1 00:00:01,400 --> 00:00:06,720 Speaker 1: From Marhart where Innovation, Money and Power Collie in Silicon Valley, NBN. 2 00:00:07,040 --> 00:00:11,080 Speaker 1: This is Bloomberg Technology with Caroline Hyde and Ed Ludlow. 3 00:00:25,280 --> 00:00:28,280 Speaker 2: I'm Caroline Hyde at Bloomberg's World headquarters in New York. 4 00:00:28,440 --> 00:00:31,160 Speaker 3: And I'm Ed Ludlow in San Francisco. This is Bloomberg 5 00:00:31,200 --> 00:00:32,360 Speaker 3: Technology coming up. 6 00:00:32,440 --> 00:00:35,280 Speaker 2: We count you down to the FED rate decision and 7 00:00:35,680 --> 00:00:37,720 Speaker 2: get a real look of you on the state of 8 00:00:37,800 --> 00:00:40,479 Speaker 2: the banking system and the FinTechs have helped speed up 9 00:00:40,520 --> 00:00:42,880 Speaker 2: deposit flight adding to crisis concerns. 10 00:00:43,760 --> 00:00:45,479 Speaker 3: We're going to dive deep into the world at Venture 11 00:00:45,479 --> 00:00:49,040 Speaker 3: Capital and speak with Sequoia general partner Constantine Beula and 12 00:00:49,120 --> 00:00:51,280 Speaker 3: Coastladventures founder Vinode Coastler. 13 00:00:51,560 --> 00:00:55,880 Speaker 2: Plus how malware operators are caching in on the AI height. 14 00:00:55,880 --> 00:00:57,320 Speaker 4: To target your businesses. 15 00:00:57,680 --> 00:01:00,400 Speaker 2: We'll talk CHATCHYPT and cybersecurity with let Us. 16 00:01:00,360 --> 00:01:02,320 Speaker 4: Head of Security later this hour. 17 00:01:02,760 --> 00:01:04,920 Speaker 2: Let's turn towards, of course, what else the Federal Reserve 18 00:01:04,959 --> 00:01:06,920 Speaker 2: is really trying to navigate how you can hight rates 19 00:01:06,920 --> 00:01:10,000 Speaker 2: amid what is an ongoing banking crisis in part brought 20 00:01:10,000 --> 00:01:13,320 Speaker 2: through of course higher interest rates. Also, though, is there 21 00:01:13,360 --> 00:01:15,520 Speaker 2: a bit of an acceleration of some of these crises 22 00:01:15,800 --> 00:01:17,800 Speaker 2: because of the ease with which we can pull cash 23 00:01:17,880 --> 00:01:20,080 Speaker 2: out of our deposits. Joining us now to discuss all 24 00:01:20,120 --> 00:01:22,800 Speaker 2: of this cannis known us OURGP management consultant. 25 00:01:22,880 --> 00:01:23,319 Speaker 4: You've got what. 26 00:01:23,360 --> 00:01:26,760 Speaker 2: Sixteen years in career of risk management, you worked at 27 00:01:26,800 --> 00:01:29,160 Speaker 2: the FDIC. Canis I put it to you of how 28 00:01:29,240 --> 00:01:32,800 Speaker 2: much of a worry is at the moment fintech is 29 00:01:32,840 --> 00:01:36,880 Speaker 2: the ability with which we can withdraw our deposits from banks. 30 00:01:36,920 --> 00:01:38,760 Speaker 2: Does that sort of speed up some of the anxiety 31 00:01:38,800 --> 00:01:40,160 Speaker 2: and the crisis cycles we're seeing? 32 00:01:41,319 --> 00:01:45,479 Speaker 5: Yeah, thanks for having me. It certainly plays a role. 33 00:01:46,240 --> 00:01:49,920 Speaker 5: I think we have fintech technology to pull out our 34 00:01:50,400 --> 00:01:53,680 Speaker 5: deposits very quickly. But what really is the problem with 35 00:01:53,840 --> 00:01:59,520 Speaker 5: this particular banking crisis is the high concentration of uninsured 36 00:01:59,560 --> 00:02:03,840 Speaker 5: deposits we have banks. Silicon Valley Bank had over ninety 37 00:02:03,880 --> 00:02:08,360 Speaker 5: percent exposure to uninsured deposits. And by their very nature, 38 00:02:08,480 --> 00:02:13,320 Speaker 5: uninsured deposits are not sensitive to the FED insurance, the 39 00:02:13,360 --> 00:02:18,000 Speaker 5: FDIC insurance. Rather, they're very opportunistic, so they move fast 40 00:02:18,480 --> 00:02:21,640 Speaker 5: on to a better opportunity, or they get spooked, and 41 00:02:21,680 --> 00:02:23,920 Speaker 5: they kind of all move in tandem. And that's what 42 00:02:23,960 --> 00:02:27,080 Speaker 5: we saw in this most recent banking crisis. 43 00:02:27,120 --> 00:02:30,840 Speaker 2: Are you expecting a change to whether all deposits end 44 00:02:30,880 --> 00:02:32,440 Speaker 2: up being ensured by the FDIC. 45 00:02:34,080 --> 00:02:36,960 Speaker 5: Well, if all deposits end up being insured by the FDIC, 46 00:02:37,240 --> 00:02:41,119 Speaker 5: that means that the FDIC bank fees. The fees that 47 00:02:41,440 --> 00:02:44,960 Speaker 5: the FDIC charges for deposit insurance is going to have 48 00:02:45,120 --> 00:02:48,400 Speaker 5: to increase and is going to have to be distributed 49 00:02:48,480 --> 00:02:53,239 Speaker 5: differently among banks. I mean, I suppose that's an option, 50 00:02:53,720 --> 00:02:56,920 Speaker 5: but I think that there are so many more prudent 51 00:02:56,960 --> 