00:00:58,440 loss for AUS company in one day, close to 600 billion in value. 17 00:00:58,440 --> 00:01:02,720 Of course, that's back up now. The NASDAQ fell 3.1%, or 612 18 00:01:02,720 --> 00:01:05,600 points, but now it's only down 1.8%. 19 00:01:05,600 --> 00:01:09,000 So it seems like the the freak out of everyone discovering this 20 00:01:09,000 --> 00:01:11,880 over the weekend is now people know what it is. 21 00:01:11,880 --> 00:01:14,360 It's out there. It's a little bit, it's calmed, 22 00:01:14,360 --> 00:01:17,280 it's cold. But yeah, it's it feels like one 23 00:01:17,280 --> 00:01:20,400 of those weeks where all of a sudden the game has changed. 24 00:01:21,440 --> 00:01:23,080 Yeah. I mean, I think, you know, some 25 00:01:23,080 --> 00:01:28,120 things have changed, right? I mean, I think 1 and this I 26 00:01:28,120 --> 00:01:29,280 think has been talked about less. 27 00:01:29,280 --> 00:01:31,480 It's like, you know, it's Cerebral Valley in November is 28 00:01:31,480 --> 00:01:33,400 like, is there a wall? Is there a wall? 29 00:01:33,400 --> 00:01:37,000 You know, is AI going to improve Well, you know, deep seek isn't 30 00:01:37,000 --> 00:01:40,640 the only one changing the game like yes, there if you can be 31 00:01:40,760 --> 00:01:44,720 smarter in a more cost effective way, then yes, the game is 32 00:01:44,720 --> 00:01:48,200 definitely changing. It feels like certain that we're 33 00:01:48,200 --> 00:01:51,800 going to see serious AI development this year. 34 00:01:51,800 --> 00:01:55,880 So I I think that's clear. It underlines, you know how 35 00:01:55,880 --> 00:01:58,760 important open source is. I think people pay attention. 36 00:01:59,160 --> 00:02:02,480 You know, Llama and others were already sort of doing credible 37 00:02:02,480 --> 00:02:04,840 things. And then 3rd, on the technology 38 00:02:04,840 --> 00:02:08,800 side, obviously, it just reminds us that China is serious. 39 00:02:08,800 --> 00:02:12,240 It's a major global power. You know, it could be during the 40 00:02:12,240 --> 00:02:15,720 Trump administration that we sort of wake up and say, oh man, 41 00:02:15,720 --> 00:02:18,680 China is surpassing us on some things. 42 00:02:18,680 --> 00:02:21,880 And we can't just sort of pretend like American tech 43 00:02:21,880 --> 00:02:26,080 giants are the only source of activity in the world. 44 00:02:26,080 --> 00:02:29,240 I mean, I think honestly in self driving cars, you're seeing a 45 00:02:29,240 --> 00:02:32,240 lot of Chinese cities deploy self driving cars, even though 46 00:02:32,240 --> 00:02:34,920 in the US we mostly talk about way MO. 47 00:02:34,920 --> 00:02:38,000 So, yeah, there's definitely a lot of areas in AI where China 48 00:02:38,480 --> 00:02:41,600 is doing stuff that I guess our markets probably don't quite 49 00:02:41,600 --> 00:02:43,040 appreciate yet. Totally. 50 00:02:43,040 --> 00:02:46,600 I think one part of this too is that in the model race leaders 51 00:02:46,600 --> 00:02:50,000 have been saying for months now that the next breakthrough is 52 00:02:50,000 --> 00:02:53,360 always, you know, just the next model release away and 53 00:02:53,520 --> 00:02:55,240 competition can be really fierce, right? 54 00:02:55,240 --> 00:02:57,920 Like open AI has had the industry lead for a while. 55 00:02:57,920 --> 00:03:01,200 But as new developments come up, I remember, you know, when 56 00:03:01,200 --> 00:03:04,240 Anthropic dropped computer use that was is the agent that 57 00:03:04,280 --> 00:03:07,120 products that just, you know, took everyone by storm at that 58 00:03:07,120 --> 00:03:09,760 point. So it's not really a given that 59 00:03:09,840 --> 00:03:12,200 it's just going to be an American tech giant that's going 60 00:03:12,200 --> 00:03:14,720 to take the whole industry. Yeah, I don't know. 61 00:03:14,720 --> 00:03:18,000 Coming out of this whole experience, to me my initial 62 00:03:18,000 --> 00:03:22,360 reaction was vindication for the application thesis. 63 00:03:22,480 --> 00:03:24,960 We've been platforming people with the application thesis 64 00:03:24,960 --> 00:03:27,400 since Cerebral Valley, New York last June. 65 00:03:27,960 --> 00:03:30,520 Right. I mean, this idea that a lot of 66 00:03:30,520 --> 00:03:33,840 the value is going to accrue to people building use cases, 67 00:03:34,160 --> 00:03:36,520 whether for consumers or enterprises. 68 00:03:36,520 --> 00:03:38,920 It's the relationship with the customer, understanding how to 69 00:03:38,920 --> 00:03:41,600 sell it to people. And sort of the more traditional 70 00:03:41,600 --> 00:03:45,000 things around software businesses that would again 71 00:03:45,000 --> 00:03:48,400 differentiate AI companies and that foundation models were 72 00:03:48,400 --> 00:03:51,960 potentially spending a lot of money to build a commoditized 73 00:03:51,960 --> 00:03:55,480 technology. Certainly deep seek made that 74 00:03:55,480 --> 00:03:59,600 seem even more likely. I will say that, you know, as 75 00:03:59,600 --> 00:04:02,680 we're recording this podcast now, the Wall Street Journal is 76 00:04:02,680 --> 00:04:06,440 reporting that open AIS and talks to raise $40 billion in a 77 00:04:06,440 --> 00:04:10,200 $340 billion valuation from SoftBank. 