1 00:00:02,480 --> 00:00:08,720 Speaker 1: Bloomberg Audio Studios, podcasts, radio news. Well, we started today 2 00:00:08,760 --> 00:00:11,640 Speaker 1: with a big pop and shares of AMD, this after 3 00:00:11,680 --> 00:00:14,160 Speaker 1: it inked the landmark deal with open Ai to build 4 00:00:14,160 --> 00:00:17,400 Speaker 1: out infrastructure, giving the chip maker a chance to challenge 5 00:00:17,520 --> 00:00:20,680 Speaker 1: in video in the computing industry. Now, this comes on 6 00:00:20,720 --> 00:00:23,360 Speaker 1: the same day as open AI's annual Developers Day of 7 00:00:23,440 --> 00:00:27,520 Speaker 1: Developers Day, where we've seen a lot of stocks including Thigma, HubSpot, Cisco, 8 00:00:27,560 --> 00:00:30,960 Speaker 1: Mattel just to name a few, also moving significantly on 9 00:00:31,080 --> 00:00:33,839 Speaker 1: based on some of the comments coming out of that conference. 10 00:00:34,159 --> 00:00:36,880 Speaker 1: Joining us right now is Ed Ludlow. He is at 11 00:00:36,880 --> 00:00:39,400 Speaker 1: that conference and he's joined right now by the CEO 12 00:00:39,600 --> 00:00:41,320 Speaker 1: of open Ai, Brad Lightcap. 13 00:00:41,479 --> 00:00:45,920 Speaker 2: Ed. Yeah, and the headline is the ability to use 14 00:00:45,960 --> 00:00:49,440 Speaker 2: third party apps within chat GPT, and the profound impact 15 00:00:49,479 --> 00:00:52,720 Speaker 2: on the market was simply those name checked saw their 16 00:00:52,760 --> 00:00:55,959 Speaker 2: stocks move in a significant way. Delighted to speak once 17 00:00:56,000 --> 00:00:58,960 Speaker 2: again with Brad Lightcap that it's just simple as that, 18 00:00:59,040 --> 00:01:02,080 Speaker 2: as that really the ability to be within chat GPT 19 00:01:02,720 --> 00:01:07,000 Speaker 2: and access Spotify or Figma as two examples. When you 20 00:01:07,040 --> 00:01:10,319 Speaker 2: were discussing the idea of that and an API access 21 00:01:10,400 --> 00:01:13,960 Speaker 2: with those partners, was it a tough negotiation to get 22 00:01:13,959 --> 00:01:15,320 Speaker 2: them on board with the idea. 23 00:01:16,280 --> 00:01:18,959 Speaker 3: No, I mean that we've seen enthusiasm for this from 24 00:01:19,000 --> 00:01:21,720 Speaker 3: the beginning. You remember, way back when we were lost 25 00:01:21,720 --> 00:01:23,280 Speaker 3: something called plugins. 26 00:01:23,400 --> 00:01:24,800 Speaker 4: Back in the early days of chat ept. 27 00:01:24,959 --> 00:01:27,280 Speaker 3: This was one of our first attempts to start to 28 00:01:27,319 --> 00:01:30,119 Speaker 3: build an ecosystem around chat GPT so that chat ept 29 00:01:30,240 --> 00:01:32,479 Speaker 3: can start to engage with and interact with the applications 30 00:01:32,520 --> 00:01:34,160 Speaker 3: that are important to you in your personal life and 31 00:01:34,200 --> 00:01:36,759 Speaker 3: at work. And now we really have a. 32 00:01:36,720 --> 00:01:39,800 Speaker 4: Much richer surface through MCP. 33 00:01:39,520 --> 00:01:42,039 Speaker 3: And other protocols to be able to bring applications into 34 00:01:42,120 --> 00:01:44,840 Speaker 3: chat ept and for really to allow you to engage 35 00:01:44,880 --> 00:01:47,800 Speaker 3: with chat GPT in the kind of work around work right, 36 00:01:48,160 --> 00:01:51,160 Speaker 3: it's the contextual aspect of I'm doing X, or I 37 00:01:51,160 --> 00:01:53,200 Speaker 3: need why I'm on a road trip and I want 38 00:01:53,240 --> 00:01:55,000 Speaker 3: to know what playlist would go well with this in 39 00:01:55,040 --> 00:01:57,480 Speaker 3: the context of my broader trip planning. That allows you 40 00:01:57,520 --> 00:01:59,720 Speaker 3: now to kind of use chat GPT to solve that 41 00:01:59,800 --> 00:02:03,240 Speaker 3: high level task and then also integrate apps contextually to 42 00:02:03,280 --> 00:02:04,480 Speaker 3: solve those specific problems. 