00:03:01,359 Speaker 5: decisions that can be made beforehand to improve revision, including 52 00:03:01,400 --> 00:03:05,040 Speaker 5: the use of technology and supervision, such that we don't 53 00:03:05,040 --> 00:03:08,040 Speaker 5: have to ensure all deposits. It kind of really doesn't 54 00:03:08,120 --> 00:03:10,639 Speaker 5: make sense when there are so many other interim things 55 00:03:10,680 --> 00:03:11,400 Speaker 5: that you can do. 56 00:03:11,800 --> 00:03:12,880 Speaker 1: The regulators can do. 57 00:03:12,919 --> 00:03:15,079 Speaker 5: And if you've read their most recent reports by both 58 00:03:15,120 --> 00:03:18,360 Speaker 5: the FED and the FDIC, they admitted their shortfalls. They 59 00:03:18,360 --> 00:03:20,560 Speaker 5: admitted their short but I think that that can be 60 00:03:20,600 --> 00:03:24,480 Speaker 5: ameliorated with the use of technology, among other things. 61 00:03:24,480 --> 00:03:26,760 Speaker 2: Interesting ed almost technology at the heart of this, but 62 00:03:26,800 --> 00:03:28,560 Speaker 2: also what's at the heart is the fact that we 63 00:03:28,600 --> 00:03:32,640 Speaker 2: have neobanks fintech able to offer higher returns. ED to 64 00:03:32,639 --> 00:03:34,239 Speaker 2: that extent, we're seeing a bit of a stampede. It 65 00:03:34,320 --> 00:03:35,200 Speaker 2: keeps being the watchword. 66 00:03:36,080 --> 00:03:39,160 Speaker 3: Yeah, stampede or herd mentality. I mean, kend this. I 67 00:03:39,200 --> 00:03:41,400 Speaker 3: go back to what happened with SVB and more recently 68 00:03:41,400 --> 00:03:45,040 Speaker 3: with First Republic, that it's an issue of confidence, right, 69 00:03:45,120 --> 00:03:49,240 Speaker 3: Depositors lose confidence and pull their money. The point that 70 00:03:49,320 --> 00:03:51,560 Speaker 3: Caroline raised this morning is so smart that now you 71 00:03:51,640 --> 00:03:56,200 Speaker 3: have so many means technologically speaking to transfer funds. I 72 00:03:56,280 --> 00:03:58,600 Speaker 3: just want you to try and give us some granularity 73 00:03:58,680 --> 00:04:01,840 Speaker 3: on how much that has changed the game. How much 74 00:04:01,960 --> 00:04:04,560 Speaker 3: is contributed through the risk of a bank run. 75 00:04:05,960 --> 00:04:08,240 Speaker 5: Oh, I think it's contributed greatly to the risk. I 76 00:04:08,280 --> 00:04:10,360 Speaker 5: mean you can have a bank run from your cell phone, 77 00:04:10,360 --> 00:04:13,320 Speaker 5: as kind of what you're suggesting. Back in the olden days, 78 00:04:13,360 --> 00:04:15,600 Speaker 5: people stood in line and waited for the banks to 79 00:04:15,680 --> 00:04:18,400 Speaker 5: open in order to pull out their physical cash. Now 80 00:04:18,440 --> 00:04:20,680 Speaker 5: you can have a bank run from your living room. 81 00:04:21,160 --> 00:04:24,320 Speaker 5: So I don't think payments are not going to slow down. 82 00:04:24,360 --> 00:04:28,919 Speaker 5: Payment technology, fintech technology, that's not going to slow down. 83 00:04:29,320 --> 00:04:32,200 Speaker 5: I think as a result of that, risk managers need 84 00:04:32,240 --> 00:04:36,160 Speaker 5: to be more prudent with regard to their concentration and 85 00:04:36,200 --> 00:04:40,600 Speaker 5: their distribution to exposures of things like uninsured deposits, so 86 00:04:41,080 --> 00:04:43,640 Speaker 5: you're going to continue to have advances and payments. I 87 00:04:43,640 --> 00:04:46,520 Speaker 5: think they're healthy for the economy in a lot of ways. 88 00:04:46,560 --> 00:04:50,880 Speaker 5: They make our lives as consumers very convenient. But that's 89 00:04:50,960 --> 00:04:53,560 Speaker 5: not going to change, and we're still going to be 90 00:04:53,600 --> 00:04:57,080 Speaker 5: exposed to the risk of rapid withdrawals from bank and 91 00:04:57,160 --> 00:05:00,280 Speaker 5: financial institutions, given our ease with which we can do it. 92 00:05:00,520 --> 00:05:03,880 Speaker 3: Candice, what have you been advising your clients since Silicon 93 00:05:04,000 --> 00:05:07,320 Speaker 3: Valley collapse, since First Republic was acquired by JP Morgan. 94 00:05:07,400 --> 00:05:10,600 Speaker 3: What have you told them to do from a tech perspective. 