78 00:04:10,240 --> 00:04:14,040 You know this, so you know. Investor appetite has not 79 00:04:14,040 --> 00:04:16,279 slowed. Sounds like they're going to do 80 00:04:16,279 --> 00:04:18,160 OK right now. Right, right. 81 00:04:18,160 --> 00:04:21,680 It's not as damning. You know, SoftBank clearly still 82 00:04:22,000 --> 00:04:26,720 a buyer for open AII. Mean now, you know, obviously 83 00:04:26,720 --> 00:04:30,320 open AI has a direct consumer product and there's part of the 84 00:04:30,320 --> 00:04:33,000 value proposition for them is that even if their technology 85 00:04:33,000 --> 00:04:37,960 gets commoditized, they are the Google, the iconic brand of the 86 00:04:37,960 --> 00:04:39,920 category. And so consumers will always 87 00:04:40,320 --> 00:04:42,320 flock to them. Right, like you could say that 88 00:04:42,320 --> 00:04:45,560 they really are the 1st to build a killer app so far. 89 00:04:45,560 --> 00:04:48,680 I mean, they, they have the app we talk about a lot about where 90 00:04:48,680 --> 00:04:50,600 the companies are going to come at the application layer, but 91 00:04:50,600 --> 00:04:53,880 they have an application that million like 300 million weekly 92 00:04:53,880 --> 00:04:56,920 active users as of last year. You know, like they have it. 93 00:04:56,920 --> 00:05:00,800 So it doesn't the the real value coming from applications also 94 00:05:00,800 --> 00:05:02,560 doesn't necessarily count out open AII. 95 00:05:03,040 --> 00:05:06,680 Think the thing that is muddled? A lot of takes and honesty makes 96 00:05:06,680 --> 00:05:10,240 it even hard for us to opine on some of it. 97 00:05:10,240 --> 00:05:14,800 Is this debate about how much deep seek is sort of a psyop or 98 00:05:14,800 --> 00:05:18,600 like how how honest are they about their access to NVIDIA 99 00:05:18,600 --> 00:05:20,240 chips and how much money is being spent? 100 00:05:20,240 --> 00:05:23,040 You're this is really where journalism fails because you're 101 00:05:23,040 --> 00:05:26,320 seeing a lot of narratives have to get written where it's like 102 00:05:26,800 --> 00:05:30,280 you classic journalism, you pick one side or the other of that, 103 00:05:30,280 --> 00:05:33,520 like either they had a lot of money or they didn't, but then 104 00:05:33,520 --> 00:05:35,160 you know, you don't know for sure. 105 00:05:35,160 --> 00:05:37,640 And so you have to sort of thread to competing narratives. 106 00:05:37,640 --> 00:05:40,160 So I do think. Yeah, the waffling is pretty 107 00:05:40,160 --> 00:05:42,680 apparent everywhere. I mean just because we need more 108 00:05:42,680 --> 00:05:44,960 information. We will not participate in that 109 00:05:44,960 --> 00:05:46,920 as talking heads. I think it's it's an open 110 00:05:46,920 --> 00:05:49,240 question. I think what's clear, you know, 111 00:05:49,240 --> 00:05:52,920 it's clear that they have NVIDIA chips, but there are, you know, 112 00:05:52,920 --> 00:05:56,160 things beyond H1 hundreds. And it's not clear exactly the 113 00:05:56,160 --> 00:05:59,200 mix and how high-powered their access is and whether it 114 00:05:59,200 --> 00:06:03,920 represents this sort of random company and a hedge fund or if 115 00:06:03,920 --> 00:06:06,960 this is part of the Chinese government's real sort of bet 116 00:06:06,960 --> 00:06:09,320 here and they're sort of backing that we don't know about. 117 00:06:09,560 --> 00:06:13,680 And then there is, I think more certainty that what Deep Seek 118 00:06:13,680 --> 00:06:18,080 has released can be run on a pretty cheap, in a cheap way, 119 00:06:18,080 --> 00:06:20,280 like you can reproduce it, you can run the model. 120 00:06:20,480 --> 00:06:23,640 And so certainly doing the thing they have done can be done 121 00:06:23,640 --> 00:06:26,120 cheaply and that will be important for AI. 122 00:06:26,520 --> 00:06:30,440 What is less clear is the amount of investment it took them to do 123 00:06:30,440 --> 00:06:34,120 it, which is people are sort of comparing apples to oranges 124 00:06:34,120 --> 00:06:37,160 where they're like, Oh my God, our foundation models spent so 125 00:06:37,160 --> 00:06:40,800 much money and they at the end of the day have produced 126 00:06:40,800 --> 00:06:42,440 something that you can do cheaply. 127 00:06:42,440 --> 00:06:45,920 But obviously if you just like follow everyone elses research 128 00:06:45,920 --> 00:06:48,600 and then do something at the end like it'll, it'll be cheap. 