43 00:02:04,520 --> 00:02:06,559 Speaker 2: And we're in a place where chat GPT has become 44 00:02:06,640 --> 00:02:10,000 Speaker 2: more of an operating system. Whether that was your ambition 45 00:02:10,240 --> 00:02:14,240 Speaker 2: or not, is that where you want to take it 46 00:02:14,280 --> 00:02:18,760 Speaker 2: to be an OS and a developer driven platform. It's 47 00:02:18,760 --> 00:02:21,440 Speaker 2: almost like an app store, you know, on paper based 48 00:02:21,440 --> 00:02:22,880 Speaker 2: on what you announced this afternoon. 49 00:02:23,639 --> 00:02:26,480 Speaker 3: Well, we've always thought of chatgybt as like a super assistant. 50 00:02:27,040 --> 00:02:28,560 Speaker 4: We never set out to build a chatbot. 51 00:02:28,600 --> 00:02:31,400 Speaker 3: We always wanted to build something that was really true 52 00:02:31,400 --> 00:02:33,679 Speaker 3: to you and what your preferences. 53 00:02:33,160 --> 00:02:35,440 Speaker 4: Are, what your goals are. They could actually help you 54 00:02:35,480 --> 00:02:36,080 Speaker 4: achieve more. 55 00:02:36,480 --> 00:02:39,040 Speaker 3: And so I think part of that is CHATGIBD having 56 00:02:39,200 --> 00:02:42,560 Speaker 3: an appreciation and understanding of the applications in your life 57 00:02:42,560 --> 00:02:45,800 Speaker 3: that are important, and I think enabling that kind of 58 00:02:45,800 --> 00:02:49,640 Speaker 3: connectivity and interoperability makes Chattibt Richard also enables a lot 59 00:02:49,680 --> 00:02:51,560 Speaker 3: of pass through for people to be able to engage 60 00:02:51,600 --> 00:02:53,560 Speaker 3: with apps they love it as well as new apps. 61 00:02:53,600 --> 00:02:55,320 Speaker 5: The pass through bit is interesting. 62 00:02:55,639 --> 00:02:58,480 Speaker 2: Were their concerns with some of those technology companies you're 63 00:02:58,480 --> 00:02:59,200 Speaker 2: partnering with that. 64 00:02:59,200 --> 00:03:00,960 Speaker 5: It would take traff they away from there? 65 00:03:01,120 --> 00:03:04,760 Speaker 3: In domains now, I think mostly people are really focused 66 00:03:04,760 --> 00:03:06,240 Speaker 3: on building into new interfaces. 67 00:03:06,320 --> 00:03:06,480 Speaker 2: Right. 68 00:03:06,520 --> 00:03:09,080 Speaker 3: This is just like mobile in some sense, where you 69 00:03:09,120 --> 00:03:11,240 Speaker 3: have a new interface, you have a new form factor. 70 00:03:11,639 --> 00:03:14,320 Speaker 3: People are going to want to use mobile form factors 71 00:03:14,320 --> 00:03:16,080 Speaker 3: on the go and apps like. 72 00:03:16,080 --> 00:03:17,880 Speaker 4: Spotify in some ways exist. 