95 00:05:11,680 --> 00:05:15,320 Speaker 5: Yeah, I think that you know, this is unlike our 96 00:05:15,440 --> 00:05:18,320 Speaker 5: last crisis, where it was a credit crisis. This is 97 00:05:18,400 --> 00:05:23,120 Speaker 5: an asset liability mismatch. And we always advise our clients 98 00:05:24,000 --> 00:05:27,080 Speaker 5: even despite this, and admittedly we've worked with the more 99 00:05:27,600 --> 00:05:31,320 Speaker 5: larger and mid size clients in my particular space, but 100 00:05:31,400 --> 00:05:35,200 Speaker 5: we always advise them to stay on top of stress 101 00:05:35,240 --> 00:05:38,480 Speaker 5: testing and be very prudent about managing your risk and 102 00:05:38,520 --> 00:05:43,360 Speaker 5: managing your exposures to shift in the macroeconomic market. Look 103 00:05:43,400 --> 00:05:47,359 Speaker 5: for macro hedges to whatever it is that you're taking 104 00:05:47,400 --> 00:05:51,200 Speaker 5: exposure to. We're still giving the same good risk management 105 00:05:51,440 --> 00:05:55,560 Speaker 5: advice despite what's going on here in the market, because 106 00:05:55,600 --> 00:05:59,920 Speaker 5: these things were foreseen and they continue to be able 107 00:05:59,920 --> 00:06:02,599 Speaker 5: to you can you can gauge them, and that's what 108 00:06:02,640 --> 00:06:04,479 Speaker 5: we're revising our big clients to do. 109 00:06:05,680 --> 00:06:05,800 Speaker 6: Right. 110 00:06:05,880 --> 00:06:08,359 Speaker 3: Candice, knowness of our GP, thank you so much for 111 00:06:08,400 --> 00:06:11,320 Speaker 3: your time. Now coming up what to make of the 112 00:06:11,440 --> 00:06:14,880 Speaker 3: rapid evolution in the generative AI space, will bring you 113 00:06:14,920 --> 00:06:19,120 Speaker 3: that conversation with Sequoia Capital partner an AI expert, Constantine Buler. 114 00:06:19,560 --> 00:06:20,200 Speaker 7: Caroline. 115 00:06:20,480 --> 00:06:22,520 Speaker 2: Yeah, and of course we have got to keep our 116 00:06:22,520 --> 00:06:25,400 Speaker 2: eyes on all things chips, on all things MD. After 117 00:06:25,640 --> 00:06:27,719 Speaker 2: a really tepid set of numbers coming out of what 118 00:06:27,839 --> 00:06:30,480 Speaker 2: is the second biggest PC chip maker, and we know 119 00:06:30,560 --> 00:06:32,839 Speaker 2: the woes, we know the concern about the PC market 120 00:06:32,839 --> 00:06:34,520 Speaker 2: we saw in an Intel. We move it to AMD, 121 00:06:34,680 --> 00:06:37,040 Speaker 2: but all eyes there from Qualcom after the bells, as 122 00:06:37,080 --> 00:06:40,280 Speaker 2: that particular chip maker has really diversified, perhaps away. 123 00:06:40,000 --> 00:06:41,600 Speaker 4: From just PCs. We're off by more than eight and 124 00:06:41,600 --> 00:06:42,159 Speaker 4: a half percent. 125 00:06:42,360 --> 00:07:00,800 Speaker 2: As a Bloomberg, let's talk artificial intelligence with AI. Jeffrey 126 00:07:00,839 --> 00:07:03,960 Speaker 2: hidden Hinton leaving Google and adding his voice to a 127 00:07:04,000 --> 00:07:07,440 Speaker 2: growing chorus of experts really warning about the dangers of AI. 128 00:07:07,800 --> 00:07:08,600 Speaker 4: What does this mean? 129 00:07:08,720 --> 00:07:11,559 Speaker 2: His departure for Alphabet's efforts and its Deep Mind AI 130 00:07:11,680 --> 00:07:14,840 Speaker 2: lab or Deep Mind CEO Demis Hasabis sat down with 131 00:07:14,960 --> 00:07:17,520 Speaker 2: Bluemberg Originals and talked about how the industry is not 132 00:07:17,600 --> 00:07:20,360 Speaker 2: just losing people but also attracting a lot of talents. 133 00:07:20,360 --> 00:07:24,679 Speaker 3: To take a listen, it's an incredibly dynamic field right now, 134 00:07:25,600 --> 00:07:29,800 Speaker 3: all sorts of new things happening, research going on, you know, literally, 135 00:07:29,840 --> 00:07:31,920 Speaker 3: I would say in the last six months, probably hundreds 136 00:07:31,920 --> 00:07:34,880 Speaker 3: of thousands of people, very talented engineers and others, have 137 00:07:34,920 --> 00:07:38,680 Speaker 3: got into the field. You know. I think they've come 138 00:07:38,720 --> 00:07:41,160 Speaker 3: from many other fields and they've decided maybe they weren't 139 00:07:41,560 --> 00:07:44,280 Speaker 3: interest in AI for long, you know, until very recently, 140 00:07:44,440 --> 00:07:46,360 Speaker 3: but they've decided this is a this is a great 141 00:07:46,360 --> 00:07:49,600 Speaker 3: growth point. And so that's brought in a lot of 142 00:07:49,720 --> 00:07:53,440 Speaker 3: energy into the space and a lot of dynamic ideas. 143 00:07:53,880 --> 00:07:56,120 Speaker 3: And so even for us, who are you know, right 144 00:07:56,160 --> 00:07:58,200 Speaker 3: in the middle of it, it's hard to keep track. 