129 00:06:48,600 --> 00:06:51,640 Of course, that's cheaper. Yeah, jumping in in like the 130 00:06:51,640 --> 00:06:54,960 last mile to say and build something a lot faster with your 131 00:06:54,960 --> 00:06:57,760 Sprint certainly looks faster than doing the underlying work, 132 00:06:57,760 --> 00:06:59,560 right? I mean, there's been a the juicy 133 00:06:59,560 --> 00:07:02,680 irony, I think, you know, that the media class has enjoyed 134 00:07:02,680 --> 00:07:05,280 that. It feels like open AI is like 135 00:07:05,800 --> 00:07:08,400 how dare you copy our technology? 136 00:07:08,400 --> 00:07:09,680 You know what? What? 137 00:07:09,920 --> 00:07:12,320 You can't inhale everything in the world. 138 00:07:12,360 --> 00:07:14,600 Only we can inhale everything in the world. 139 00:07:14,600 --> 00:07:18,440 The irony is not lost on those of us who create original 140 00:07:18,440 --> 00:07:21,200 content that has become, you know, the high quality training 141 00:07:21,200 --> 00:07:25,520 data that these models use that now open AI and and you know, 142 00:07:25,560 --> 00:07:28,640 these these closed models especially are getting very 143 00:07:28,640 --> 00:07:30,600 angry about something being trained on top of them. 144 00:07:30,640 --> 00:07:32,760 Yeah. So I, you know, I think open 145 00:07:32,760 --> 00:07:37,200 source remains super powerful. I think Anthropic and Open AI 146 00:07:37,400 --> 00:07:40,080 are really going to have to work to show that they can stay 147 00:07:40,080 --> 00:07:41,960 ahead. And so this is going to put 148 00:07:41,960 --> 00:07:45,560 pressure on them to release new versions faster. 149 00:07:45,560 --> 00:07:49,240 And I think we've seen what you could fairly call defensiveness 150 00:07:49,240 --> 00:07:53,520 from both Open AI and Anthropic. Dario and Anthropic wrote a long 151 00:07:53,840 --> 00:07:56,800 note on sort of a tangential point that I think we can talk 152 00:07:56,800 --> 00:07:59,560 about, but sort of, you know, responding to everything that 153 00:07:59,560 --> 00:08:02,520 was going on. Open AI has been tweeting about 154 00:08:02,520 --> 00:08:04,120 this. You know this, this is the new 155 00:08:04,120 --> 00:08:07,080 world where everybody, like you don't ignore your competitors 156 00:08:07,120 --> 00:08:09,320 like big news. You sort of like get in the 157 00:08:09,320 --> 00:08:10,680 Twitter trenches. You have to. 158 00:08:10,680 --> 00:08:12,440 Jump in with your take. Yeah, exactly. 159 00:08:12,440 --> 00:08:14,000 What did you think of the reaction? 160 00:08:14,120 --> 00:08:16,600 Found the reaction from the little tech boosters actually to 161 00:08:16,600 --> 00:08:19,080 be one of the more positive sides of this. 162 00:08:19,360 --> 00:08:23,160 Alex Rampell over at A16Z called this a Sputnik moment, but also, 163 00:08:23,160 --> 00:08:27,120 you know, compared that to how Russians getting ahead of the US 164 00:08:27,120 --> 00:08:30,360 and the space race led to like a massive wave of US innovation 165 00:08:30,360 --> 00:08:33,360 and space technology. And so I think there's a lot to 166 00:08:33,360 --> 00:08:36,120 be said for this really opening up the application layer. 167 00:08:36,120 --> 00:08:38,880 If foundation models get cheaper, there's a ton of value 168 00:08:38,880 --> 00:08:41,280 to be added if you can build killer apps for it. 169 00:08:41,280 --> 00:08:44,840 So I don't really see this, you know, as like a death knell for 170 00:08:44,840 --> 00:08:48,240 the venture space that's gone all in on AI, if you will. 171 00:08:48,240 --> 00:08:51,000 I think there's plenty of funds that have much to invest in on 172 00:08:51,000 --> 00:08:53,200 the application layer that this actually opens up a lot more 173 00:08:53,200 --> 00:08:56,400 opportunity for founders and investors to really get into the 174 00:08:56,400 --> 00:08:58,840 space. If if this if the foundation 175 00:08:58,840 --> 00:09:01,880 model layer is where it is now and everything's open, you can 176 00:09:01,880 --> 00:09:04,280 build on top. I like I like the Andreessen 177 00:09:04,600 --> 00:09:08,320 Horowitz Impala essay. I know the Sputnik moment was 178 00:09:08,320 --> 00:09:09,960 smart. Madeline, you've been talking to 179 00:09:09,960 --> 00:09:12,240 investors all week about this. What's what? 180 00:09:12,240 --> 00:09:14,960 What's your read from them? I mean, a lot of them are kind 181 00:09:14,960 --> 00:09:17,880 of taking, I think some of them can get credit for this, some of 182 00:09:17,880 --> 00:09:19,680 them can't. But a lot of people are saying, 183 00:09:19,680 --> 00:09:22,360 you know, we always knew that the value really wasn't going to 184 00:09:22,360 --> 00:09:24,600 accrue at the model layer. You know, it's really at the 185 00:09:24,600 --> 00:09:26,720 application layer. We knew this the whole time. 186 00:09:26,720 --> 00:09:29,840 So I've gotten a couple, you know, takes of people maybe 187 00:09:29,960 --> 00:09:32,640 talking their own books a bit about how smart they were about 188 00:09:32,640 --> 00:09:33,920 this whole thing. But what? 189 00:09:33,920 --> 00:09:38,080 What firms are most expensive those two foundation model bets? 190 00:09:38,360 --> 00:09:40,600 Well, certainly Thrive is pretty exposed. 