73 00:03:17,800 --> 00:03:20,280 Speaker 3: Almost because they really nail mobile and so we think, 74 00:03:20,320 --> 00:03:23,280 Speaker 3: actually there's an opportunity for builders to create entirely new 75 00:03:23,320 --> 00:03:26,280 Speaker 3: applications that are even native to chat GPT, and of 76 00:03:26,320 --> 00:03:29,360 Speaker 3: course for services you love to be able to benefit 77 00:03:29,360 --> 00:03:29,680 Speaker 3: there too. 78 00:03:30,080 --> 00:03:33,160 Speaker 2: Is there a revenue sharing agreement with those third parties 79 00:03:33,240 --> 00:03:35,200 Speaker 2: whose apps are accessible to we chat GPT. 80 00:03:35,760 --> 00:03:37,640 Speaker 3: So we're going to figure out the economics of this 81 00:03:37,680 --> 00:03:41,000 Speaker 3: over time. You know, we're brand new here. Plugins was 82 00:03:41,080 --> 00:03:43,240 Speaker 3: the first version of this and that was even an experiment, 83 00:03:43,240 --> 00:03:45,240 Speaker 3: and so like everything at Opening Eye, we take this 84 00:03:45,320 --> 00:03:47,840 Speaker 3: very experimental mindset to making. 85 00:03:47,680 --> 00:03:48,440 Speaker 4: Sure we get it right. 86 00:03:48,520 --> 00:03:50,480 Speaker 3: But the idea is we do we do want to 87 00:03:50,480 --> 00:03:53,120 Speaker 3: build something that's useful for developers, and of course there's 88 00:03:53,160 --> 00:03:54,800 Speaker 3: going to be you know, have to be some exchange 89 00:03:54,840 --> 00:03:58,080 Speaker 3: in there of economics and value, and we'll have to 90 00:03:58,080 --> 00:03:59,040 Speaker 3: figure out to get that right. 91 00:03:59,360 --> 00:04:02,720 Speaker 2: You have hit eight hundred million active weekly users. You 92 00:04:02,760 --> 00:04:05,200 Speaker 2: announced on stage. Actually Greg Brockman told me that at 93 00:04:05,240 --> 00:04:08,400 Speaker 2: eight o'clock this morning, and maybe we missed it, but 94 00:04:08,560 --> 00:04:11,920 Speaker 2: it's a significant milestone. All the time, I'm asked by 95 00:04:11,920 --> 00:04:14,480 Speaker 2: all kinds of people, do we have any sense of 96 00:04:14,520 --> 00:04:17,640 Speaker 2: within that eight hundred million, how many are base level 97 00:04:17,760 --> 00:04:21,360 Speaker 2: free users and how many are premium level paid subscribers. 98 00:04:22,120 --> 00:04:25,080 Speaker 3: Yeah, we have a very healthy, you know, funnel of 99 00:04:25,120 --> 00:04:27,200 Speaker 3: people that choose to pay for chatchubut. 100 00:04:28,080 --> 00:04:30,200 Speaker 4: You know, it's surpassed for my expectations. 101 00:04:30,200 --> 00:04:33,599 Speaker 3: Frankly, were people have this kind of conception that consumers 102 00:04:33,920 --> 00:04:36,839 Speaker 3: tend to not pay for software and you know, similar 103 00:04:36,920 --> 00:04:39,000 Speaker 3: even to what I was saying before around how do 104 00:04:39,040 --> 00:04:42,560 Speaker 3: you co develop the product alongside the business model. Chatchbt 105 00:04:42,680 --> 00:04:45,120 Speaker 3: is a great example of that, where the subscription model 106 00:04:45,120 --> 00:04:47,320 Speaker 3: I think has been really a testament to how valuable 107 00:04:47,360 --> 00:04:50,640 Speaker 3: it is for more users than I think we expected 108 00:04:50,680 --> 00:04:52,680 Speaker 3: to be willing to pay for it. So we don't, 109 00:04:52,720 --> 00:04:54,960 Speaker 3: I think, just close the exact number, but it's a 110 00:04:55,320 --> 00:04:56,880 Speaker 3: healthy amount and more every day. 111 00:04:57,440 --> 00:05:01,919 Speaker 2: Open AI in the beginning went off the consumer for 112 00:05:02,080 --> 00:05:05,839 Speaker 2: uses aggressively. You are now very focused on the enterprise business. 113 00:05:06,360 --> 00:05:08,560 Speaker 2: What is the strategy for that and how do you 114 00:05:08,600 --> 00:05:10,320 Speaker 2: prioritize your enterprise business? 