1 00:00:02,560 --> 00:00:13,560 Speaker 1: Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is alive 2 00:00:13,600 --> 00:00:17,400 Speaker 1: from coast to coast with Caroline Hide in New York 3 00:00:17,720 --> 00:00:20,040 Speaker 1: and ever though in sentences. 4 00:00:19,600 --> 00:00:24,560 Speaker 2: Go this is Bloomberg Tech coming up. Wall Street can't 5 00:00:24,560 --> 00:00:28,280 Speaker 2: get enough of SpaceX. With demand from big institutional investors 6 00:00:28,440 --> 00:00:32,480 Speaker 2: and the biggest IPO in history, way over subscribed, class. 7 00:00:32,159 --> 00:00:35,519 Speaker 3: Google backstops and thropping data centers, a Silicon valley raises 8 00:00:35,600 --> 00:00:38,839 Speaker 3: to build AI infrastructure with ever more intertwined deals. 9 00:00:38,840 --> 00:00:41,400 Speaker 2: An Oracle reports after the closing bell, it's a race 10 00:00:41,520 --> 00:00:44,720 Speaker 2: between building data centers and booking AI cloud. 11 00:00:44,479 --> 00:00:49,280 Speaker 4: Revenues, AI AI AI and space SpaceX's over subscribed IPO 12 00:00:49,440 --> 00:00:52,840 Speaker 4: is where we have to start ed because the geographical 13 00:00:53,200 --> 00:00:54,920 Speaker 4: reach of the level of demand. 14 00:00:55,280 --> 00:00:58,080 Speaker 5: We've been mesmerized by this record breaking. 15 00:00:57,920 --> 00:00:59,960 Speaker 2: Yeah, it's out of this world. I don't apologize for 16 00:01:00,080 --> 00:01:01,960 Speaker 2: that for one bit. The state of players this right. 17 00:01:02,160 --> 00:01:05,319 Speaker 2: The order book for institution investors closes four pm Eastern today, 18 00:01:05,560 --> 00:01:09,200 Speaker 2: and as we've reported, there are several long only asset 19 00:01:09,280 --> 00:01:11,959 Speaker 2: managers basically that want ten billion dollars worth of shares. 20 00:01:12,000 --> 00:01:13,399 Speaker 5: It's seventy five billion dollars worth. 21 00:01:13,520 --> 00:01:15,760 Speaker 2: So somebody is going to miss out. Now the retail 22 00:01:15,800 --> 00:01:19,360 Speaker 2: investor can still place orders I think through Thursday on 23 00:01:19,400 --> 00:01:22,360 Speaker 2: whatever platforms are available, but they're not guaranteed to get 24 00:01:22,360 --> 00:01:24,559 Speaker 2: hold of those shares either. So that's the state of play. 25 00:01:24,640 --> 00:01:26,679 Speaker 2: And believe it or not, there is a roadshow happening 26 00:01:26,680 --> 00:01:27,479 Speaker 2: in the background, and. 27 00:01:27,440 --> 00:01:29,800 Speaker 5: We're learning ever more on that roadshow. That's the entire point. 28 00:01:29,840 --> 00:01:32,840 Speaker 3: We're understanding the transparency, the business model. We're learning much 29 00:01:32,840 --> 00:01:34,840 Speaker 3: about those orbital data centers. 30 00:01:34,920 --> 00:01:36,640 Speaker 2: Yeah, I think the focus in the pitch has still 31 00:01:36,640 --> 00:01:39,360 Speaker 2: been let us explain orbital data centers. That brings us 32 00:01:39,400 --> 00:01:42,479 Speaker 2: to today's big number, two hundred and fifty billion. That's 33 00:01:42,520 --> 00:01:46,120 Speaker 2: a total amount of SpaceX IPO orders we've reported this morning, 34 00:01:46,200 --> 00:01:50,120 Speaker 2: one to five billion of which is coming from Saudi Qate, 35 00:01:50,320 --> 00:01:53,760 Speaker 2: other Middle East funds, sovereign funds. That's according to sources. 36 00:01:54,000 --> 00:01:57,000 Speaker 2: That's the absolute latest. Joining us now to talk all 37 00:01:57,040 --> 00:02:01,240 Speaker 2: things SpaceX. It's IPO. Also the general landscape Peter Singlehurst, 38 00:02:01,440 --> 00:02:03,520 Speaker 2: head of private Companies are Bailey giff and we just know, 39 00:02:03,920 --> 00:02:06,200 Speaker 2: you know SpaceX is a really important investment for you 40 00:02:06,320 --> 00:02:10,760 Speaker 2: guys prior to the offering. Let's start there. You know, 41 00:02:10,919 --> 00:02:15,200 Speaker 2: what does this the biggest IPO in history represent to 42 00:02:15,320 --> 00:02:19,080 Speaker 2: you and to the firm and to I guess support 43 00:02:19,120 --> 00:02:22,000 Speaker 2: the thesis when you first made the investment way back when. 44 00:02:24,320 --> 00:02:27,400 Speaker 6: I think that the SpaceX IPO needs to be seen 45 00:02:27,680 --> 00:02:30,840 Speaker 6: as the culmination of a trend which has been playing 46 00:02:30,880 --> 00:02:34,840 Speaker 6: out now for fifteen years or longer, of companies staying 47 00:02:35,000 --> 00:02:37,400 Speaker 6: private for longer. And this is something that we started 48 00:02:37,440 --> 00:02:40,880 Speaker 6: to see in twenty twelve when we first started investing 49 00:02:41,440 --> 00:02:45,440 Speaker 6: in private companies. Now, we didn't think the companies would 50 00:02:45,520 --> 00:02:48,600 Speaker 6: get this big and stay private this long, But here 51 00:02:48,639 --> 00:02:52,040 Speaker 6: we are with, you know, SpaceX going public at something 52 00:02:52,080 --> 00:02:55,680 Speaker 6: like a one point eight trillion dollar valuation. That's nine 53 00:02:55,800 --> 00:02:59,600 Speaker 6: hundred times larger and more valuable than Tesla was when 54 00:02:59,600 --> 00:03:02,720 Speaker 6: it went public in twenty twelve. So, on the one hand, 55 00:03:02,760 --> 00:03:05,239 Speaker 6: this is a story of a truly exceptional company which 56 00:03:05,320 --> 00:03:09,800 Speaker 6: has compounded its growth at a very high rate. On 57 00:03:09,840 --> 00:03:11,959 Speaker 6: the other hand, it's a story of a bigger structural 58 00:03:12,000 --> 00:03:14,480 Speaker 6: trend of companies staying private longer and more and more 59 00:03:14,480 --> 00:03:17,000 Speaker 6: return to accruing within the high growth private. 60 00:03:16,720 --> 00:03:18,240 Speaker 5: Market and peter to that end. 61 00:03:18,320 --> 00:03:21,240 Speaker 3: When you think about Tesla after it's gone public, it 62 00:03:21,280 --> 00:03:25,000 Speaker 3: was a volatile ride, but it's twenty five thousand percent 63 00:03:25,160 --> 00:03:29,120 Speaker 3: higher than when it listed, And so will we see 64 00:03:29,280 --> 00:03:31,680 Speaker 3: a level of returns do you think in the public 65 00:03:31,720 --> 00:03:34,360 Speaker 3: market or does that have to be in some ways 66 00:03:34,400 --> 00:03:38,160 Speaker 3: pushed against the meat and bones of returns going to 67 00:03:38,200 --> 00:03:40,000 Speaker 3: have happened to private investors. 68 00:03:40,960 --> 00:03:43,680 Speaker 6: I think it's mathematically it's very hard to see how 69 00:03:43,760 --> 00:03:47,360 Speaker 6: you could see SpaceX delivering the same kind of returns 70 00:03:47,400 --> 00:03:50,000 Speaker 6: as a public company as Tesla did. But I think 71 00:03:50,000 --> 00:03:53,800 Speaker 6: what this speaks to is a requirement for investors to 72 00:03:53,920 --> 00:03:58,040 Speaker 6: have exposure to growth in both the private and the 73 00:03:58,080 --> 00:04:01,840 Speaker 6: public markets. Has been set up to almost divide these 74 00:04:02,200 --> 00:04:04,400 Speaker 6: things and say there's kind of private growth and there's 75 00:04:04,440 --> 00:04:07,320 Speaker 6: public growth, and these things are different, and we've taken 76 00:04:07,360 --> 00:04:09,640 Speaker 6: a different approach. We've sort of taken the view that actually, 77 00:04:09,640 --> 00:04:11,400 Speaker 6: if you want to do growth equity investing, you want 78 00:04:11,440 --> 00:04:13,400 Speaker 6: to do it properly, you have to do it in 79 00:04:13,440 --> 00:04:15,680 Speaker 6: the private markets, and you have to do it in 80 00:04:15,720 --> 00:04:19,440 Speaker 6: the public markets. And what our clients are and beneficiaries 81 00:04:19,440 --> 00:04:21,920 Speaker 6: who are predominantly pension funds, what they need and what 82 00:04:21,960 --> 00:04:24,480 Speaker 6: they ask from us is that we give them exposures 83 00:04:24,520 --> 00:04:27,479 Speaker 6: to the world's best growth stage companies starting in the 84 00:04:27,480 --> 00:04:30,000 Speaker 6: private markets. Earning the returns that we can generate there, 85 00:04:30,400 --> 00:04:32,360 Speaker 6: and then also only those in the public markets from 86 00:04:32,400 --> 00:04:34,240 Speaker 6: within our public funds to make sure that they're still 87 00:04:34,320 --> 00:04:37,919 Speaker 6: capturing that growth even once companies transition into the public markets. 