191 00:09:40,600 --> 00:09:44,200 They've just given so much and invested so much into open AI. 192 00:09:44,480 --> 00:09:47,560 Menlo Ventures is probably is one of the largest backers of 193 00:09:47,560 --> 00:09:51,320 Anthropic outside of the big tech companies like Google and 194 00:09:51,320 --> 00:09:54,200 Amazon. A16Z obviously has backed a lot 195 00:09:54,200 --> 00:09:56,360 of application companies, but they have. 196 00:09:56,360 --> 00:09:58,760 That's the opposite. That, but well, they've written 197 00:09:58,960 --> 00:10:02,720 really big checks into Miss Straw OK open AI through the 198 00:10:02,720 --> 00:10:05,240 secondary sale from employee tender. 199 00:10:05,240 --> 00:10:06,560 So they're kind of wrapped up in. 200 00:10:06,560 --> 00:10:08,520 Everything, yeah. Exactly. 201 00:10:08,720 --> 00:10:10,240 Yeah. So they're they're kind of on 202 00:10:10,240 --> 00:10:12,200 the application side and the model layer. 203 00:10:12,200 --> 00:10:15,080 So they're and maybe that's a too big to fail situation, but 204 00:10:15,080 --> 00:10:18,160 they have a lot of exposure. I know Sequoia and Tiger Global 205 00:10:18,160 --> 00:10:21,080 also bought in the open AI share, so share sales. 206 00:10:21,080 --> 00:10:24,880 So we're going to see, I mean the big firms obviously do have 207 00:10:24,880 --> 00:10:27,800 a lot wrapped up in this. So maybe this is, you know, the 208 00:10:27,800 --> 00:10:29,960 little tech victory lap, if you will. 209 00:10:30,560 --> 00:10:34,160 This should almost go without being stated, but I think I need 210 00:10:34,160 --> 00:10:37,880 to say it. It's like AI being better and 211 00:10:37,880 --> 00:10:41,160 cheaper is obviously good for AI. 212 00:10:41,240 --> 00:10:44,840 Broadly construed right? Bad bad for people only 213 00:10:44,840 --> 00:10:49,080 differentiate by having the smartest and fastest AI in town. 214 00:10:49,080 --> 00:10:52,120 But in the fact consumers are more likely to use something 215 00:10:52,120 --> 00:10:55,920 that is better and cheaper to use and so I do think some of 216 00:10:55,920 --> 00:10:58,280 the whiplash confusing reactions. 217 00:10:58,280 --> 00:11:01,000 And yeah, isn't this the the Jevons paradox point? 218 00:11:01,040 --> 00:11:04,440 Yeah, Something gets cheaper, more people use it and demand 219 00:11:04,440 --> 00:11:06,400 goes up. I mean, I feel like that's also 220 00:11:06,400 --> 00:11:10,320 been just seen with, you know, different service adoption for 221 00:11:10,360 --> 00:11:11,840 tools throughout startup history. 222 00:11:12,080 --> 00:11:14,400 Yeah. So it was an interesting, it was 223 00:11:14,400 --> 00:11:17,840 a market test. It was definitely, I think 224 00:11:17,840 --> 00:11:20,560 Silicon Valley was rightly like by the dip, and I think we're 225 00:11:20,560 --> 00:11:23,800 seeing that borne out. It was sort of like a freak out. 226 00:11:24,560 --> 00:11:27,320 Felt like outside of Silicon Valley, even though Silicon 227 00:11:27,320 --> 00:11:30,120 Valley reacted, it was like an interesting cultural moment. 228 00:11:30,600 --> 00:11:33,480 Well, I think it's important to circle back to Dario's point, 229 00:11:33,480 --> 00:11:36,840 all really being about export controls and how the leadership 230 00:11:36,840 --> 00:11:40,160 here isn't isn't necessarily a given that America will have 231 00:11:40,160 --> 00:11:42,480 this leadership. And obviously he's team safety 232 00:11:42,480 --> 00:11:46,400 with anthropics whole, you know, safe AI controlled top down kind 233 00:11:46,400 --> 00:11:48,320 of strategy. But he's saying, no, actually 234 00:11:48,320 --> 00:11:50,640 the export controls are working because I didn't get quite as 235 00:11:50,640 --> 00:11:53,040 many chips. And if we, you know, get, you 236 00:11:53,040 --> 00:11:57,320 know, strengthen control over this part of the stack and keep 237 00:11:57,320 --> 00:12:00,120 it really in the West, then we can stay ahead. 238 00:12:00,200 --> 00:12:04,760 And I feel like in this sort of open opening up moment and, you 239 00:12:04,760 --> 00:12:09,160 know, big, you know, revoking of the controlling Biden AI 240 00:12:09,440 --> 00:12:11,680 executive order that, you know, the administration's doing 241 00:12:11,680 --> 00:12:17,040 everyone's all, all in EACC all the way, you know, like like at 242 00:12:17,040 --> 00:12:20,640 this top down kind of safety mindset isn't necessarily. 