115 00:05:10,520 --> 00:05:12,360 Speaker 4: Yeah, I'm glad you asked about that. 116 00:05:12,560 --> 00:05:15,880 Speaker 3: So really today's announcements actually I think target what are 117 00:05:16,120 --> 00:05:18,760 Speaker 3: an important set of use cases for the enterprise things 118 00:05:19,000 --> 00:05:22,159 Speaker 3: we've been hearing enterprises ask us about now for some time, 119 00:05:22,320 --> 00:05:26,159 Speaker 3: so specifically, one is we now have an ability for 120 00:05:26,279 --> 00:05:29,360 Speaker 3: enterprises to build agents in a much more visual, much 121 00:05:29,360 --> 00:05:32,240 Speaker 3: more intuitive way. You've heard us say twenty twenty five 122 00:05:32,279 --> 00:05:34,360 Speaker 3: has been the year of agents. We think that's true. 123 00:05:34,720 --> 00:05:36,960 Speaker 3: Codex has been a great example for us of that. 124 00:05:37,320 --> 00:05:40,360 Speaker 3: Our coding agent now available through an API. 125 00:05:40,520 --> 00:05:42,719 Speaker 5: Also, the lead time to make software is a lot shorter. 126 00:05:42,960 --> 00:05:43,919 Speaker 4: It's gotten a lot shorter. 127 00:05:44,040 --> 00:05:46,280 Speaker 3: I think you saw today we demoed live demo, I 128 00:05:46,279 --> 00:05:48,360 Speaker 3: think three or four different things that we've built in 129 00:05:48,400 --> 00:05:48,880 Speaker 3: real time. 130 00:05:49,200 --> 00:05:51,240 Speaker 4: We expect that to continue to be to the trend. 131 00:05:51,720 --> 00:05:54,560 Speaker 3: Things like Agent Builder allow enterprises to be able to 132 00:05:54,560 --> 00:05:59,479 Speaker 3: build agentic experiences, powerful identic experiences on the go, iteratively 133 00:06:00,040 --> 00:06:03,120 Speaker 3: and connected into the tools and sources of information that 134 00:06:03,160 --> 00:06:03,839 Speaker 3: matter for the business. 135 00:06:03,920 --> 00:06:06,040 Speaker 2: The data point that jumps out me is your API 136 00:06:06,160 --> 00:06:09,880 Speaker 2: is handling more than six billion tokens per minute, and 137 00:06:09,920 --> 00:06:13,800 Speaker 2: that helps explain why the AMD deal, you know, which 138 00:06:13,800 --> 00:06:16,120 Speaker 2: is focused on inference. You are involved in all of 139 00:06:16,160 --> 00:06:19,120 Speaker 2: these domains of the company. I've already asked Greg, but 140 00:06:19,160 --> 00:06:21,359 Speaker 2: I've got to ask you, how are you going to 141 00:06:21,480 --> 00:06:24,880 Speaker 2: finance yet another infrastructure project? Like is there going to 142 00:06:24,920 --> 00:06:28,880 Speaker 2: be some debt here specific for the AMD capacity, and 143 00:06:28,920 --> 00:06:31,480 Speaker 2: how do you move quickly to get it online? 144 00:06:31,800 --> 00:06:34,480 Speaker 3: Yeah, well, the high level thing is we are tremendously 145 00:06:34,480 --> 00:06:37,359 Speaker 3: compute constrained. It feels like we're in this kind of 146 00:06:37,360 --> 00:06:40,039 Speaker 3: recurring theme of being compute constrained. And I think the 147 00:06:40,080 --> 00:06:42,120 Speaker 3: reason for that is the answer to the question you ask, 148 00:06:42,200 --> 00:06:45,000 Speaker 3: which is demand. Right, we see there are multiples of 149 00:06:45,000 --> 00:06:49,400 Speaker 3: demand that are lateent and untapped from what we have today. 