145 00:07:58,080 --> 00:07:59,560 Speaker 7: Of all of that at once. 146 00:07:59,600 --> 00:08:01,360 Speaker 3: So you have to try and take into account and 147 00:08:01,840 --> 00:08:04,560 Speaker 3: look at the main trends and the main threads and 148 00:08:04,680 --> 00:08:07,640 Speaker 3: sort of focus on that. I would say, but it's 149 00:08:08,120 --> 00:08:10,080 Speaker 3: almost like you need an AI system to help you 150 00:08:10,400 --> 00:08:12,920 Speaker 3: keep track of all of the developments that are going on. 151 00:08:15,080 --> 00:08:18,040 Speaker 3: Google deep Mind CEO Demis Hasavis and you can watch 152 00:08:18,200 --> 00:08:22,160 Speaker 3: more of that interview on Bloomberg's Originals AI IRL and 153 00:08:22,200 --> 00:08:25,679 Speaker 3: stream new episodes Wednesdays eight thirty pm Eastern five thirty 154 00:08:25,920 --> 00:08:26,640 Speaker 3: Pacific time. 155 00:08:26,960 --> 00:08:27,520 Speaker 7: Let's keep the. 156 00:08:27,440 --> 00:08:31,080 Speaker 3: Conversation going about AI with Sequoia Capital partner Constantine Buler, 157 00:08:31,400 --> 00:08:33,720 Speaker 3: who's been immersed in the world of AI for a 158 00:08:33,760 --> 00:08:36,440 Speaker 3: decade already. We've got some catching up to do with you. 159 00:08:37,120 --> 00:08:39,880 Speaker 3: That was an interesting move because the history of AI 160 00:08:40,280 --> 00:08:42,760 Speaker 3: is very closely tied with deep Mind. They've kind of 161 00:08:42,880 --> 00:08:45,760 Speaker 3: brought Google and deep Mind together. You heard Demis there. 162 00:08:45,840 --> 00:08:48,600 Speaker 3: Can I just get your reaction to the landscape and 163 00:08:48,640 --> 00:08:50,400 Speaker 3: that in particular. 164 00:08:50,480 --> 00:08:53,839 Speaker 8: Absolutely, First, thanks for having us to pleasure to be 165 00:08:53,880 --> 00:08:57,480 Speaker 8: here always. So we are in the middle of an 166 00:08:57,480 --> 00:09:01,880 Speaker 8: incredible revolution in AI. It is probably most similar to 167 00:09:02,000 --> 00:09:05,560 Speaker 8: the personal computer revolution of forty years ago, and Demis 168 00:09:05,679 --> 00:09:08,240 Speaker 8: alludes to this briefly in his comment when he says 169 00:09:08,320 --> 00:09:10,680 Speaker 8: we're in the middle of something very big here. Why 170 00:09:10,720 --> 00:09:14,160 Speaker 8: do I say personal computer revolution because of the impact 171 00:09:14,160 --> 00:09:17,960 Speaker 8: it's having. Imagine, right before the personal computer revolution, you 172 00:09:18,040 --> 00:09:20,679 Speaker 8: were a typist and every time you made a mistake 173 00:09:20,960 --> 00:09:22,880 Speaker 8: or you had to make a change, you'd throw away 174 00:09:22,880 --> 00:09:27,400 Speaker 8: the paper. You have to put the type exactly exactly. 175 00:09:27,480 --> 00:09:29,480 Speaker 8: You throw away the paper, you have to put a 176 00:09:29,480 --> 00:09:31,800 Speaker 8: new one in and you have to start all over again. 177 00:09:32,240 --> 00:09:37,040 Speaker 8: And then the personal computer comes around and everything is faster, 178 00:09:37,720 --> 00:09:42,920 Speaker 8: it's easier, it's way more streamlined. But there are some changes, 179 00:09:42,960 --> 00:09:45,320 Speaker 8: there's some retraining that needs to be done. You have 180 00:09:45,400 --> 00:09:48,760 Speaker 8: to learn this new technology. Same kind of thing is 181 00:09:48,760 --> 00:09:53,040 Speaker 8: happening here in AI. So there's this massive potential. We're 182 00:09:53,080 --> 00:09:55,400 Speaker 8: right in the middle of it. It's been the cornerstone 183 00:09:55,400 --> 00:09:58,760 Speaker 8: of my career for thirteen years. Sequoya, it's been a 184 00:09:58,760 --> 00:10:02,720 Speaker 8: cornerstone for us for thirty years, and it is just 185 00:10:02,840 --> 00:10:05,000 Speaker 8: the early innings of what's going to be very. 186 00:10:04,920 --> 00:10:06,120 Speaker 6: Brings us to hear in now though. 187 00:10:06,160 --> 00:10:10,120 Speaker 3: So you are principally a seed stage and series A investor. Yes, 188 00:10:10,280 --> 00:10:12,880 Speaker 3: you know there are many that echo your sentiments. How 189 00:10:12,920 --> 00:10:16,200 Speaker 3: do you actually invest? Yes in that moment. 190 00:10:16,760 --> 00:10:20,720 Speaker 6: So the first thing is founder driven, Okay. 191 00:10:21,000 --> 00:10:24,240 Speaker 8: Founders are always the lifeblood of the venture capital industry, 192 00:10:24,440 --> 00:10:27,720 Speaker 8: the technology industry at large, and definitely the AI industry. 193 00:10:27,960 --> 00:10:32,400 Speaker 8: So think back to when Sequoia started its venture in ai, 194 00:10:32,480 --> 00:10:35,480 Speaker 8: which was nineteen ninety two when we backed at the 195 00:10:35,520 --> 00:10:39,520 Speaker 8: Series A, the first investment in arguably, if not the 196 00:10:39,559 --> 00:10:43,319 Speaker 8: most important AI hardware company in the world in Nvidia. 