243 00:12:21,960 --> 00:12:25,040 Right, well there there are two pieces of this like you know, 244 00:12:25,040 --> 00:12:28,080 the export controls. First of all, it's hard not to 245 00:12:28,080 --> 00:12:29,640 read his talking his own book right. 246 00:12:29,640 --> 00:12:33,560 Like certainly open AI and Anthropic would benefit from 1 247 00:12:33,560 --> 00:12:39,080 less major competitor in if we deny China NVIDIA chips, they'll 248 00:12:39,080 --> 00:12:40,720 have a harder time being competitors. 249 00:12:40,720 --> 00:12:42,240 Therefore we have less competition. 250 00:12:42,480 --> 00:12:47,240 Great for us very anti free market capitalism, but but 251 00:12:47,800 --> 00:12:51,440 potentially, you know, I mean, I, I, you know, probably agree 252 00:12:51,440 --> 00:12:56,200 with the export controls just in the sense that like everything 253 00:12:56,200 --> 00:12:59,480 with China seems to flow one way where they get to take all our 254 00:12:59,480 --> 00:13:02,000 stuff and then our companies can't operate there and they 255 00:13:02,280 --> 00:13:04,000 they, they benefit from our work. 256 00:13:04,040 --> 00:13:08,240 I mean, that said, I do think the open source nature of this 257 00:13:08,240 --> 00:13:12,800 makes things a little different because, I mean, deep seek is 258 00:13:12,800 --> 00:13:15,320 giving it away, you know? And like, that's the ultimate 259 00:13:15,800 --> 00:13:18,560 sort of good thing. Ben Shapiro had a terrible tweet 260 00:13:18,560 --> 00:13:22,960 where he's like, the deep seek surprise is a fantastic reminder 261 00:13:22,960 --> 00:13:26,000 that our enemies are motivated and aggressive and that our 262 00:13:26,000 --> 00:13:28,320 technological security is not guaranteed, I guess. 263 00:13:28,360 --> 00:13:30,760 OK, that's a fair take. But Bill Gurley, I thought, had 264 00:13:30,760 --> 00:13:36,040 a very moving response, which was, why are they enemies? 265 00:13:36,280 --> 00:13:39,480 No one that works at deep seek is an enemy of mine. 266 00:13:39,680 --> 00:13:42,560 The fact that they love open source makes me think I would 267 00:13:42,560 --> 00:13:46,440 gel with them quite well. And I mean, I, I think, you 268 00:13:46,440 --> 00:13:52,040 know, the right China message is that troubled by Chinese 269 00:13:52,120 --> 00:13:56,200 government's obviously free speech policies, control and 270 00:13:56,200 --> 00:13:58,080 then also their industrial policies. 271 00:13:58,440 --> 00:14:02,280 But Silicon Valley probably has more in common with the average 272 00:14:02,520 --> 00:14:04,960 tech person in China than they do with the average American. 273 00:14:04,960 --> 00:14:07,640 You know, like that's, that's the ongoing contradiction. 274 00:14:07,640 --> 00:14:11,440 Like while deep seek was able to operate fairly under the radar 275 00:14:11,440 --> 00:14:14,760 as like, I don't know, hardcore technologist and then give it 276 00:14:14,760 --> 00:14:17,600 all away. Yeah, they're they're very In 277 00:14:17,600 --> 00:14:20,480 Sync with a lot of I was like. That's like the the classic 278 00:14:20,480 --> 00:14:23,040 Silicon Valley story. That's the classic Silicon 279 00:14:23,040 --> 00:14:26,200 Valley ethos right there. But you know, then as soon as 280 00:14:26,200 --> 00:14:30,040 deep seek becomes, I don't know, China's most important AI 281 00:14:30,040 --> 00:14:33,800 company, then it becomes about sort of nation states competing 282 00:14:33,800 --> 00:14:36,000 and that, you know, these things get complicated. 283 00:14:36,000 --> 00:14:39,800 We also worth mentioning that now Alibaba is like coming in 284 00:14:39,800 --> 00:14:41,760 claiming that they have like something even better. 285 00:14:41,760 --> 00:14:44,280 So it'll be interesting. Again, we don't know what's 286 00:14:44,280 --> 00:14:46,240 happening in China. So it's like, is this all like, 287 00:14:46,240 --> 00:14:49,560 oh, they're ready to like announce their sort of big hits 288 00:14:49,560 --> 00:14:51,640 to make a message or who, who knows? 289 00:14:51,640 --> 00:14:55,240 But I agree with Gurley fundamentally that if they're 290 00:14:55,240 --> 00:14:59,680 giving it away, how, how evil could not be, you know? 291 00:15:00,000 --> 00:15:03,160 Yeah, I mean, I was talking to LON at basis set yesterday who 292 00:15:03,160 --> 00:15:06,440 just said the best line to me, which was, you know, people to 293 00:15:06,440 --> 00:15:08,880 make this too much of a big deal because they told you the 294 00:15:08,880 --> 00:15:12,360 recipe, just take it and make the dish, like, right, they gave 295 00:15:12,360 --> 00:15:13,680 it away. You know, why? 296 00:15:13,800 --> 00:15:16,440 Why are we spreading over how it happened? 297 00:15:16,440 --> 00:15:19,400 And they've given us a gift. We can build on top of it now. 298 00:15:19,520 --> 00:15:21,880 Right. The other piece of Dario is that 299 00:15:21,880 --> 00:15:26,080 say, you know, the safety ISM. I mean, I probably would have 300 00:15:26,080 --> 00:15:27,840 agreed with him. I probably agree with him. 301 00:15:27,840 --> 00:15:32,160 But like, man, that feels like a lost a lost fight. 