88 00:04:38,120 --> 00:04:40,599 Speaker 2: Peter, I think it's important to pose the question why 89 00:04:40,720 --> 00:04:43,640 Speaker 2: is SpaceX going public? And when Elon Musk spoke to 90 00:04:43,720 --> 00:04:46,680 Speaker 2: Jamie Diamond, he eventually got to the answer, which is 91 00:04:47,200 --> 00:04:51,039 Speaker 2: they need capital for this growth phase. But what we 92 00:04:51,080 --> 00:04:53,840 Speaker 2: are seeing outside of just this fixation on IPOs is 93 00:04:53,880 --> 00:04:58,200 Speaker 2: a race for capital through equity. How comfortable do you 94 00:04:58,360 --> 00:05:02,359 Speaker 2: feel as a firm at Bailey gifed, whatever mechanism it 95 00:05:02,480 --> 00:05:05,520 Speaker 2: is raising money at that volume, but it basically then 96 00:05:05,600 --> 00:05:09,160 Speaker 2: goes directly into capital expenditure. That's what's happening here. 97 00:05:11,600 --> 00:05:13,800 Speaker 6: You want to invest in companies that are able to 98 00:05:13,839 --> 00:05:16,560 Speaker 6: deploy capital or high rates of return. So I don't 99 00:05:16,560 --> 00:05:19,679 Speaker 6: think there's anything wrong, per se in investing in capital 100 00:05:19,720 --> 00:05:22,520 Speaker 6: intensive businesses. In fact, what you want as a company 101 00:05:22,520 --> 00:05:25,200 Speaker 6: that can deploy large amounts of capital, but where you 102 00:05:25,240 --> 00:05:27,800 Speaker 6: can earn high returns on that capital. And ultimately that's 103 00:05:27,839 --> 00:05:30,320 Speaker 6: what separates a good business from a bad business. It's 104 00:05:30,360 --> 00:05:33,040 Speaker 6: return on equity. And so when we're looking at a company, 105 00:05:33,040 --> 00:05:35,800 Speaker 6: whether it's SpaceX or Anthropic, what any other company that 106 00:05:35,839 --> 00:05:38,760 Speaker 6: we invest in privately or publicly. Ultimately, what we're asking 107 00:05:38,839 --> 00:05:41,880 Speaker 6: is how do you get to high returns on equity? 108 00:05:41,920 --> 00:05:43,800 Speaker 6: And it's building those thesis that then leads us to 109 00:05:43,880 --> 00:05:47,560 Speaker 6: invest in companies. And in the case of SpaceX, increasingly 110 00:05:47,560 --> 00:05:50,279 Speaker 6: that thesis is going to have to rely on AI. 111 00:05:50,920 --> 00:05:53,960 Speaker 6: They've shown that they can invest capital or high rates 112 00:05:53,960 --> 00:05:57,440 Speaker 6: of return in rockets in starlink, and of course the 113 00:05:57,480 --> 00:05:59,960 Speaker 6: next leg of that is going to be in AI 114 00:06:00,040 --> 00:06:02,960 Speaker 6: data center build out, quite possibly in space but that's. 115 00:06:02,800 --> 00:06:06,400 Speaker 3: Where it becomes so fascinating, particularly Peter for Bailey Gifford, 116 00:06:06,440 --> 00:06:09,760 Speaker 3: which in the private markets backspace X on a thesis 117 00:06:09,760 --> 00:06:12,840 Speaker 3: of SPACE, Backtindthropic on a thesis of AI, and now 118 00:06:12,880 --> 00:06:16,200 Speaker 3: they're all overlapping in terms of business models. What is 119 00:06:16,240 --> 00:06:19,520 Speaker 3: your perspective of commoditization or a winner takes all or 120 00:06:19,640 --> 00:06:20,960 Speaker 3: is there room for all. 121 00:06:20,839 --> 00:06:23,240 Speaker 5: Of these giant AI players? Do we winning in. 122 00:06:23,240 --> 00:06:25,440 Speaker 3: The technology as well as perhaps in the public markets. 123 00:06:26,800 --> 00:06:29,640 Speaker 6: I think what your question gets to is this very 124 00:06:29,680 --> 00:06:35,560 Speaker 6: important question of where does value accrue in the AI stack? Now, 125 00:06:35,600 --> 00:06:37,320 Speaker 6: hopefully lots of value is going to accrue to the 126 00:06:37,440 --> 00:06:41,040 Speaker 6: end customers. That has to happen. Historically, we've seen value 127 00:06:41,040 --> 00:06:44,560 Speaker 6: accruing to the chip manufacturers, initially with Nvidio, but now 128 00:06:44,560 --> 00:06:47,480 Speaker 6: increasing league so memory manufacturers. But I think what we're 129 00:06:47,480 --> 00:06:49,320 Speaker 6: also starting to see is value a crew at the 130 00:06:49,320 --> 00:06:53,520 Speaker 6: foundational model level. And I suppose with the Grock acquisition, 131 00:06:54,440 --> 00:06:56,840 Speaker 6: SpaceX is making a bet not just on the foundational 132 00:06:56,839 --> 00:06:58,960 Speaker 6: model level, but also on the infrastructure level. And I 133 00:06:59,000 --> 00:07:01,320 Speaker 6: think what we're seeing with the deal that they did 134 00:07:01,360 --> 00:07:03,640 Speaker 6: recently with Anthropic is that they have options in terms 135 00:07:03,680 --> 00:07:07,039 Speaker 6: of how they can monetize in the AI transition, both 136 00:07:07,200 --> 00:07:11,600 Speaker 6: through their own models, but also importantly through the infrastructure itself. 137 00:07:11,680 --> 00:07:14,160 Speaker 2: I have so many questions about this. You know, let's 138 00:07:14,160 --> 00:07:17,200 Speaker 2: be honest. The hedge that SpaceX has put in place 139 00:07:17,240 --> 00:07:19,840 Speaker 2: in the interim is to become a hyperscaler and sell 140 00:07:19,880 --> 00:07:23,760 Speaker 2: compute played a blinder with that. We got the design, 141 00:07:24,080 --> 00:07:26,600 Speaker 2: or at least the renderings of Orbital Data Center. I 142 00:07:26,640 --> 00:07:27,800 Speaker 2: think the team are going to put them up on 143 00:07:27,800 --> 00:07:30,840 Speaker 2: the screen now in that presentation that Elon must make, 144 00:07:31,160 --> 00:07:34,000 Speaker 2: like there's the body, there's the solar arrays, there's the radiator, 145 00:07:35,160 --> 00:07:38,320 Speaker 2: which part of the thesis, Peter is most important to you, right, 146 00:07:38,680 --> 00:07:42,080 Speaker 2: It is a long way from the tam of twenty 147 00:07:42,120 --> 00:07:45,160 Speaker 2: six point five trillion that they're basically packaging it as 148 00:07:45,280 --> 00:07:49,120 Speaker 2: enterprise AI and in the interim, this plan for orbital 149 00:07:49,200 --> 00:07:51,320 Speaker 2: data center like it needs to work. That's what they're 150 00:07:51,360 --> 00:07:52,560 Speaker 2: telling people on the road show. 151 00:07:53,800 --> 00:07:58,080 Speaker 6: So there's absolutely no question that the orbital Data center 152 00:07:59,240 --> 00:08:04,840 Speaker 6: strategic that they're making widens the range of outcomes for SpaceX. 153 00:08:05,240 --> 00:08:07,320 Speaker 6: On the one hand, if it works, it increases the 154 00:08:07,320 --> 00:08:10,080 Speaker 6: potential upsides for the business. On the other hand, if 155 00:08:10,120 --> 00:08:12,240 Speaker 6: this doesn't work, it's going to increase the downside for 156 00:08:12,320 --> 00:08:15,280 Speaker 6: this company. And investing ultimately is about ranges of outcomes. 157 00:08:15,320 --> 00:08:18,320 Speaker 6: It's about probabilities, and it's about payoffs in those range 158 00:08:18,320 --> 00:08:21,800 Speaker 6: of outcomes. What we've seen with SpaceX over the years 159 00:08:21,880 --> 00:08:26,680 Speaker 6: is that they have continuously tested and validated a series 160 00:08:26,680 --> 00:08:30,280 Speaker 6: of outlandish hypotheses. The very notion of the business starting 161 00:08:30,280 --> 00:08:32,640 Speaker 6: off as a private rocket company was a self an 162 00:08:32,640 --> 00:08:36,040 Speaker 6: outlandish hypothesis that they've validated. Then the idea that you 163 00:08:36,080 --> 00:08:39,600 Speaker 6: could have reusable rockets was also an outlanded hypothesis, and 164 00:08:39,640 --> 00:08:43,480 Speaker 6: they validated it. They did the same with starlink with 165 00:08:44,600 --> 00:08:47,840 Speaker 6: satellite based broadband. They've done the same with rockets on 166 00:08:47,920 --> 00:08:51,599 Speaker 6: the scale of Starship and the orbital data centillate is 167 00:08:51,640 --> 00:08:55,320 Speaker 6: that is the next hypothesis that they are seeking to validate. 168 00:08:55,480 --> 00:08:58,160 Speaker 6: But everybody should be totally aware of the risks that 169 00:08:58,200 --> 00:09:00,679 Speaker 6: are involved in this. It is unproven. In the event 170 00:09:00,720 --> 00:09:03,640 Speaker 6: that they prove it, the payoffs will be large. But 171 00:09:03,679 --> 00:09:05,480 Speaker 6: as we've already touched on the amount of cattle that 172 00:09:05,559 --> 00:09:07,320 Speaker 6: is going into this means that the event that they 173 00:09:07,320 --> 00:09:09,800 Speaker 6: don't validate it, it's going to increase the scope of 174 00:09:09,840 --> 00:09:12,120 Speaker 6: downside in the investments as well, and investors just need 175 00:09:12,160 --> 00:09:14,560 Speaker 6: to understand the range of outcomes and the payoffs that 176 00:09:14,600 --> 00:09:15,120 Speaker 6: go with that. 177 00:09:15,559 --> 00:09:18,880 Speaker 3: Can I ask about payoffs, Peter, because I don't want 178 00:09:18,920 --> 00:09:20,760 Speaker 3: to go into the granularity of how much SpaceX E 179 00:09:20,760 --> 00:09:22,280 Speaker 3: suppose you have, et cetera, but how long do you 180 00:09:22,280 --> 00:09:25,320 Speaker 3: think you'll hold it and how much do you think 181 00:09:25,600 --> 00:09:27,719 Speaker 3: it's a warrior or an anxiety that all these other 182 00:09:27,720 --> 00:09:30,880 Speaker 3: big public companies are selling equity into this market at 183 00:09:30,880 --> 00:09:33,880 Speaker 3: the same time Alphabet trying to fund its own capex 184 00:09:34,000 --> 00:09:36,199 Speaker 3: in the equity market, Meta might be doing as well, 185 00:09:36,240 --> 00:09:38,240 Speaker 3: And does that take oxygen out the room. 186 00:09:39,800 --> 00:09:42,280 Speaker 6: I think that's probably part of the thinking that's going 187 00:09:42,320 --> 00:09:44,880 Speaker 6: on for these different companies trying to raise these large 188 00:09:44,880 --> 00:09:46,760 Speaker 6: amounts of cattle. They're sort of trying to soak up 189 00:09:46,760 --> 00:09:51,680 Speaker 6: what available castle there is. But to your first question, 190 00:09:52,920 --> 00:09:54,839 Speaker 6: different funds within Bailey Gifford are going to be in 191 00:09:54,920 --> 00:09:58,400 Speaker 6: very different positions. For those funds that have owned SpaceX 192 00:09:58,440 --> 00:10:01,200 Speaker 6: since twenty eighteen, since it was a thirty billion dollar company, 193 00:10:01,320 --> 00:10:05,960 Speaker 6: those funds have very very large exposure, large positions in SpaceX. 