150 00:06:49,400 --> 00:06:52,120 Speaker 3: And even today, obviously by any standard, demand in revenue 151 00:06:52,120 --> 00:06:56,679 Speaker 3: growth has been torrid in its pace, and so really 152 00:06:56,720 --> 00:06:57,920 Speaker 3: we have to invest ahead of that. 153 00:06:57,960 --> 00:06:59,080 Speaker 4: And I think that's going to be the. 154 00:06:59,040 --> 00:07:01,560 Speaker 3: Ray limitter for us to be able to go capture demand, 155 00:07:01,560 --> 00:07:03,880 Speaker 3: whether it's consumer or enterprise, and for us to be 156 00:07:03,920 --> 00:07:07,960 Speaker 3: able to build new models, paralyze more experiences, more product experiences, 157 00:07:08,120 --> 00:07:11,080 Speaker 3: and then enable users specifically to be able to use those. 158 00:07:10,920 --> 00:07:14,360 Speaker 4: Products more actively in their daily life at work and 159 00:07:14,400 --> 00:07:14,720 Speaker 4: at home. 160 00:07:14,800 --> 00:07:17,600 Speaker 3: And so, you know, even things like Sora, the app 161 00:07:17,640 --> 00:07:20,760 Speaker 3: we just launched, we wish we could invite more people 162 00:07:20,760 --> 00:07:22,840 Speaker 3: onto it now, but we just need more compute. So 163 00:07:22,880 --> 00:07:27,400 Speaker 3: the AMD deal we're excited about being you know, directionally 164 00:07:27,920 --> 00:07:28,640 Speaker 3: a way for us. 165 00:07:28,520 --> 00:07:28,960 Speaker 5: To do that. 166 00:07:29,800 --> 00:07:32,360 Speaker 2: I've got to ask about the report that open AI 167 00:07:32,520 --> 00:07:36,119 Speaker 2: closed secondary or the ability for employees to sell shares 168 00:07:36,120 --> 00:07:39,200 Speaker 2: at a five hundred billion dollar valuation. I already asked 169 00:07:39,200 --> 00:07:41,680 Speaker 2: you this question, but what is the metric we're supposed 170 00:07:41,680 --> 00:07:44,960 Speaker 2: to judge your success by the five hundred billion dollar valuation? 171 00:07:45,360 --> 00:07:48,440 Speaker 2: The six billion tokens per minute? To you, Frad, what 172 00:07:48,600 --> 00:07:48,800 Speaker 2: is it? 173 00:07:49,520 --> 00:07:52,720 Speaker 3: For me? It's it's actually kind of a metric that 174 00:07:52,880 --> 00:07:57,400 Speaker 3: we we talked about is tokens. It's you mentioned six 175 00:07:57,440 --> 00:08:01,880 Speaker 3: billion tokens per minute on our That is the purest 176 00:08:02,320 --> 00:08:02,600 Speaker 3: for me. 177 00:08:02,720 --> 00:08:05,880 Speaker 4: The kind of essence of utility is that consumption metric. 178 00:08:05,960 --> 00:08:09,760 Speaker 3: And so we've actively tracked that metric to see how 179 00:08:09,760 --> 00:08:12,800 Speaker 3: people's consumption of AI is growing over time. And you 180 00:08:12,880 --> 00:08:16,080 Speaker 3: see this happen in amazing ways. So things like Codex, 181 00:08:16,080 --> 00:08:19,400 Speaker 3: for example, we've seen grow ten x since August purely 182 00:08:19,440 --> 00:08:22,880 Speaker 3: on consumption of tokens around coding, and you start to 183 00:08:22,920 --> 00:08:25,880 Speaker 3: see that same pattern emerge across multiple lanes of use 184 00:08:26,040 --> 00:08:28,400 Speaker 3: and across multiple areas of work. And that's the metric 185 00:08:28,440 --> 00:08:30,280 Speaker 3: I look at because if that number is going up, 186 00:08:30,320 --> 00:08:32,080 Speaker 3: it means people are using us for more things, and 187 00:08:32,120 --> 00:08:32,920 Speaker 3: that's the ultimate Goal. 188 00:08:33,520 --> 00:08:35,920 Speaker 5: Brad Lightcap is open ai Coo. 189 00:08:36,280 --> 00:08:38,240 Speaker 2: There has been a fire hose of headlines and it 190 00:08:38,360 --> 00:08:39,839 Speaker 2: is moved markets all day long.