302 00:15:32,160 --> 00:15:34,800 You know, it's like if we can't even keep the executive order or 303 00:15:34,800 --> 00:15:38,400 this sort of the idea that we're going to like hold on to AI, 304 00:15:38,400 --> 00:15:42,600 keep it in sort of a tightly held set of hands and make sure 305 00:15:42,600 --> 00:15:45,960 it doesn't get out of control is feeling less and less likely. 306 00:15:45,960 --> 00:15:49,040 I mean, I do think there's a world in which we look back at 307 00:15:49,040 --> 00:15:53,720 100 years over this time and we happened to build AI in this 308 00:15:53,720 --> 00:15:57,680 very anti elite moment, which meant it was totally ungoverned 309 00:15:57,680 --> 00:16:00,640 and we sort of let it run wild. And we're like, Oh my God, the 310 00:16:00,640 --> 00:16:04,360 one time we should have been sort of more sort of top down 311 00:16:04,360 --> 00:16:06,840 and sort of thoughtful about how this all came out. 312 00:16:07,000 --> 00:16:09,800 We just happened for totally other cultural reasons to be in 313 00:16:09,800 --> 00:16:13,960 a like, you know, let everybody do everything wild moment and 314 00:16:13,960 --> 00:16:17,520 like so we create, you know, the all powerful AI that that 315 00:16:17,520 --> 00:16:19,640 destroys us all. You know, hopefully that doesn't 316 00:16:19,640 --> 00:16:22,720 happen, but but it certainly makes it a little more likely. 317 00:16:22,840 --> 00:16:24,560 It's feeling very science fiction novels. 318 00:16:24,560 --> 00:16:27,560 See what they did take, you know, prediction but. 319 00:16:27,680 --> 00:16:29,720 Exactly. Yeah, you get right exactly. 320 00:16:29,720 --> 00:16:33,200 It's classic that you get sort of destroyed by your vice, 321 00:16:33,200 --> 00:16:34,800 right? Your vice, you know, it's like 322 00:16:35,200 --> 00:16:37,600 we're in this sort of like burn it all down mode. 323 00:16:37,600 --> 00:16:41,240 And the thing that destroys us is the thing that we didn't 324 00:16:41,240 --> 00:16:44,440 quite appreciate coming that needed a little bit more of 325 00:16:44,680 --> 00:16:48,320 collective control. Marc Andreessen would hate that 326 00:16:48,320 --> 00:16:52,440 take with all of his being, but. This is not the Marc Andreessen 327 00:16:52,440 --> 00:16:57,080 podcast, so we can. Escode it freely well. 328 00:16:57,080 --> 00:17:00,920 Moving on from China and deep seek and the crumbling of the 329 00:17:00,920 --> 00:17:03,800 American empire, let's talk about fund returns. 330 00:17:03,800 --> 00:17:06,319 Let's talk about funding data. Eric, you got an exclusive 331 00:17:06,319 --> 00:17:10,119 report from Angelus this week. Wasn't the prettiest picture, 332 00:17:10,119 --> 00:17:13,240 right, of the venture market? I I don't know if this is moving 333 00:17:13,240 --> 00:17:16,720 on from the crumbling of the crumbling of the American empire 334 00:17:16,720 --> 00:17:20,079 at all. You know, I mean, you know, 335 00:17:20,079 --> 00:17:23,920 basically, yeah. Angel List gave me their annual 336 00:17:23,920 --> 00:17:27,599 report, which I find super useful on the state of venture 337 00:17:28,000 --> 00:17:30,360 from their platform. You know, obviously some of the 338 00:17:30,360 --> 00:17:34,080 firms on Angel List are smaller, but they see more and more of 339 00:17:34,080 --> 00:17:37,120 the sort of market and generally have a good sense that I think 340 00:17:37,120 --> 00:17:38,840 matches with what everybody else sees. 341 00:17:39,120 --> 00:17:43,280 And you know, one big take is that basically if you wanted to 342 00:17:43,280 --> 00:17:47,400 mark up your venture capital firm, you needed to be active in 343 00:17:47,400 --> 00:17:50,720 2021 and the time around that like everybody got their markups 344 00:17:50,720 --> 00:17:54,280 then and then it's been flat. If you started after 2021, you 345 00:17:54,280 --> 00:17:55,840 basically don't have many markups. 346 00:17:56,160 --> 00:17:58,760 And if you started before it, you had all your markups, then 347 00:17:58,760 --> 00:18:02,280 you're sort of killing time. I think the other big take for 348 00:18:02,280 --> 00:18:07,080 venture capital firms is it they have, you know, TVPI, they've 349 00:18:07,160 --> 00:18:09,600 they've marked up their holdings, but they don't have 350 00:18:09,600 --> 00:18:11,680 much DPI. They don't have distributions 351 00:18:11,680 --> 00:18:13,840 and you know, you can't eat TVPI. 352 00:18:13,840 --> 00:18:19,000 Ultimately, LP's want DPI. And so I called this year a 353 00:18:19,000 --> 00:18:22,000 Crucible year. You know, it's because one of 354 00:18:22,000 --> 00:18:24,040 two things is going to have to happen. 