194 00:10:06,240 --> 00:10:08,600 Speaker 6: Now it might make sense post lock up for those 195 00:10:08,600 --> 00:10:11,040 Speaker 6: funds to start selling down, even if they want to 196 00:10:11,080 --> 00:10:16,160 Speaker 6: maintain a meaningful exposure, because ultimately, we are beholding to 197 00:10:16,160 --> 00:10:18,560 Speaker 6: our clients and we have to provide them with a 198 00:10:18,640 --> 00:10:21,560 Speaker 6: level of diversification within their funds. And then, of course 199 00:10:21,600 --> 00:10:24,800 Speaker 6: funds that don't own it, funds that are solely public funds, 200 00:10:24,960 --> 00:10:26,720 Speaker 6: they then faced with a question of whether to buy 201 00:10:26,800 --> 00:10:29,760 Speaker 6: it for their funds. So it might well be that 202 00:10:29,800 --> 00:10:32,840 Speaker 6: you see different funds within Bailey Gifford doing different things 203 00:10:32,960 --> 00:10:34,800 Speaker 6: over the coming months, and that will be a function 204 00:10:34,840 --> 00:10:37,360 Speaker 6: of the history and the portfolio context. I think there's 205 00:10:37,520 --> 00:10:41,440 Speaker 6: universe agreement that SpaceX has been an exceptional company. The 206 00:10:41,480 --> 00:10:44,240 Speaker 6: real question from here is what is the right price 207 00:10:44,679 --> 00:10:48,360 Speaker 6: and what is the right position size in SpaceX. 208 00:10:48,320 --> 00:10:51,360 Speaker 3: From thirty billion to potentially one point eight trillion this 209 00:10:51,400 --> 00:10:54,600 Speaker 3: week Peter Senglehurst A Bailey Gifford A joy to have 210 00:10:54,720 --> 00:10:57,880 Speaker 3: you on about all things SpaceX and the IPO landscape. 211 00:10:57,960 --> 00:11:01,400 Speaker 3: More broadly, let's get though, also to political tensions. They 212 00:11:01,440 --> 00:11:04,720 Speaker 3: are continuing to whipsaw markets. We are down a percentage 213 00:11:04,760 --> 00:11:06,640 Speaker 3: point again on the NASA one hundred s and p 214 00:11:06,800 --> 00:11:09,880 Speaker 3: is under pressure. You're seeing a really hardware of by 215 00:11:09,880 --> 00:11:12,480 Speaker 3: two percent if you're looking at the semiconductor index. President 216 00:11:12,480 --> 00:11:15,680 Speaker 3: Trump is saying that Iran would pay the price. 217 00:11:15,840 --> 00:11:17,400 Speaker 5: For delaying peace negotiations. 218 00:11:17,480 --> 00:11:19,280 Speaker 3: Let's get you up to speakably the most Tyler Kendall, 219 00:11:19,320 --> 00:11:20,400 Speaker 3: the latest sim of the White House. 220 00:11:20,440 --> 00:11:22,960 Speaker 7: What do we need to know, hey, Caroline, Well, at 221 00:11:23,000 --> 00:11:26,320 Speaker 7: this point, President Trump is renewing his threat, really just 222 00:11:26,480 --> 00:11:30,040 Speaker 7: underscoring that this White House has mounting frustration with the 223 00:11:30,120 --> 00:11:33,760 Speaker 7: ongoing negotiations, as the US has repeatedly maintained that they 224 00:11:33,760 --> 00:11:38,400 Speaker 7: were trying to prioritize a diplomatic solution to end the conflict. Now, 225 00:11:38,400 --> 00:11:42,280 Speaker 7: President Trump's remarks aren't totally clear if this means that 226 00:11:42,280 --> 00:11:45,400 Speaker 7: we're going to see an end to the ceasefire agreement 227 00:11:45,520 --> 00:11:47,920 Speaker 7: after we saw the worst flare up in fighting between 228 00:11:48,080 --> 00:11:51,200 Speaker 7: the sides just overnight with both the US and Iran 229 00:11:51,280 --> 00:11:56,320 Speaker 7: exchanging strikes after an American military helicopter was shot down. 230 00:11:56,600 --> 00:12:00,920 Speaker 7: After the strike from the US then on Iranian military assets, 231 00:12:01,000 --> 00:12:04,200 Speaker 7: we saw Iran put forward some strikes and attempts to 232 00:12:04,280 --> 00:12:07,720 Speaker 7: hit American military assets. It's really been escalating from here. 233 00:12:07,760 --> 00:12:11,000 Speaker 7: But our own analysts at Bloomberg Economics, perhaps this is 234 00:12:11,040 --> 00:12:14,079 Speaker 7: a bid to escalate in a bid to de escalate. 235 00:12:14,360 --> 00:12:17,480 Speaker 7: In one positive sign for the negotiation front, Irani and 236 00:12:17,520 --> 00:12:20,400 Speaker 7: state media reported within the last hour that a Katari 237 00:12:20,559 --> 00:12:23,520 Speaker 7: delegation has landed in Tehran in a bid to keep 238 00:12:23,559 --> 00:12:26,240 Speaker 7: diplomacy on track. But Ed and Caroline, I want to 239 00:12:26,360 --> 00:12:29,440 Speaker 7: highlight this renewed risk that we are seeing moments ago 240 00:12:29,760 --> 00:12:33,040 Speaker 7: flashing across the Bloomberg terminal. India is now condemning an 241 00:12:33,040 --> 00:12:37,320 Speaker 7: apparent attack on a commercial vessel near the Strait of 242 00:12:37,320 --> 00:12:40,600 Speaker 7: her Moves off the coast of Oman. So definitely still 243 00:12:40,760 --> 00:12:43,760 Speaker 7: very high intentions contributing to the situation that we're seeing 244 00:12:43,840 --> 00:12:44,400 Speaker 7: on the ground. 245 00:12:45,559 --> 00:12:48,040 Speaker 2: Goodvoks Tyler Kendall, Thank you very much. So coming up, 246 00:12:48,080 --> 00:12:51,400 Speaker 2: Google steps up to backstop a massive thirty five billion 247 00:12:51,440 --> 00:12:54,360 Speaker 2: dollar financing deal from Propic. We had the details. Next, 248 00:12:54,480 --> 00:13:03,240 Speaker 2: this is Bloomberg Tech. Super Micro is targeting public markets 249 00:13:03,240 --> 00:13:05,920 Speaker 2: for a massive seven billion dollar equity raise. Sales to 250 00:13:06,000 --> 00:13:09,800 Speaker 2: super Micro servers fitted with video chips have surged for 251 00:13:09,960 --> 00:13:13,160 Speaker 2: AI workloads. The server manufacturer is moving quickly to fund 252 00:13:13,200 --> 00:13:16,560 Speaker 2: a staggering thirty nine billion dollars in orders, using the 253 00:13:16,600 --> 00:13:19,800 Speaker 2: fresh cash injection to pay for the equipment needed to 254 00:13:19,840 --> 00:13:23,760 Speaker 2: make the servers. That's greasing the wheels. Google is backstopping 255 00:13:23,800 --> 00:13:27,319 Speaker 2: a massive thirty five billion dollar financing deal from Fropic, 256 00:13:27,600 --> 00:13:30,480 Speaker 2: The creator of Claud, is leasing AI chips across five 257 00:13:30,760 --> 00:13:33,760 Speaker 2: different data centers with help from the long established tech 258 00:13:33,800 --> 00:13:37,800 Speaker 2: giant Doomscott Carpenter joins us now to break down the mechanics, like, 259 00:13:38,280 --> 00:13:41,839 Speaker 2: you know, to our audience, what does backstopping mean? I 260 00:13:41,880 --> 00:13:44,559 Speaker 2: think what we're saying is guaranteeing the funds for it 261 00:13:44,920 --> 00:13:46,360 Speaker 2: and the event that something goes wrong. 262 00:13:46,440 --> 00:13:50,360 Speaker 8: But but just go with the basics, right, So, first 263 00:13:50,400 --> 00:13:55,240 Speaker 8: of all, Broadcom is providing a huge guarantee on the 264 00:13:55,440 --> 00:14:00,439 Speaker 8: chips themselves, as the biggest part of the thirty five years. Yes, 265 00:14:00,480 --> 00:14:03,120 Speaker 8: these are Google's TPUs that are going to be involved, 266 00:14:03,600 --> 00:14:06,319 Speaker 8: So Broadcom is backstopping the debt itself. 