197 00:10:43,880 --> 00:10:46,880 Speaker 8: That was a founder driven investment. That was a founder 198 00:10:46,960 --> 00:10:50,920 Speaker 8: driven investment because Jensen Wang was an exceptional engineer from 199 00:10:50,920 --> 00:10:54,680 Speaker 8: a previous Sequoia company, So founder driven as always with Sequoia, 200 00:10:54,800 --> 00:10:57,760 Speaker 8: and then in nineteen ninety nine, arguably the most important 201 00:10:57,960 --> 00:11:02,360 Speaker 8: at maturity AI company, Google was also a Sequoia investment, 202 00:11:02,480 --> 00:11:04,520 Speaker 8: and that brings us all the way to today, which 203 00:11:04,559 --> 00:11:07,320 Speaker 8: is a lot of great AI companies that are coming up, 204 00:11:07,559 --> 00:11:10,800 Speaker 8: hundreds at this point that we're meeting. We're excited to 205 00:11:10,800 --> 00:11:13,120 Speaker 8: get to know and finding the most brilliant founders for 206 00:11:13,160 --> 00:11:15,520 Speaker 8: what they're building in the future is what Sequoia is 207 00:11:15,679 --> 00:11:16,200 Speaker 8: all about. 208 00:11:16,400 --> 00:11:19,840 Speaker 3: You know, Caroline's Constantine's point. We see that activity on 209 00:11:19,880 --> 00:11:23,319 Speaker 3: the show every single day, the hundreds of new companies 210 00:11:23,320 --> 00:11:24,080 Speaker 3: that he's talking about. 211 00:11:24,120 --> 00:11:24,680 Speaker 6: That energy. 212 00:11:25,240 --> 00:11:27,720 Speaker 3: It came from open Ai in November of last year, 213 00:11:28,000 --> 00:11:30,320 Speaker 3: and then Microsoft's investment at the beginning of the year. 214 00:11:30,240 --> 00:11:32,000 Speaker 4: And of course the Quoia backs open Ai. 215 00:11:32,120 --> 00:11:34,280 Speaker 2: But I'm interested to dig in hugging Face and then 216 00:11:34,320 --> 00:11:37,400 Speaker 2: others in your portfolio. Constantine, it feels as though the 217 00:11:37,440 --> 00:11:40,360 Speaker 2: rest of the VC community is rushing. It feels that 218 00:11:40,440 --> 00:11:43,240 Speaker 2: regulators are having to be forced to rush. It feels 219 00:11:43,280 --> 00:11:45,480 Speaker 2: as though this has taken us all by complete surprise. 220 00:11:45,840 --> 00:11:48,640 Speaker 2: But you've been studying at Stanford AI. As you say, 221 00:11:48,679 --> 00:11:51,439 Speaker 2: it's been Cornerstone for thirteen years. Why have we suddenly 222 00:11:51,640 --> 00:11:53,600 Speaker 2: been sort of had the rug pulled a bit? 223 00:11:54,480 --> 00:11:57,599 Speaker 6: Okay, so you're exactly right, Caroline. 224 00:11:57,720 --> 00:12:01,360 Speaker 8: This is actually decades in the making, and for us, 225 00:12:01,400 --> 00:12:05,079 Speaker 8: it's sekoya for me, for our team. 226 00:12:04,360 --> 00:12:04,840 Speaker 6: There's. 227 00:12:06,240 --> 00:12:09,000 Speaker 8: This is something that we've been training for and excited 228 00:12:09,040 --> 00:12:12,480 Speaker 8: for for decades. But there is a big, exciting change 229 00:12:12,520 --> 00:12:15,200 Speaker 8: and it is exactly what you described, which is this 230 00:12:15,400 --> 00:12:19,320 Speaker 8: new user interface. When I talk about the user interface, 231 00:12:19,760 --> 00:12:22,920 Speaker 8: I'm harkening back to the personal computer when you had 232 00:12:22,960 --> 00:12:26,800 Speaker 8: the graphical user interface come out, and that was the 233 00:12:26,880 --> 00:12:30,960 Speaker 8: advent of the first Microsoft Windows, that was the advent 234 00:12:31,000 --> 00:12:33,680 Speaker 8: of a lot of the first Apple computers. And the 235 00:12:33,800 --> 00:12:37,920 Speaker 8: point of the first graphical user interface in the personal 236 00:12:37,960 --> 00:12:42,840 Speaker 8: computer revolution is that anyone could just use their mouse 237 00:12:42,880 --> 00:12:46,560 Speaker 8: and click, as opposed to the old terminal based MS 238 00:12:46,600 --> 00:12:51,320 Speaker 8: doss that only engineers could use. Similarly, the lms that 239 00:12:51,400 --> 00:12:54,520 Speaker 8: open Ai and others have introduced make it so that 240 00:12:54,720 --> 00:12:55,880 Speaker 8: anyone can use AI. 241 00:12:56,000 --> 00:12:57,520 Speaker 6: You don't have to be an AI engineer. 242 00:12:57,960 --> 00:13:01,760 Speaker 8: That's why we're in an incredibly important moment because all 243 00:13:01,760 --> 00:13:05,199 Speaker 8: of a sudden, AI is accessible to everyone, and that's 244 00:13:05,240 --> 00:13:05,720 Speaker 8: what shows. 