355 00:18:24,320 --> 00:18:28,920 Either those marks turn into exits and they can pat 356 00:18:28,920 --> 00:18:31,840 themselves on the back or, you know, they're going to have to 357 00:18:31,840 --> 00:18:36,120 eat crow at least to some extent and say, oh, you know, we 358 00:18:36,360 --> 00:18:38,680 thought they were more valuable, but actually they're never going 359 00:18:38,680 --> 00:18:41,680 to turn into distributions. And if they don't turn into 360 00:18:41,680 --> 00:18:46,200 distributions, you know, that's potentially going to undermine 361 00:18:46,760 --> 00:18:49,920 LP's view of the the venture capital model. 362 00:18:49,920 --> 00:18:53,720 So we really have these two forks in the road, which is some 363 00:18:53,720 --> 00:18:57,320 of these holdings are valuable around what the VCs say they 364 00:18:57,320 --> 00:19:01,280 are. And we get some exits or we 365 00:19:01,280 --> 00:19:02,680 don't. But it feels like it's been 366 00:19:02,680 --> 00:19:07,320 going on long enough that 2025 should be the year where we have 367 00:19:07,320 --> 00:19:10,560 that reckoning. I'm skeptical that these. 368 00:19:10,560 --> 00:19:13,560 Yeah, I was going to say yeah. Distributions. 369 00:19:14,120 --> 00:19:16,880 Yeah, I was going to say how, what's your take on this? 370 00:19:16,880 --> 00:19:19,760 I mean, obviously there was some startups that also kind of, you 371 00:19:19,760 --> 00:19:23,080 know, scraped along and if, you know, tried to hold off before 372 00:19:23,320 --> 00:19:26,640 raising again just because due to the poor market conditions, 373 00:19:26,640 --> 00:19:30,520 like are there many signals that you could see us, you know, 374 00:19:30,520 --> 00:19:33,200 moving into a better exit year this year? 375 00:19:33,200 --> 00:19:36,840 I mean, the IPOs that we've had so far not been promising. 376 00:19:37,480 --> 00:19:39,840 Yeah, I mean everybody thinks we're getting better M&A. 377 00:19:39,840 --> 00:19:43,520 So maybe you know we'll see some activity there and like that 378 00:19:43,520 --> 00:19:45,880 could help, you know your billion dollar company could 379 00:19:45,880 --> 00:19:49,320 sell for 500 million though with preference tax and anything as 380 00:19:49,320 --> 00:19:51,080 we've talked about on other episodes. 381 00:19:51,080 --> 00:19:53,480 You know that's not necessarily a good outcome. 382 00:19:53,480 --> 00:19:57,000 IPOs so far haven't looked good. I mean, basically, you know, if 383 00:19:57,000 --> 00:20:01,360 you hold data bricks, you know, open AI stripe, you know, the 384 00:20:01,360 --> 00:20:05,880 big winners and so those. Those funds, you know, firm, you 385 00:20:05,880 --> 00:20:08,520 know, LP's can see that they're fairly liquid. 386 00:20:08,520 --> 00:20:12,240 They do secondary deals like I mean we talked about in your 387 00:20:12,240 --> 00:20:15,320 Utemco piece, which people should subscribe to newcomer for 388 00:20:15,320 --> 00:20:17,920 and go check out. You know, Thrive is doing well 389 00:20:17,920 --> 00:20:21,360 partially because they're doing these late stage deals where 390 00:20:21,360 --> 00:20:24,320 they're raising again and they get to mark them up. 391 00:20:24,680 --> 00:20:28,120 But like if you're a firm that's not in one of those big names, I 392 00:20:28,120 --> 00:20:31,680 don't think people are very optimistic that the random, you 393 00:20:31,680 --> 00:20:35,720 know, $3 billion Unicorn is going to deliver for you. 394 00:20:35,720 --> 00:20:38,840 And obviously a couple of them will, but I think a lot of them 395 00:20:38,840 --> 00:20:41,200 won't. Yeah, it seems kind of like the 396 00:20:41,200 --> 00:20:44,000 power law, but on steroids, if you will. 397 00:20:44,120 --> 00:20:46,000 Right, Exactly right. Yeah. 398 00:20:46,320 --> 00:20:49,840 How little DPI is like, I don't know on some of these funds, 399 00:20:49,840 --> 00:20:55,960 like even like the 75th core percentile from their oldest 400 00:20:55,960 --> 00:21:00,760 vintage funds, 2017 has only returned .64 X. 401 00:21:00,760 --> 00:21:04,160 So you don't even have all your money yet, like eight years 402 00:21:04,160 --> 00:21:07,560 later, you know? Yeah, if you're going on a 10 403 00:21:07,560 --> 00:21:09,520 year cycle, that's not feeling right. 404 00:21:10,120 --> 00:21:12,840 I know nobody wants to be on a 10 year cycle anymore. 405 00:21:13,800 --> 00:21:17,560 So you know the only other, you know, AI, obviously lots of AI 406 00:21:17,560 --> 00:21:19,560 deals. There is the counter intuitive 407 00:21:19,560 --> 00:21:23,880 point that valuations are basically like 10% up, but you 408 00:21:23,880 --> 00:21:26,560 have to remember that's because fewer people are raising 409 00:21:27,120 --> 00:21:28,560 starting companies and raising money. 410 00:21:28,600 --> 00:21:29,880 I don't know what was your biggest. 