267 00:14:07,080 --> 00:14:09,000 Speaker 2: Now, the chips, when. 268 00:14:08,880 --> 00:14:12,440 Speaker 8: They are delivered and they have to be manufactured, are 269 00:14:12,520 --> 00:14:15,400 Speaker 8: going to be used in these five data centers that 270 00:14:15,600 --> 00:14:20,080 Speaker 8: we identify in the story. The leases on those five 271 00:14:20,200 --> 00:14:24,200 Speaker 8: data centers are backstopped by Google. So you could think 272 00:14:24,200 --> 00:14:28,560 Speaker 8: of it as two different forms of guarantees being involved 273 00:14:29,040 --> 00:14:32,240 Speaker 8: in this to put together this deal. There's the Broadcom 274 00:14:32,280 --> 00:14:34,080 Speaker 8: one and there's the Google ones underneath. 275 00:14:34,280 --> 00:14:37,880 Speaker 3: So in a way, Broadcom's saying almost Alphabet's going to 276 00:14:37,880 --> 00:14:40,560 Speaker 3: get its money from Anthropic for buying the chips. 277 00:14:40,800 --> 00:14:41,280 Speaker 5: Is that right? 278 00:14:41,400 --> 00:14:44,080 Speaker 3: Meanwhile, who's getting the money for the leases and who 279 00:14:44,160 --> 00:14:46,480 Speaker 3: therefore is Alphabet saying like, you're good for the money, 280 00:14:46,480 --> 00:14:48,920 Speaker 3: don't worry. Is that the people actually constructing the data 281 00:14:48,920 --> 00:14:50,160 Speaker 3: centers owning the land. 282 00:14:50,600 --> 00:14:54,880 Speaker 8: It's the leases are to Fluid Stack, which is a 283 00:14:54,920 --> 00:14:58,680 Speaker 8: company that Anthropic has said it's been working with to 284 00:14:58,800 --> 00:15:03,080 Speaker 8: develop these data seen. So you see how it's it's complicated, right, 285 00:15:03,480 --> 00:15:04,400 Speaker 8: there's many. 286 00:15:04,400 --> 00:15:05,560 Speaker 2: See how it's complicated. 287 00:15:05,640 --> 00:15:08,480 Speaker 8: Yeah, yeah, I mean to pull off a deal of 288 00:15:08,520 --> 00:15:12,040 Speaker 8: this magnitude, which I think is the largest private credit 289 00:15:12,040 --> 00:15:16,160 Speaker 8: deal in history, definitely the largest chip deal. There's a 290 00:15:16,160 --> 00:15:18,560 Speaker 8: lot of moving pieces. One of the key things is 291 00:15:18,560 --> 00:15:20,840 Speaker 8: that these chips are not I mean, they need to 292 00:15:20,920 --> 00:15:21,720 Speaker 8: be created. 293 00:15:21,760 --> 00:15:27,200 Speaker 2: They don't exist right now. But yeah, there's a lot 294 00:15:27,280 --> 00:15:27,960 Speaker 2: that goes into this. 295 00:15:28,600 --> 00:15:33,880 Speaker 3: Who's the manufacturer of questions around Intel? VISs TSMC absolutely fascinating. 296 00:15:33,920 --> 00:15:36,440 Speaker 3: Scott Carpenter. He broke it down so clearly, we so 297 00:15:36,480 --> 00:15:39,400 Speaker 3: appreciate it. Meanwhile, let's turn our attention to soft Bank. 298 00:15:39,520 --> 00:15:42,080 Speaker 3: It's attempt to leverage its massive AI bets. It's hitting 299 00:15:42,200 --> 00:15:44,680 Speaker 3: a bit of a wall. Sources told Bloomberg that talks 300 00:15:44,680 --> 00:15:46,960 Speaker 3: are stalled with potential creditors to raise at least six 301 00:15:46,960 --> 00:15:50,200 Speaker 3: billion dollars from a margin loan backed by its opening 302 00:15:50,240 --> 00:15:53,360 Speaker 3: eye steak. So it's unclear why the pores comes just 303 00:15:53,400 --> 00:15:56,200 Speaker 3: weeks after soft Bank slashed it's fundraising target from ten 304 00:15:56,240 --> 00:15:59,080 Speaker 3: billion dollars, and soft Bank shares have tumbled nearly ten percent. 305 00:15:59,120 --> 00:16:01,960 Speaker 3: On the news today, A says the firm ways alternative 306 00:16:02,240 --> 00:16:04,400 Speaker 3: funding options coming up. 307 00:16:04,840 --> 00:16:05,760 Speaker 5: We are going to be speaking a. 308 00:16:05,840 --> 00:16:10,240 Speaker 3: Saphia Noble, Professor, director of the Center of Resilience and 309 00:16:10,280 --> 00:16:13,800 Speaker 3: Digital Justice, at the UCLA to discuss bias discrimination within 310 00:16:13,840 --> 00:16:14,520 Speaker 3: this world of AI. 311 00:16:14,560 --> 00:16:16,720 Speaker 5: We keep talking more on that next as a Blueberg tech. 312 00:16:24,040 --> 00:16:26,680 Speaker 3: This week, as Open AI filed its S one confidentially, 313 00:16:26,760 --> 00:16:29,600 Speaker 3: CEO Samaltman was also out with a sweeping long term 314 00:16:29,640 --> 00:16:33,160 Speaker 3: vision for generative AIS alignment with humanity and warned that 315 00:16:33,440 --> 00:16:37,600 Speaker 3: transformative technologies such as AI can concentrate power, stating Open 316 00:16:37,640 --> 00:16:40,400 Speaker 3: AIS quote clear ride about the risks as it aims 317 00:16:40,440 --> 00:16:43,680 Speaker 3: to build powerful systems that remain safe subject to human control. 318 00:16:43,720 --> 00:16:47,600 Speaker 3: But critics have long voiced concerns about AI risks such 319 00:16:47,600 --> 00:16:51,280 Speaker 3: as algorithmic bias in equality joining us now, Sofia Noble, 320 00:16:51,440 --> 00:16:55,800 Speaker 3: UCLA professor author of the acclaimed book Algorithms of Oppression 321 00:16:56,800 --> 00:16:59,800 Speaker 3: clear Ride, is that enough. Are we seeing some of 322 00:16:59,840 --> 00:17:04,200 Speaker 3: the risks being digested and answered for within these models? 323 00:17:04,640 --> 00:17:06,960 Speaker 9: I don't think so. I don't think we're anywhere near 324 00:17:08,160 --> 00:17:12,400 Speaker 9: a call for or an ability to realize safe AI. 325 00:17:13,359 --> 00:17:17,160 Speaker 9: What we see, in fact, are chapbot technologies and large 326 00:17:17,200 --> 00:17:21,000 Speaker 9: language models that for the most part don't have markets. 327 00:17:21,359 --> 00:17:24,560 Speaker 9: I mean, they were built for Corporate America to reduce 328 00:17:24,640 --> 00:17:27,760 Speaker 9: labor costs, but corporate America is moving away from them 329 00:17:27,800 --> 00:17:31,320 Speaker 9: because they're very expensive, they're not really reliable, they are 330 00:17:31,560 --> 00:17:35,160 Speaker 9: incredibly negatively impactful on the environment. 331 00:17:35,400 --> 00:17:37,360 Speaker 3: Moving away from them, you think corporate America is moving 332 00:17:37,359 --> 00:17:38,439 Speaker 3: away from lagenguge porp. 333 00:17:38,520 --> 00:17:38,760 Speaker 2: I do. 334 00:17:38,840 --> 00:17:42,200 Speaker 9: We've been seeing studies where companies are saying that it's 335 00:17:42,280 --> 00:17:46,639 Speaker 9: more expensive for them to use these chatbots because human 336 00:17:46,680 --> 00:17:50,200 Speaker 9: beings have to check the efficacy and the reliability. There 337 00:17:50,200 --> 00:17:54,520 Speaker 9: are so many errors, factual errors that are proliferating through 338 00:17:54,520 --> 00:17:58,600 Speaker 9: these technologies. So if the technology itself is that flawed 339 00:17:59,320 --> 00:18:02,960 Speaker 9: and it's being now on the public as some type 340 00:18:03,000 --> 00:18:06,119 Speaker 9: of solution, I think we're in trouble. And of course 341 00:18:06,240 --> 00:18:10,760 Speaker 9: we know that we have the racial bias, the gender bias, 342 00:18:11,040 --> 00:18:14,960 Speaker 9: the kind of geographic and political concerns about what comes 343 00:18:14,960 --> 00:18:18,840 Speaker 9: out of these technologies. I think that we are moving 344 00:18:18,880 --> 00:18:24,120 Speaker 9: into very dangerous territory trying to bolster our society on 345 00:18:24,359 --> 00:18:26,280 Speaker 9: large language models fascinatable. 346 00:18:26,560 --> 00:18:29,680 Speaker 2: The large body of your work looked at the data 347 00:18:29,720 --> 00:18:33,560 Speaker 2: issues for commercial search engines. Basically the net result is 348 00:18:34,119 --> 00:18:38,800 Speaker 2: that the search engines, as per your books title, reinforce racism. Yes, 349 00:18:39,520 --> 00:18:43,840 Speaker 2: what was the underlying issue in the search engine case study? 350 00:18:44,119 --> 00:18:47,480 Speaker 2: And what is different or the same about the large 351 00:18:47,520 --> 00:18:50,560 Speaker 2: language models that I think you're saying yield a similar 352 00:18:50,600 --> 00:18:51,560 Speaker 2: result they do. 353 00:18:52,280 --> 00:18:56,480 Speaker 9: So what we've seen over the last fifteen twenty years 354 00:18:56,640 --> 00:19:00,000 Speaker 9: is that all of the discrimination that's in our society, 355 00:19:00,400 --> 00:19:04,520 Speaker 9: all of the kind of stereotyping, all of the inequality, 356 00:19:05,160 --> 00:19:08,920 Speaker 9: just gets packaged up and then used to train models. 357 00:19:09,240 --> 00:19:11,000 Speaker 2: So it's within the data. 358 00:19:10,640 --> 00:19:13,480 Speaker 9: Within the data, but it's also the people who are 359 00:19:13,480 --> 00:19:16,960 Speaker 9: designing the models are really not aware. They don't understand 360 00:19:17,040 --> 00:19:21,320 Speaker 9: the kind of social, historical, economic processes. These are software 361 00:19:21,320 --> 00:19:23,640 Speaker 9: engineers who don't even think about they don't even ask 362 00:19:23,680 --> 00:19:26,480 Speaker 9: the kinds of questions that let's say, a sociologist like 363 00:19:26,520 --> 00:19:30,240 Speaker 9: I would ask. And so we have discriminatory data that 364 00:19:30,359 --> 00:19:34,680 Speaker 9: is training models. But what's different now is that these 365 00:19:34,800 --> 00:19:40,880 Speaker 9: models obfuscate the inequality. They appear to be factual and reliable. 