245 00:13:06,480 --> 00:13:09,360 Speaker 2: The data which it runs on is still in many 246 00:13:09,400 --> 00:13:13,040 Speaker 2: ways not perfect. BI sees within how are you thinking 247 00:13:13,080 --> 00:13:17,360 Speaker 2: about the startups you invest in and the overall regulatory 248 00:13:17,440 --> 00:13:19,200 Speaker 2: environment that's going to have to play catch up in 249 00:13:19,240 --> 00:13:19,800 Speaker 2: some way. 250 00:13:20,559 --> 00:13:25,320 Speaker 8: Yes, no doubt, it's the data and you're referring to 251 00:13:25,520 --> 00:13:28,920 Speaker 8: hallucination is the term that's often used. This is absolutely 252 00:13:29,000 --> 00:13:31,480 Speaker 8: top of mind for the enterprises that we're talking to, 253 00:13:31,600 --> 00:13:35,000 Speaker 8: and there's several ways that are currently being addressed from 254 00:13:35,040 --> 00:13:38,760 Speaker 8: an engineering and a company building perspective to improve that. 255 00:13:39,360 --> 00:13:42,560 Speaker 8: And in terms of policy, there's a couple responses. One 256 00:13:42,600 --> 00:13:46,320 Speaker 8: is to step back and say hey, too powerful, too dangerous, 257 00:13:46,640 --> 00:13:49,439 Speaker 8: not me. Another is to just say hey, let it run. 258 00:13:49,920 --> 00:13:52,840 Speaker 8: And the third way, which is what we take at Sequoia, 259 00:13:53,040 --> 00:13:56,280 Speaker 8: what I believe is the right approach, is to say, hey, 260 00:13:56,640 --> 00:14:01,760 Speaker 8: this is incredibly powerful, let's be in evolved let's take 261 00:14:01,920 --> 00:14:05,720 Speaker 8: actions to help form it safely. Let's let's be a 262 00:14:05,760 --> 00:14:08,320 Speaker 8: part of this evolution of this technology so that it 263 00:14:08,360 --> 00:14:11,120 Speaker 8: can be a powerful force of good for humanity. 264 00:14:11,480 --> 00:14:14,200 Speaker 2: You're sounding positive, you're sounding optimistic. In fact, we went 265 00:14:14,200 --> 00:14:16,920 Speaker 2: to our own viewers. We did a Twitter poll, as 266 00:14:16,960 --> 00:14:19,160 Speaker 2: we tend to do every single day, conserting. Just take 267 00:14:19,160 --> 00:14:20,800 Speaker 2: a look at the results, because we ask them whether 268 00:14:20,800 --> 00:14:24,760 Speaker 2: they're pessimist or optimists around AI great for humanity. 269 00:14:24,840 --> 00:14:25,520 Speaker 4: Actually, Ed. 270 00:14:27,200 --> 00:14:30,320 Speaker 3: Said yes, yeah, yeah, And look, they're too early to tell. 271 00:14:30,360 --> 00:14:32,400 Speaker 3: We hear it so often, so constant. You've been in 272 00:14:32,440 --> 00:14:35,520 Speaker 3: this field for a long time. Broadly society is catching up. 273 00:14:35,560 --> 00:14:38,160 Speaker 3: His guest Caroline's point. There's a parallel I want to 274 00:14:38,240 --> 00:14:41,640 Speaker 3: draw with crypto because you are been an active investor 275 00:14:41,680 --> 00:14:45,200 Speaker 3: in the crypto industry. I think the latest was edX markets, right, 276 00:14:45,960 --> 00:14:50,640 Speaker 3: that is a froth and bubble burst cycle. Why will 277 00:14:50,680 --> 00:14:53,480 Speaker 3: we not see the same thing play out here in AI? 278 00:14:54,040 --> 00:14:56,960 Speaker 8: Yes, so first I am a believer in crypto as well, 279 00:14:57,240 --> 00:15:00,480 Speaker 8: But that is a separate trend for you know, further discussion, 280 00:15:00,480 --> 00:15:03,600 Speaker 8: because there's a lot of depth there too on AI. 281 00:15:03,920 --> 00:15:08,000 Speaker 8: This too early to tell point is a critical point, 282 00:15:08,040 --> 00:15:12,280 Speaker 8: and I encourage your audience to use AI and to 283 00:15:12,320 --> 00:15:15,360 Speaker 8: actually start making it a part of their lives. You know, 284 00:15:15,400 --> 00:15:20,080 Speaker 8: there's this kind of trite quote which is AI is 285 00:15:20,120 --> 00:15:21,080 Speaker 8: not going to take your job. 286 00:15:21,320 --> 00:15:22,400 Speaker 6: A human using. 287 00:15:22,200 --> 00:15:25,000 Speaker 8: AI will, and I think that is a lot of 288 00:15:25,120 --> 00:15:27,840 Speaker 8: there's a lot of truth to that. The action that 289 00:15:27,880 --> 00:15:30,600 Speaker 8: should be taken is to start harnessing the power of AI. 290 00:15:30,960 --> 00:15:34,040 Speaker 8: If you're not using AI tools, try them out, see 291 00:15:34,040 --> 00:15:36,560 Speaker 8: how they can make you better internally. 292 00:15:36,520 --> 00:15:37,040 Speaker 6: It's aquoia. 293 00:15:37,160 --> 00:15:41,080 Speaker 8: We talk a lot about augmented intelligence as opposed to 294 00:15:41,280 --> 00:15:42,560 Speaker 8: artificial intelligence. 295 00:15:44,080 --> 00:15:47,280 Speaker 2: Constantine absolutely great to have you. Thank you, Saquiah Capital 296 00:15:47,320 --> 00:15:50,640 Speaker 2: partner Constantine Bula with the optimistic point of view. 297 00:15:50,400 --> 00:15:54,560 Speaker 9: There. 