411 00:21:30,040 --> 00:21:34,320 We've done 2 Utemco pieces. What was your top take or one or 412 00:21:34,320 --> 00:21:37,520 two takes from those pieces? The biggest take away I've 413 00:21:37,520 --> 00:21:40,560 found, at least from, you know, doing this last year and then 414 00:21:40,560 --> 00:21:44,160 now doing it again a year later to look at this data is really 415 00:21:44,160 --> 00:21:48,200 just that the market downturn like is much more visible now. 416 00:21:48,240 --> 00:21:50,880 And obviously you can see exceptions, but it's people who 417 00:21:50,880 --> 00:21:53,400 are doing things a little bit differently, like thrive, like 418 00:21:53,400 --> 00:21:56,800 getting into these growth stage rounds and not performing like a 419 00:21:56,800 --> 00:21:58,960 traditional venture firm, if you will. 420 00:22:00,000 --> 00:22:02,520 And otherwise like, you know, kind of doing these more growth 421 00:22:02,520 --> 00:22:05,760 stage deals to where it their returns look better faster, you 422 00:22:05,760 --> 00:22:08,640 know, I. Think your take, your take was 423 00:22:08,640 --> 00:22:12,960 just going to be Union Square Ventures is a singular firm. 424 00:22:13,040 --> 00:22:15,080 I don't know. You look, you look at the Utemco 425 00:22:15,080 --> 00:22:18,200 performance and they're just like they have so many good 426 00:22:18,200 --> 00:22:22,080 funds like you just wrote about Blue Yard, a great firm that has 427 00:22:22,080 --> 00:22:25,360 a really strong fund that's competitive with USVI. 428 00:22:25,760 --> 00:22:27,840 Think they were in the creator of Filecoin. 429 00:22:28,600 --> 00:22:30,320 Kudos. And it seems like they're doing 430 00:22:30,320 --> 00:22:32,800 well their 20. 16 fund is great. Yeah, right. 431 00:22:32,960 --> 00:22:34,840 It's. But USB, it's just like hit 432 00:22:34,840 --> 00:22:38,920 after hit, like, yeah, 2004 is the best one on our chart. 433 00:22:39,240 --> 00:22:43,920 I think what USB opportunity from 2011, looking at it might 434 00:22:43,920 --> 00:22:47,240 be the second best on our chart. You know, it's just like they 435 00:22:47,240 --> 00:22:50,160 are, yeah. They're pretty strong. 436 00:22:50,200 --> 00:22:53,600 It'll be great, yeah. Well, I was gonna say also I 437 00:22:53,640 --> 00:22:56,640 will say New York based IA Ventures also incredible 438 00:22:56,640 --> 00:23:00,200 returns, not quite to the level of USV on the charts per 439 00:23:00,200 --> 00:23:04,320 Utemco's internal calculations on IRR, but still really 440 00:23:04,320 --> 00:23:06,360 consistently up there with each of their funds. 441 00:23:06,560 --> 00:23:09,360 And so maybe this is big New York, maybe that's my other 442 00:23:09,360 --> 00:23:10,640 take. You know, I'm actually, I'm 443 00:23:10,640 --> 00:23:15,120 calling it now bullish on New York, IA Ventures and USV most 444 00:23:15,120 --> 00:23:18,160 steadiest, biggest returns in the UTEMCO portfolio. 445 00:23:18,960 --> 00:23:24,040 Turns out financial mind is still valuable when making 446 00:23:24,040 --> 00:23:28,800 investments and not just like, yeah, well, yeah, definitely 447 00:23:28,800 --> 00:23:32,280 worth subscribing. Yeah, check it out, Check it 448 00:23:32,280 --> 00:23:34,240 out. Cool, cool. 449 00:23:34,400 --> 00:23:36,080 Our deal of the week. Deal of the Week What's? 450 00:23:36,080 --> 00:23:38,880 Our deal of the week. The deal of the Week exciting 451 00:23:38,880 --> 00:23:42,600 deal of the week 11 Labs finally closed its own funding. 452 00:23:42,600 --> 00:23:45,240 You know, we're big fans over here, Voice AI, although 453 00:23:45,240 --> 00:23:47,360 different voice AI companies to be fair, but. 454 00:23:47,440 --> 00:23:50,080 No, I think volley volley, you know, our partner is Ruble 455 00:23:50,080 --> 00:23:52,320 Valley of my good friends voice games. 456 00:23:52,320 --> 00:23:55,040 I think they use 11 labs or they're I think they're bullish 457 00:23:55,040 --> 00:23:56,280 on them. I don't think it's. 458 00:23:56,280 --> 00:23:58,160 Awesome. It's then it's all love. 459 00:23:58,160 --> 00:24:04,520 Fantastic 11 labs raised 180 million series C round which 460 00:24:04,800 --> 00:24:09,520 raised its valuation big up round to 3.3 billion and it was 461 00:24:09,520 --> 00:24:15,000 Co led by iconic growth and A16Z yeah voice AI is hot lots of 462 00:24:15,000 --> 00:24:19,000 market opportunity. 11 labs is growing quickly, doing well. 463 00:24:19,080 --> 00:24:20,960 Big up round. That's our show. 464 00:24:20,960 --> 00:24:25,440 We had what, a couple weeks ago? Red Note taking over the US App 465 00:24:25,440 --> 00:24:27,400 Store. Now it's deep seek. 466 00:24:27,400 --> 00:24:30,760 Deep seek what? What Chinese app are we bringing 467 00:24:30,760 --> 00:24:33,720 to you next week on the show? We'll wait and see. 468 00:24:33,880 --> 00:24:35,800 All right, see you next week. See ya.