366 00:19:41,160 --> 00:19:43,920 Speaker 9: And if you don't know, if you don't have deep expertise, 367 00:19:44,200 --> 00:19:45,880 Speaker 9: you're not going to know that the kinds of things 368 00:19:45,920 --> 00:19:49,640 Speaker 9: that are being served up in these products are actually faulty. 369 00:19:49,840 --> 00:19:55,040 Speaker 9: And imagine putting your whole business enterprise, your public institutions, 370 00:19:55,080 --> 00:19:59,240 Speaker 9: your schools, your libraries, making that the backbone. I mean 371 00:19:59,320 --> 00:20:02,160 Speaker 9: that is to me quite dangerous Sophia. 372 00:20:03,040 --> 00:20:05,800 Speaker 3: We can talk at length about the risks and the problems. 373 00:20:06,200 --> 00:20:08,479 Speaker 3: What about the solutions here because we saw but at 374 00:20:08,560 --> 00:20:10,400 Speaker 3: least two years ago, I think it was when alphabet 375 00:20:10,480 --> 00:20:13,520 Speaker 3: was struggling to ensure that some of the images AI 376 00:20:13,640 --> 00:20:18,080 Speaker 3: generated images didn't overcompensate for some of the worries about 377 00:20:18,480 --> 00:20:20,480 Speaker 3: racism and bias within the algorithm. 378 00:20:20,640 --> 00:20:23,879 Speaker 5: So what have you been done that works. Let's not 379 00:20:23,920 --> 00:20:26,840 Speaker 5: just beautify the problem, let's give us the solution. 380 00:20:27,560 --> 00:20:30,000 Speaker 9: Well, I think that we don't want to give up 381 00:20:30,240 --> 00:20:37,040 Speaker 9: what it means to have human expertise, human journalists, fact checkers, teachers, thinkers. 382 00:20:37,440 --> 00:20:39,560 Speaker 10: This is our most powerful asset. 383 00:20:39,680 --> 00:20:43,280 Speaker 9: These human beings are people, and we can't replace people 384 00:20:43,320 --> 00:20:46,320 Speaker 9: with these kinds of machines. So to me, you know, 385 00:20:46,880 --> 00:20:50,240 Speaker 9: having deep knowledge in the humanities and social sciences, these 386 00:20:50,240 --> 00:20:52,719 Speaker 9: are the things that are going to really be important 387 00:20:52,760 --> 00:20:56,840 Speaker 9: as we go forward in society. And we're over investing, 388 00:20:56,960 --> 00:20:59,199 Speaker 9: I think, in the wrong things. We need to be 389 00:20:59,280 --> 00:21:04,960 Speaker 9: investing in putting resources into pro social, pro rights respecting technology. 390 00:21:05,280 --> 00:21:09,320 Speaker 9: There's a whole world of small language models and different 391 00:21:09,400 --> 00:21:13,680 Speaker 9: kinds of very interesting kinds of technologies that women are 392 00:21:13,720 --> 00:21:17,040 Speaker 9: thinking about that people of color are working on and 393 00:21:17,080 --> 00:21:19,800 Speaker 9: these are the least invested in, but they are I 394 00:21:19,840 --> 00:21:23,800 Speaker 9: think the kinds of technologies that are going to help 395 00:21:23,920 --> 00:21:25,160 Speaker 9: us find a way forward. 396 00:21:25,240 --> 00:21:29,359 Speaker 2: Sophea, How conscious of and open about are the companies 397 00:21:29,760 --> 00:21:32,879 Speaker 2: on the issue, And you know, in research and writing 398 00:21:32,920 --> 00:21:34,720 Speaker 2: your book, but your ongoing work, how much do they 399 00:21:34,720 --> 00:21:35,760 Speaker 2: engage with you on it? 400 00:21:36,440 --> 00:21:40,320 Speaker 9: The companies for the most part want to deny, deny 401 00:21:40,920 --> 00:21:44,560 Speaker 9: the most dangerous dimensions of their products, and of course 402 00:21:44,640 --> 00:21:48,840 Speaker 9: they are only interested in regulation that they're writing. We've 403 00:21:48,920 --> 00:21:54,480 Speaker 9: just saw the landmark ruling against Meta, where they knew 404 00:21:54,760 --> 00:21:58,879 Speaker 9: that their products were harmful, especially to girls and to women. 405 00:21:59,160 --> 00:22:01,199 Speaker 9: And of course this includes all of the kind of 406 00:22:01,240 --> 00:22:06,600 Speaker 9: deep fake technologies that these companies are invested toute. 407 00:22:06,119 --> 00:22:07,960 Speaker 2: That right, and we covered that case in detail on 408 00:22:08,000 --> 00:22:08,520 Speaker 2: the program. 409 00:22:08,600 --> 00:22:11,720 Speaker 9: But yeah, well, I think you know what we have 410 00:22:12,080 --> 00:22:15,560 Speaker 9: is more and more litigation against these companies because there's 411 00:22:15,720 --> 00:22:16,920 Speaker 9: evidence of harm. 412 00:22:17,440 --> 00:22:20,800 Speaker 2: Sophear Noble, Professor and director of the Center and Resilience 413 00:22:20,800 --> 00:22:23,080 Speaker 2: and Digital Justice. You see, La, thank you very much 414 00:22:23,080 --> 00:22:25,800 Speaker 2: for joining us coming up on the show. The excitement 415 00:22:25,840 --> 00:22:30,560 Speaker 2: around SpaceX's IPO is putting pressure on market operators to 416 00:22:30,600 --> 00:22:34,360 Speaker 2: make sure this goes smoothly, we get really in the weeds, 417 00:22:34,520 --> 00:22:37,199 Speaker 2: very technical about what pulling off the biggest IPO and 418 00:22:37,280 --> 00:22:40,760 Speaker 2: history means for the market. That's next, That's what markets 419 00:22:40,760 --> 00:22:43,320 Speaker 2: look like. Stay with us. It's half time and this 420 00:22:43,359 --> 00:22:44,200 Speaker 2: is Bloomberg Tech. 421 00:22:57,440 --> 00:22:59,120 Speaker 5: Welcome back to Bloomberg Tech. 422 00:22:59,160 --> 00:23:01,080 Speaker 3: We check in on these market which are under pressure 423 00:23:01,119 --> 00:23:04,720 Speaker 3: as we await the biggest IPO in history. Then as 424 00:23:04,720 --> 00:23:07,120 Speaker 3: that one hundred is off five a percentage point, there's 425 00:23:07,160 --> 00:23:11,000 Speaker 3: geopolitical tensions, risks and terms about yet further conflict in 426 00:23:11,040 --> 00:23:12,840 Speaker 3: the Middle East between Iran the United States. 427 00:23:12,880 --> 00:23:15,119 Speaker 5: We see the semic conduct To Index hardware. 428 00:23:14,720 --> 00:23:17,359 Speaker 3: Once again, having risen so much, gets pulled back somewhat 429 00:23:17,400 --> 00:23:18,119 Speaker 3: of by two percent. 430 00:23:18,160 --> 00:23:21,840 Speaker 5: Magnificent seven also down, but some aren't. For men. ASML 431 00:23:22,160 --> 00:23:23,560 Speaker 5: just finishing trading. 432 00:23:23,200 --> 00:23:27,400 Speaker 3: In Europe record high I since nineteen ninety five. We're 433 00:23:27,480 --> 00:23:30,840 Speaker 3: up another percentage point on ASML and its European trading 434 00:23:30,880 --> 00:23:31,240 Speaker 3: on the day. 435 00:23:31,280 --> 00:23:33,000 Speaker 5: But we really do shine light and what's. 436 00:23:32,840 --> 00:23:35,320 Speaker 3: Been happening more broadly in the American indices and there is. 437 00:23:35,280 --> 00:23:35,840 Speaker 5: Some concern there. 438 00:23:35,920 --> 00:23:39,120 Speaker 2: Yeah, tech trians are driving us lower, but SpaceX there 439 00:23:39,160 --> 00:23:42,120 Speaker 2: is going to be an element of volatility whatever happens 440 00:23:42,119 --> 00:23:46,000 Speaker 2: and outside demand for SpaceX shares has market operators stress 441 00:23:46,080 --> 00:23:49,679 Speaker 2: testing their systems to ensure smooth trading for the largest 442 00:23:49,720 --> 00:23:52,560 Speaker 2: IPO in history. Bloomberg yzabel Lee has been speaking with 443 00:23:52,600 --> 00:23:54,760 Speaker 2: some of those firms. And the way that you put 444 00:23:54,760 --> 00:23:58,840 Speaker 2: it third paragraph of a critically important story is when 445 00:23:58,880 --> 00:24:02,040 Speaker 2: this IPO hits, you're talking millions and millions of orders, 446 00:24:02,359 --> 00:24:05,240 Speaker 2: and with those orders comes millions and millions of messages 447 00:24:05,560 --> 00:24:10,720 Speaker 2: and transactions. This becomes a technology story. How does that work? 448 00:24:10,960 --> 00:24:13,280 Speaker 2: What is it that they're stress testing right now? 449 00:24:13,920 --> 00:24:15,840 Speaker 11: Thank you for reading, and that's proof that you read 450 00:24:15,840 --> 00:24:17,920 Speaker 11: this story. But indeed, I think much has been said 451 00:24:17,920 --> 00:24:21,160 Speaker 11: about the excitement surrounding SpaceX IPO, but what is often 452 00:24:21,240 --> 00:24:24,520 Speaker 11: left un said is the plumbing that powers this IPO, 453 00:24:24,560 --> 00:24:26,840 Speaker 11: because for the IPO to be successful, the plumbing has 454 00:24:26,880 --> 00:24:27,560 Speaker 11: to work smoothly. 