298 00:15:58,600 --> 00:15:59,600 Speaker 7: Time for Tokien tech. 299 00:15:59,600 --> 00:16:03,360 Speaker 3: Across the pond, the United Kingdom's Competition of Markets Authority 300 00:16:03,400 --> 00:16:07,080 Speaker 3: has officially started a merger inquiry into Adobe's twenty billion 301 00:16:07,120 --> 00:16:10,320 Speaker 3: dollar purchase of startup Figma. The agency set a June 302 00:16:10,400 --> 00:16:13,960 Speaker 3: thirtieth deadline for its phase one decision. The deal, of course, 303 00:16:14,000 --> 00:16:17,440 Speaker 3: also ander scrutiny from regulators here in the US. Now, 304 00:16:17,480 --> 00:16:19,960 Speaker 3: TSMC is drawing up plans to build its first chip 305 00:16:19,960 --> 00:16:23,440 Speaker 3: fabrication plant in Europe. According to sources, the chip maker 306 00:16:23,680 --> 00:16:25,600 Speaker 3: is in talks with partners to spend as much as 307 00:16:25,640 --> 00:16:28,400 Speaker 3: eleven billion dollars on the new plant, which would be 308 00:16:28,400 --> 00:16:32,680 Speaker 3: based in Saxony, Germany. And finally, Goldman Sachs is just 309 00:16:32,760 --> 00:16:36,600 Speaker 3: named Kim Posnet as Global head of TMT Banking. According 310 00:16:36,600 --> 00:16:38,960 Speaker 3: to a company memo, she is a veteran deal maker 311 00:16:39,120 --> 00:16:42,960 Speaker 3: whose recent work includes Endeavors Group Indeavor Group's nine point 312 00:16:43,000 --> 00:16:46,400 Speaker 3: three billion dollar takeover of WWE Carrot. 313 00:16:46,800 --> 00:16:48,440 Speaker 4: A great array of news there. 314 00:16:48,480 --> 00:16:51,080 Speaker 2: Meanwhile, let's dig into another part of the news agenda 315 00:16:51,120 --> 00:16:53,040 Speaker 2: that we've got to touch. Because the latest in the 316 00:16:53,040 --> 00:16:55,520 Speaker 2: war on Ukraine, Russia actually saying it is able to 317 00:16:55,560 --> 00:16:58,000 Speaker 2: avert an attack by a pair of drones aimed to 318 00:16:58,040 --> 00:17:01,080 Speaker 2: hit President Vladimir Putin's residence in Most Tuesday night, and 319 00:17:01,400 --> 00:17:04,720 Speaker 2: blame was immediately placed on Ukraine, but without providing evidence. 320 00:17:04,920 --> 00:17:08,520 Speaker 2: State Department Secretary Anthony Blincoln said that he couldn't validate 321 00:17:08,520 --> 00:17:11,120 Speaker 2: the reports and warned to take any Kremlin claims with. 322 00:17:11,119 --> 00:17:14,000 Speaker 4: A large shaker of salt. For now. 323 00:17:14,320 --> 00:17:25,399 Speaker 2: Selinsky has said that this is untrue. Welcome back to 324 00:17:25,400 --> 00:17:27,360 Speaker 2: Blue Blow Technology. I'm Caroline Hyde and New. 325 00:17:27,359 --> 00:17:29,639 Speaker 3: York and I met Lovelo in San Francisco. 326 00:17:29,800 --> 00:17:32,800 Speaker 2: Turning back now to AI ed, of course, we've got 327 00:17:32,840 --> 00:17:35,080 Speaker 2: to do it. Microsoft chief, in fact, the chief economist 328 00:17:35,160 --> 00:17:37,639 Speaker 2: for that particular company, Michael Schwartz, was out with a 329 00:17:37,720 --> 00:17:40,960 Speaker 2: warning about the extent to which artificial intelligence will actould 330 00:17:40,960 --> 00:17:43,000 Speaker 2: be dangerous in the hands of bad actors. He spoke 331 00:17:43,040 --> 00:17:45,280 Speaker 2: earlier at the World Economic Forum panel in Geneva. 332 00:17:45,960 --> 00:17:51,240 Speaker 10: I'm quite confident that yes, AI will be used by 333 00:17:51,320 --> 00:17:56,600 Speaker 10: bad actors, and yes it will cause real damage, and 334 00:17:56,760 --> 00:17:59,960 Speaker 10: yes we have to be very careful and very big 335 00:18:00,520 --> 00:18:05,639 Speaker 10: to avoid they all the means possible, we have to 336 00:18:05,640 --> 00:18:06,680 Speaker 10: put saveguards. 337 00:18:07,960 --> 00:18:11,280 Speaker 2: Let's talk about that mix of cyber and AI. Meta 338 00:18:11,520 --> 00:18:14,200 Speaker 2: out with its quarterly threat report, which showed that malware 339 00:18:14,240 --> 00:18:19,320 Speaker 2: operations using for example, chat GPT, AI help tools as 340 00:18:19,560 --> 00:18:22,320 Speaker 2: phishing links to target businesses and consumers. Joining us now 341 00:18:22,400 --> 00:18:25,119 Speaker 2: as metas head of Security Policy, Nathaniel Gleischer, I just 342 00:18:25,160 --> 00:18:28,160 Speaker 2: want to dig into how new malware strains you're finding 343 00:18:28,480 --> 00:18:34,359 Speaker 2: are basically including posing as chatchypts, browser extensions or productivity tools. 