455 00:24:27,600 --> 00:24:29,399 Speaker 10: And we talked to a couple of those players. 456 00:24:29,520 --> 00:24:33,160 Speaker 11: The TCC, for one, they're the Depository Trust and Clearing Corporation. 457 00:24:33,480 --> 00:24:35,439 Speaker 11: They like to say they're the most important company that 458 00:24:35,480 --> 00:24:36,080 Speaker 11: no one has. 459 00:24:35,920 --> 00:24:36,639 Speaker 10: Ever heard of. 460 00:24:36,920 --> 00:24:39,440 Speaker 11: Think of them as the central plumbing that is basically 461 00:24:39,440 --> 00:24:42,919 Speaker 11: in charge of virtually all transactions that processes, clears and 462 00:24:42,960 --> 00:24:46,400 Speaker 11: settles us financial assets in the US. We also talk 463 00:24:46,480 --> 00:24:48,960 Speaker 11: to the S andp's equity book builder. They think of 464 00:24:49,000 --> 00:24:52,440 Speaker 11: them as like the financial technology used by global investment 465 00:24:52,480 --> 00:24:55,560 Speaker 11: banks to really power a lot of these underwritings like IPO. 466 00:24:55,760 --> 00:24:59,119 Speaker 11: So they have been preparing for weeks for this SpaceX 467 00:24:59,119 --> 00:25:00,800 Speaker 11: a IPO for DTCC. 468 00:25:00,880 --> 00:25:03,000 Speaker 10: They're going to have a watch party over the weekend. 469 00:25:03,240 --> 00:25:05,320 Speaker 11: For the SMP, they're using AI to make sure that 470 00:25:05,359 --> 00:25:06,479 Speaker 11: all systems are smooth. 471 00:25:07,320 --> 00:25:11,119 Speaker 3: One are the biggest fas is it just slowness or 472 00:25:11,160 --> 00:25:14,080 Speaker 3: is it anything that could go more deeply awry here 473 00:25:14,119 --> 00:25:14,480 Speaker 3: as well? 474 00:25:15,160 --> 00:25:17,679 Speaker 11: For DTCC, they said, it's not one big risk, but 475 00:25:17,760 --> 00:25:20,119 Speaker 11: it's just how interconnected everything is, so it could be 476 00:25:20,160 --> 00:25:23,280 Speaker 11: one small broker dealer or maybe one small market maker 477 00:25:23,280 --> 00:25:25,040 Speaker 11: that may not be as prepared and it will be 478 00:25:25,119 --> 00:25:28,160 Speaker 11: just a huge domino effect that will really affect everything. 479 00:25:28,200 --> 00:25:31,800 Speaker 11: Because they really are kind of worried about what happened 480 00:25:31,840 --> 00:25:33,840 Speaker 11: in Facebook in twenty twelve, and a lot of retail 481 00:25:33,840 --> 00:25:36,160 Speaker 11: investors were left in the dark as well as investment 482 00:25:36,200 --> 00:25:38,679 Speaker 11: bankers because it was marred by a lot of technical 483 00:25:38,720 --> 00:25:42,280 Speaker 11: failures that left some traders really uncertain. So I think 484 00:25:42,320 --> 00:25:44,080 Speaker 11: they've learned from that. It's been more than a decade 485 00:25:44,080 --> 00:25:46,800 Speaker 11: since technology has grown leaps and bounds, or they're really 486 00:25:47,000 --> 00:25:49,520 Speaker 11: preparing for it. At SMP, they have what they call 487 00:25:49,600 --> 00:25:52,040 Speaker 11: a pre mortem, which is the opposite of post mortem, 488 00:25:52,280 --> 00:25:54,720 Speaker 11: so they're really ensuring that everything is really going to 489 00:25:54,720 --> 00:25:57,119 Speaker 11: go smoothly. For example, they made sure that the tripling 490 00:25:57,160 --> 00:25:59,920 Speaker 11: of order handling capacity is going to be possible and 491 00:26:00,160 --> 00:26:04,960 Speaker 11: fourfold improvement in response time, so it really allows to digest. 492 00:26:05,040 --> 00:26:06,800 Speaker 11: But I want to join our watch party over the 493 00:26:06,800 --> 00:26:08,640 Speaker 11: weekend they're going to be online twenty four to seven. 494 00:26:08,680 --> 00:26:09,760 Speaker 10: They said, we'll. 495 00:26:09,600 --> 00:26:12,280 Speaker 3: Start our own as VALI, who has just been so 496 00:26:12,400 --> 00:26:16,320 Speaker 3: watchful on this IPO for us, thank you very much. Indeed, Meanwhile, 497 00:26:16,480 --> 00:26:19,520 Speaker 3: the AI economy, well it's moving fast and now we 498 00:26:19,600 --> 00:26:22,119 Speaker 3: may have a clearer way to track how it's changing 499 00:26:22,160 --> 00:26:25,080 Speaker 3: the way people work. ADP has partnered with the Stanford 500 00:26:25,119 --> 00:26:28,399 Speaker 3: Digital Economy Lab to launch the Canaries Dashboard. 501 00:26:28,600 --> 00:26:30,280 Speaker 5: It's a real time indicator. 502 00:26:29,840 --> 00:26:33,840 Speaker 3: Designed to show how AI is reshaping different occupations based 503 00:26:33,880 --> 00:26:37,560 Speaker 3: on actual labor market data. Joining us some More is 504 00:26:37,600 --> 00:26:40,200 Speaker 3: one of the researchers involved in this project. Nila Richardson, 505 00:26:40,280 --> 00:26:43,560 Speaker 3: chief economist at ADP, former senior economist actually at Bloomberg 506 00:26:43,600 --> 00:26:48,199 Speaker 3: and NILA present. You are monitoring the present, not the past. 507 00:26:48,680 --> 00:26:50,080 Speaker 3: But what data do you take in? 508 00:26:50,240 --> 00:26:51,240 Speaker 5: How are you monitoring this? 509 00:26:51,600 --> 00:26:51,879 Speaker 10: Well? 510 00:26:51,960 --> 00:26:55,400 Speaker 12: First of all, we are thrilled because AI is said 511 00:26:55,440 --> 00:26:58,840 Speaker 12: to be one of the most consequential technologies the world 512 00:26:58,920 --> 00:27:02,760 Speaker 12: has ever seen, and yet we have very limited ways 513 00:27:02,800 --> 00:27:05,679 Speaker 12: of measuring impact. And so this is why I'm so 514 00:27:05,760 --> 00:27:10,040 Speaker 12: excited to partner with Stanford Digital CONNOMU Lab led by 515 00:27:10,160 --> 00:27:13,399 Speaker 12: Eric Burne Jolson on these AI indicators and namely the 516 00:27:13,440 --> 00:27:18,600 Speaker 12: Canaries Dashboard, which tracks the impact of AI on occupations 517 00:27:18,640 --> 00:27:22,760 Speaker 12: in almost real time. This is about moving the conversation 518 00:27:22,880 --> 00:27:26,440 Speaker 12: about AI's impact from what we think to what we 519 00:27:26,600 --> 00:27:30,200 Speaker 12: know and what we can measure with the data Nila. 520 00:27:30,000 --> 00:27:33,119 Speaker 2: There was a section of the labor market that really 521 00:27:33,200 --> 00:27:34,919 Speaker 2: just jumps off the screen at me, and that is 522 00:27:35,000 --> 00:27:38,239 Speaker 2: the early career workers. These are people aged twenty two 523 00:27:38,320 --> 00:27:40,920 Speaker 2: to twenty five. And there is a I don't know 524 00:27:40,920 --> 00:27:43,960 Speaker 2: how you would put it, a bifurcation, right, industries that 525 00:27:44,000 --> 00:27:46,920 Speaker 2: have exposure to AI, industries that have nothing to do 526 00:27:47,040 --> 00:27:49,400 Speaker 2: with it whatsoever. What is the data telling us there? 527 00:27:49,520 --> 00:27:52,560 Speaker 2: And if you can the why the why? 528 00:27:52,880 --> 00:27:55,679 Speaker 12: Well, the important part of this data series is that 529 00:27:55,760 --> 00:28:00,159 Speaker 12: it is able to categorize over seven hundred occupations by 530 00:28:00,200 --> 00:28:02,960 Speaker 12: AI exposure in a very granular way. 531 00:28:03,240 --> 00:28:04,399 Speaker 10: But it's more than. 532 00:28:04,280 --> 00:28:09,000 Speaker 12: That because it's not just how the AI is affecting occupations, 533 00:28:09,240 --> 00:28:14,040 Speaker 12: but it's about how AIS is affecting workers people young career. 534 00:28:14,119 --> 00:28:16,679 Speaker 12: So we're able to use the demographics and the ADP 535 00:28:16,800 --> 00:28:19,800 Speaker 12: payroll data and segment it by early career twenty two 536 00:28:19,880 --> 00:28:23,160 Speaker 12: to twenty six, and look, AI, you know this as 537 00:28:23,200 --> 00:28:26,879 Speaker 12: well as anyone has two different roles in a business context, 538 00:28:27,160 --> 00:28:30,359 Speaker 12: it can augment. That's the old school story. That's the 539 00:28:30,440 --> 00:28:35,320 Speaker 12: dinosaur story of technology. How do you sorry automate work? 540 00:28:35,640 --> 00:28:38,160 Speaker 12: That's the old school story. The new school story. The 541 00:28:38,200 --> 00:28:41,320 Speaker 12: frontier is how to augment. And so what AI does 542 00:28:41,400 --> 00:28:44,280 Speaker 12: for early career in AI exposed fields, it looks like 543 00:28:44,320 --> 00:28:46,160 Speaker 12: it's automating certain tasks. 544 00:28:46,520 --> 00:28:48,560 Speaker 10: So the trick is how do we move. 545 00:28:48,440 --> 00:28:54,280 Speaker 12: From automation to augmentation and look for those higher value 546 00:28:54,440 --> 00:28:56,360 Speaker 12: tasks higher value work? 547 00:28:56,520 --> 00:28:59,760 Speaker 2: And the result is that in that space employment is contracting. 548 00:29:00,200 --> 00:29:03,760 Speaker 12: Right, So for AI exposed careers, there is a contraction 549 00:29:03,880 --> 00:29:07,840 Speaker 12: for early career in like software developers. So let's take 550 00:29:07,880 --> 00:29:11,440 Speaker 12: software developers. Since the rollout of chat GPT. In November 551 00:29:11,440 --> 00:29:14,800 Speaker 12: twenty twenty two, the dashboard shows there's been a twenty 552 00:29:14,920 --> 00:29:20,040 Speaker 12: percent decline in early career software developers, but when you 553 00:29:20,080 --> 00:29:23,480 Speaker 12: look at older workers, no decline at all. In fact, 554 00:29:23,520 --> 00:29:26,120 Speaker 12: you're seeing growth. That shows you that there is a 555 00:29:26,320 --> 00:29:30,520 Speaker 12: disparate impact here. For skills and tasks that are easily automated, 556 00:29:30,560 --> 00:29:33,360 Speaker 12: you're seeing in an effect that's the early career, But 557 00:29:33,400 --> 00:29:36,720 Speaker 12: for work where it's more complex, AI becomes a helpmate, 558 00:29:36,880 --> 00:29:41,880 Speaker 12: a coworker, an augmentation tool as opposed to an automation tool. 559 00:29:42,200 --> 00:29:44,840 Speaker 3: This is going to be released every Wednesday after Jobs Week, 560 00:29:44,880 --> 00:29:47,920 Speaker 3: so it's monthly data. How are you thinking about when 561 00:29:47,960 --> 00:29:51,240 Speaker 3: you realize an industry is becoming AI exposed At the moment, 562 00:29:51,280 --> 00:29:53,920 Speaker 3: you've been so fascinating with the fact that we've got developers, 563 00:29:54,000 --> 00:29:58,000 Speaker 3: customer services, but which one to understand luhees the next 564 00:29:58,280 --> 00:30:01,480 Speaker 3: How are you seeing low AIX posed operations starting to 565 00:30:01,600 --> 00:30:03,480 Speaker 3: change will become AI exposed. 566 00:30:03,600 --> 00:30:06,160 Speaker 12: That's a great question, and that's really the purpose and 567 00:30:06,240 --> 00:30:09,320 Speaker 12: mission of this work. It is to track value creation 568 00:30:09,560 --> 00:30:11,760 Speaker 12: in real time. And the thing about it is you 569 00:30:11,800 --> 00:30:14,320 Speaker 12: can't track it in a macro way. You can't track 570 00:30:14,360 --> 00:30:18,880 Speaker 12: it in the markets those yeah, market IPO creation. Value 571 00:30:18,880 --> 00:30:21,320 Speaker 12: creation is very different to how it affects the real 572 00:30:21,400 --> 00:30:24,959 Speaker 12: economy and what people are really experiencing at work. And 573 00:30:25,040 --> 00:30:28,080 Speaker 12: so at the task level is where you see value creation, 574 00:30:28,520 --> 00:30:32,680 Speaker 12: and you see that in certain complex jobs. So let's 575 00:30:32,720 --> 00:30:37,320 Speaker 12: move on from software developers. Maybe look at radiologists where 576 00:30:37,720 --> 00:30:41,680 Speaker 12: AI becomes a really important diagnostic tool and you can 577 00:30:41,760 --> 00:30:45,800 Speaker 12: see that value creation and delivery helping them concentrate on 578 00:30:45,840 --> 00:30:48,920 Speaker 12: the work that is necessary for human to human interaction 579 00:30:49,280 --> 00:30:53,680 Speaker 12: as opposed to simply diagnosing different patterns, which AI is 580 00:30:53,720 --> 00:30:56,720 Speaker 12: good at. So the key for employers is how to 581 00:30:57,000 --> 00:31:01,280 Speaker 12: extend human capability, not limit it, not replace it, but 582 00:31:01,400 --> 00:31:05,000 Speaker 12: extend that capability to new task and new value creation. 583 00:31:05,560 --> 00:31:08,200 Speaker 2: We have a good case study for that. Bloomboats remain Bostic. 584 00:31:08,360 --> 00:31:11,280 Speaker 2: Just spoke to the IBM CEO about this exact point. 585 00:31:12,080 --> 00:31:13,440 Speaker 2: Let's listen to what he had to say, and then 586 00:31:13,480 --> 00:31:15,120 Speaker 2: you can say if it shows up in the data. 587 00:31:15,240 --> 00:31:19,120 Speaker 13: Okay, using AI tools, now we can probably add ten 588 00:31:19,200 --> 00:31:23,440 Speaker 13: points of profit energy on day one because the amount 589 00:31:23,440 --> 00:31:25,560 Speaker 13: of time it used to take to move contracts over, 590 00:31:25,760 --> 00:31:28,400 Speaker 13: to do all of the sales automation, to do all 591 00:31:28,440 --> 00:31:32,920 Speaker 13: of the revenue forecasting. All of that now using AI 592 00:31:33,160 --> 00:31:35,840 Speaker 13: can be shrunk down literally a few weeks. 593 00:31:36,680 --> 00:31:39,680 Speaker 2: When I listen to that, I just can't draw a conclusion. 594 00:31:39,680 --> 00:31:43,440 Speaker 2: Are we talking about role elimination? Are we talking about 595 00:31:43,840 --> 00:31:46,120 Speaker 2: a boost of productivity? Which Caroline and I were told 596 00:31:46,120 --> 00:31:49,640 Speaker 2: by SFF president Mary Daily last week, isn't showing up 597 00:31:49,640 --> 00:31:52,120 Speaker 2: in the data yet, Like, how do you read the 598 00:31:52,200 --> 00:31:54,720 Speaker 2: corporate speak in this job market? 599 00:31:55,800 --> 00:31:58,800 Speaker 12: The way I read it is this AI has the 600 00:31:58,880 --> 00:32:03,040 Speaker 12: ability to reach shape work, and yet it's still a 601 00:32:03,080 --> 00:32:08,040 Speaker 12: tool that decision lies with the employer, and so this 602 00:32:08,320 --> 00:32:12,040 Speaker 12: data is about empowering employers to make that decision. Is 603 00:32:12,080 --> 00:32:16,080 Speaker 12: AI going to be your efficiency tool? There's a clear 604 00:32:16,160 --> 00:32:19,080 Speaker 12: case for that. There's a clear use case that AI 605 00:32:19,200 --> 00:32:22,240 Speaker 12: is making certain work more efficient. You can do more 606 00:32:22,320 --> 00:32:25,360 Speaker 12: with less. But I think AI has the potential for 607 00:32:25,480 --> 00:32:28,400 Speaker 12: more than that. It is a productivity tool in the 608 00:32:28,440 --> 00:32:32,160 Speaker 12: sense that it enhances work, it makes work better, it 609 00:32:32,200 --> 00:32:36,040 Speaker 12: makes problems easier to solve, and therefore you can tackle 610 00:32:36,080 --> 00:32:40,840 Speaker 12: more problems. If it's an augmentation tool, that is something 611 00:32:40,920 --> 00:32:45,080 Speaker 12: completely different, and hopefully data will help employers find that 612 00:32:45,200 --> 00:32:48,479 Speaker 12: value creation within their own businesses. So right now, I 613 00:32:48,480 --> 00:32:52,480 Speaker 12: think the narrative is really really wide. The data can 614 00:32:52,560 --> 00:32:54,960 Speaker 12: help anchor it on the truth of the moment. 615 00:32:54,840 --> 00:32:57,960 Speaker 3: And Avin would actually say they are hiring more graduates 616 00:32:58,240 --> 00:33:00,360 Speaker 3: than they happened in the past, more than ever the moment. 617 00:33:00,400 --> 00:33:03,000 Speaker 5: So it is interesting where the new grudge is coming 618 00:33:03,040 --> 00:33:03,840 Speaker 5: out and whether or not. 619 00:33:04,000 --> 00:33:06,320 Speaker 2: Now we have real time data to keep the monest 620 00:33:06,360 --> 00:33:08,880 Speaker 2: on it. Anita Richardson chief promise that ADP is great 621 00:33:08,880 --> 00:33:10,720 Speaker 2: to have you back on the show. Thank you very much. 622 00:33:10,760 --> 00:33:13,880 Speaker 2: Now coming up, Anthropic CEO Dario M. O Day says 623 00:33:13,920 --> 00:33:17,360 Speaker 2: his company's AI doesn't spell the death of software, that 624 00:33:17,760 --> 00:33:21,040 Speaker 2: some firms will be losers at that part of the conversation. Next, 625 00:33:21,080 --> 00:33:22,160 Speaker 2: this is an a big tech. 626 00:33:29,240 --> 00:33:32,800 Speaker 14: Now early on others focused on fund splashy consumer apps. 627 00:33:33,160 --> 00:33:36,640 Speaker 14: You made a bet on coding and enterprise. Why did 628 00:33:36,640 --> 00:33:40,400 Speaker 14: you make that bet? Was it a values decision or 629 00:33:40,440 --> 00:33:41,360 Speaker 14: a business decision? 630 00:33:41,800 --> 00:33:44,800 Speaker 15: Look, if you pick a business model that fundamentally conflicts 631 00:33:44,840 --> 00:33:47,840 Speaker 15: with your values, you're going to have a hard time, right, 632 00:33:48,080 --> 00:33:51,880 Speaker 15: either you betray your own values or you become irrelevant. 633 00:33:52,120 --> 00:33:54,080 Speaker 15: And so when we thought about it, we said, look, 634 00:33:54,320 --> 00:33:56,720 Speaker 15: you know, we've seen the world of social media, the 635 00:33:56,840 --> 00:34:01,360 Speaker 15: consumer world. It really seems to you know, encourage engagement 636 00:34:01,840 --> 00:34:05,240 Speaker 15: even addiction. You know, the slop we've seen with AI 637 00:34:05,320 --> 00:34:07,479 Speaker 15: video models, It's like, what's going on? Is it want 638 00:34:07,520 --> 00:34:10,120 Speaker 15: to maximize the number of minutes that you're you're paying 639 00:34:10,160 --> 00:34:14,200 Speaker 15: attention to because that's the advertising revenue driven incentive. Whereas 640 00:34:14,200 --> 00:34:17,279 Speaker 15: if we look at enterprise, look, I mean, you know, 641 00:34:17,640 --> 00:34:20,319 Speaker 15: we want to make these models useful to people. We 642 00:34:20,400 --> 00:34:24,480 Speaker 15: want to use AI to you know, cure diseases that 643 00:34:24,480 --> 00:34:27,040 Speaker 15: we couldn't cure before. Right, Well, that's working with biotech, 644 00:34:27,080 --> 00:34:30,640 Speaker 15: it's working with pharma, it's working with academic research groups. 645 00:34:30,680 --> 00:34:32,200 Speaker 15: All of those are enterprises. 646 00:34:32,320 --> 00:34:32,480 Speaker 6: Right. 647 00:34:32,760 --> 00:34:35,279 Speaker 15: We want to use AI to like, you know, to 648 00:34:35,400 --> 00:34:37,280 Speaker 15: make energy cheaper and more efficient. 649 00:34:37,440 --> 00:34:38,760 Speaker 2: That's that's all enterprise. 650 00:34:39,080 --> 00:34:40,880 Speaker 15: And so I think it served us well to have 651 00:34:41,000 --> 00:34:44,280 Speaker 15: this business model that largely aligns with our values. 652 00:34:44,719 --> 00:34:47,560 Speaker 14: Soon after Claude Cowork was released, two hundred and eighty 653 00:34:47,560 --> 00:34:51,680 Speaker 14: five billion dollars in market value vanished overnight. Traders called 654 00:34:51,719 --> 00:34:52,839 Speaker 14: it the SaaS apocalypse. 655 00:34:53,040 --> 00:34:55,760 Speaker 15: This kind of white collar wipeout story in the software 656 00:34:55,800 --> 00:34:57,360 Speaker 15: set to terrifying. 657 00:34:57,480 --> 00:34:59,359 Speaker 3: Some of those are down for nine days in a row, 658 00:34:59,400 --> 00:35:01,600 Speaker 3: So clearly the is building if. 659 00:35:01,480 --> 00:35:05,360 Speaker 14: AI continues improving at this pace, how much of traditional 660 00:35:05,400 --> 00:35:09,359 Speaker 14: software gets replaced and how fast I think. 661 00:35:09,480 --> 00:35:12,400 Speaker 15: With AI, like the pie is getting bigger. Right, so 662 00:35:12,920 --> 00:35:16,839 Speaker 15: the existing incumbents may be smaller and relative terms, some 663 00:35:16,920 --> 00:35:19,040 Speaker 15: of them may may go down in value. Some of 664 00:35:19,040 --> 00:35:21,080 Speaker 15: them may even may even go out of business if 665 00:35:21,120 --> 00:35:23,239 Speaker 15: they don't, if they don't adapt in the right way. 666 00:35:23,400 --> 00:35:26,640 Speaker 15: But like I would guess that the software industry gets larger, 667 00:35:26,760 --> 00:35:29,600 Speaker 15: not smaller, although there will be some big losers, those 668 00:35:29,640 --> 00:35:32,360 Speaker 15: who don't kind of see what's coming, who don't identify 669 00:35:32,400 --> 00:35:33,719 Speaker 15: the motes they have, they're going to have a really 670 00:35:33,760 --> 00:35:34,240 Speaker 15: hard time. 671 00:35:35,600 --> 00:35:39,320 Speaker 3: That was Bloombg Examine Chang speaking with Anthropic CEO Daria Amiday, 672 00:35:39,400 --> 00:35:41,440 Speaker 3: and you can catch part one of this two but 673 00:35:41,680 --> 00:35:44,360 Speaker 3: episode of the circuit it comes out later today. It 674 00:35:44,400 --> 00:35:48,280 Speaker 3: airs on Bloombg TV at six pm Eastern and sticking 675 00:35:48,280 --> 00:35:50,360 Speaker 3: with Anthropic, but the company has released a new model 676 00:35:50,400 --> 00:35:53,600 Speaker 3: called Claude Fable five for the capabilities of it SMITH 677 00:35:53,680 --> 00:35:57,320 Speaker 3: or SAI, but includes godrails so prevent it from responding 678 00:35:57,320 --> 00:36:00,359 Speaker 3: to queries on topics including cybersecurity and biology. 679 00:36:00,400 --> 00:36:01,960 Speaker 5: Now you'll remember Anthropic. 680 00:36:01,600 --> 00:36:05,480 Speaker 3: Initially released with us only to select organizations after warning 681 00:36:05,520 --> 00:36:07,920 Speaker 3: that it could exploit cyber vulnerabilities. 682 00:36:08,680 --> 00:36:12,760 Speaker 2: AI startup Poetic has emerged from sealth with fifty million 683 00:36:12,800 --> 00:36:16,160 Speaker 2: dollars in funding and a half a billion dollar valuation. 684 00:36:16,400 --> 00:36:20,480 Speaker 2: Right out of the gate. Poetic system helps businesses streamline complex, 685 00:36:20,800 --> 00:36:24,400 Speaker 2: long running tasks. Its founder and CEO, Markey Wagner, previously 686 00:36:24,440 --> 00:36:27,880 Speaker 2: launched AI consultancy Delphi Labs and worked on machine learning 687 00:36:28,080 --> 00:36:32,600 Speaker 2: at Google and Waimo and Markey joins us. Now this 688 00:36:32,880 --> 00:36:35,480 Speaker 2: was one of the I guess we call it a 689 00:36:35,520 --> 00:36:38,799 Speaker 2: coconut round or a mango seed round. But right out 690 00:36:38,840 --> 00:36:42,560 Speaker 2: the gate open AI Kleine Perkins founder's fund are backing you. 691 00:36:43,760 --> 00:36:46,240 Speaker 2: What is it they know that we don't yet about Poetic? 692 00:36:46,280 --> 00:36:47,279 Speaker 2: What is Poetic up to? 693 00:36:48,920 --> 00:36:51,040 Speaker 16: Yeah, so a bit about Poetic. 694 00:36:51,239 --> 00:36:54,680 Speaker 17: So Poetic is in an AI system that can learn 695 00:36:54,719 --> 00:36:58,920 Speaker 17: and execute extremely complex, multi hour processes at some of 696 00:36:58,960 --> 00:37:01,520 Speaker 17: the biggest companies in the place with over ninety nine 697 00:37:01,560 --> 00:37:04,400 Speaker 17: percent accuracy and ten times less tokens. And so what 698 00:37:04,440 --> 00:37:06,480 Speaker 17: they've seen is they know our customers and they've seen 699 00:37:06,520 --> 00:37:09,920 Speaker 17: the results, and they've seen a lot of AI pilots 700 00:37:09,920 --> 00:37:12,400 Speaker 17: that have gone well and poorly. And you know, we've 701 00:37:12,800 --> 00:37:15,520 Speaker 17: scaled up every single customer you've had into production in 702 00:37:15,600 --> 00:37:16,560 Speaker 17: a time when a. 703 00:37:16,480 --> 00:37:19,000 Speaker 16: Lot of these things are getting stuck in demo land. 704 00:37:19,120 --> 00:37:20,319 Speaker 16: And so that's what they've seen. 705 00:37:20,640 --> 00:37:23,960 Speaker 3: I mean, marketing people describe you to me as the 706 00:37:24,000 --> 00:37:27,799 Speaker 3: AI whisperer to some of the most important companies out there. 707 00:37:28,040 --> 00:37:30,399 Speaker 3: And Clin Perkins has put out a blog about them 708 00:37:30,440 --> 00:37:32,520 Speaker 3: backing you, and they reference Anthony Noto. 709 00:37:32,600 --> 00:37:33,120 Speaker 5: It's so far. 710 00:37:33,200 --> 00:37:35,840 Speaker 3: The CEO just saying how in weeks your company is 711 00:37:35,840 --> 00:37:39,400 Speaker 3: sort of turned around food processes and to end, how 712 00:37:39,440 --> 00:37:42,239 Speaker 3: and how are you doing it with not really that 713 00:37:42,320 --> 00:37:45,760 Speaker 3: much compute, that much token being used at that time. 714 00:37:46,800 --> 00:37:48,840 Speaker 17: Yeah, so we have a bit of a different approach 715 00:37:48,920 --> 00:37:51,839 Speaker 17: than what most folks are doing right now. And so 716 00:37:52,400 --> 00:37:54,400 Speaker 17: we have this system that is kind of the synthesis 717 00:37:54,480 --> 00:37:57,960 Speaker 17: of both AI and code. So, you know code today 718 00:37:58,000 --> 00:38:00,400 Speaker 17: is you know, it's very static, and so the innovations 719 00:38:00,400 --> 00:38:03,320 Speaker 17: of the past written in code. If something small changes 720 00:38:03,360 --> 00:38:05,600 Speaker 17: like a column name, it would break an AI. On 721 00:38:05,640 --> 00:38:09,000 Speaker 17: the other hand, agents are incredible, but they figure out 722 00:38:09,000 --> 00:38:10,640 Speaker 17: what to do step by step and they can easily 723 00:38:10,680 --> 00:38:11,440 Speaker 17: go off the rails. 724 00:38:11,840 --> 00:38:12,319 Speaker 2: We have this. 725 00:38:12,320 --> 00:38:15,080 Speaker 16: System that takes the best of both. 726 00:38:15,520 --> 00:38:18,080 Speaker 17: So a task is written in English similarly to an 727 00:38:18,120 --> 00:38:19,839 Speaker 17: operating procedure right, and our. 728 00:38:19,760 --> 00:38:21,399 Speaker 16: System will turn that into code under the hood. 729 00:38:22,120 --> 00:38:25,000 Speaker 3: I'm terribly sorry, Maki Wagne, but this President of the 730 00:38:25,080 --> 00:38:25,919 Speaker 3: United States is talking. 731 00:38:25,960 --> 00:38:26,880 Speaker 5: At this moment. We must go on. 732 00:38:26,920 --> 00:38:29,640 Speaker 3: Iran CEO founder a poetic We tend our attention to 733 00:38:29,680 --> 00:38:30,360 Speaker 3: President Trump 734 00:38:31,320 --> 00:38:32,799 Speaker 16: At speeds that you wouldn't want to go.