344 00:18:34,440 --> 00:18:37,040 Speaker 2: It builds as though malware bad actors are getting in 345 00:18:37,080 --> 00:18:38,040 Speaker 2: on the AI hype. 346 00:18:39,480 --> 00:18:41,720 Speaker 1: Yeah, thank you so much for having me. I'm here 347 00:18:41,760 --> 00:18:44,840 Speaker 1: today because we are releasing a series of threat reports 348 00:18:44,840 --> 00:18:48,240 Speaker 1: on threats we've countered around the world will help keep 349 00:18:48,280 --> 00:18:50,359 Speaker 1: people safe. One of the threats that I think is 350 00:18:50,400 --> 00:18:53,159 Speaker 1: particularly useful for this business audience is malware, and to 351 00:18:53,200 --> 00:18:56,359 Speaker 1: be thinking about malware are malicious tools that are designed 352 00:18:56,359 --> 00:19:00,000 Speaker 1: to look innofant, but focus on compromising personal devices, compromise 353 00:19:00,119 --> 00:19:02,639 Speaker 1: in your accounts, and we know that they target businesses. 354 00:19:03,000 --> 00:19:06,400 Speaker 1: And as you say, malware developers always try to get 355 00:19:06,400 --> 00:19:09,320 Speaker 1: it on buzzwords. They're looking for ways to trick people 356 00:19:09,359 --> 00:19:12,120 Speaker 1: into downloading malware. We've seen to do it with cryptocurrency 357 00:19:12,160 --> 00:19:14,000 Speaker 1: in twenty twenty two. They try to get people to 358 00:19:14,000 --> 00:19:16,320 Speaker 1: click on links suggesting that you would get access to 359 00:19:16,320 --> 00:19:19,560 Speaker 1: new cryptocurrency apps. Today they're trying to use generative AI, 360 00:19:19,720 --> 00:19:21,720 Speaker 1: suggesting you could click on a link, you might get 361 00:19:21,760 --> 00:19:24,480 Speaker 1: access to chat, GPT or another genderate AI tool, and 362 00:19:24,640 --> 00:19:26,320 Speaker 1: actually you'd be downloading malware. 363 00:19:26,520 --> 00:19:30,240 Speaker 2: Fascinating just who within a business is the most vulnerable, 364 00:19:30,280 --> 00:19:32,199 Speaker 2: who are going to be being sent these links. 365 00:19:33,400 --> 00:19:37,600 Speaker 1: I think businesses should be conscious that malware actors target indiscriminately. 366 00:19:37,760 --> 00:19:40,320 Speaker 1: They will target across the business, including maybe people who 367 00:19:40,320 --> 00:19:44,000 Speaker 1: are connected to prominent leaders or individuals that might just 368 00:19:44,560 --> 00:19:47,080 Speaker 1: be operating in the business and for example, maintaining or 369 00:19:47,119 --> 00:19:50,480 Speaker 1: administering the businesses social media pages. The good news for 370 00:19:50,520 --> 00:19:53,120 Speaker 1: all of this, though, is that there's a number of things, 371 00:19:53,200 --> 00:19:56,159 Speaker 1: fairly simple things that people can do and businesses can 372 00:19:56,160 --> 00:19:58,560 Speaker 1: do to keep themselves safe, whether that's turning on two 373 00:19:58,600 --> 00:20:02,440 Speaker 1: factor authentication or using strong and unique passwords across their 374 00:20:02,480 --> 00:20:05,560 Speaker 1: online accounts. And we are rolling out a series of 375 00:20:05,600 --> 00:20:09,000 Speaker 1: new protections for businesses to help them strengthen their defenses 376 00:20:09,320 --> 00:20:12,600 Speaker 1: on meta and also to help them protect against malware 377 00:20:12,640 --> 00:20:15,280 Speaker 1: wherever it might target them, whether that's on their devices, 378 00:20:15,560 --> 00:20:18,399 Speaker 1: on their other online accounts, or on our platforms. 379 00:20:19,640 --> 00:20:22,000 Speaker 3: The fact out this is a story about bad actors 380 00:20:22,160 --> 00:20:26,800 Speaker 3: posting ururls that purport to take you to attract GPT 381 00:20:27,080 --> 00:20:30,240 Speaker 3: like product. Right, that's what you've detected. What I want 382 00:20:30,240 --> 00:20:34,240 Speaker 3: to know is have you detected threat actors who are 383 00:20:34,280 --> 00:20:38,399 Speaker 3: actually using the technology using generative AI to make the 384 00:20:38,400 --> 00:20:41,600 Speaker 3: content they post on your platforms harder to detect. 385 00:20:43,840 --> 00:20:48,520 Speaker 1: We haven't seen significant sophisticated campaigns relying on genitive AI today. 386 00:20:48,960 --> 00:20:52,560 Speaker 1: We do know that bad actors try to abuse every 387 00:20:52,600 --> 00:20:55,480 Speaker 1: new piece of technology, so we expect that they're going 388 00:20:55,520 --> 00:20:58,160 Speaker 1: to try to abuse generative AI, just like they've tried 389 00:20:58,200 --> 00:21:01,400 Speaker 1: to abuse other innovations in recent years. That's why it's 390 00:21:01,400 --> 00:21:05,040 Speaker 1: particularly important that defender teams at Meta and at other 391 00:21:05,040 --> 00:21:08,360 Speaker 1: companies around the world, including in government and civil society, 392 00:21:08,400 --> 00:21:11,719 Speaker 1: are thinking about how these abuses might happen and steps 393 00:21:11,760 --> 00:21:14,880 Speaker 1: we can take that protect people and counter the bad actors. 394 00:21:16,440 --> 00:21:21,639 Speaker 3: Cisco warned that AI would make phishing attacks harder to detect. 395 00:21:21,840 --> 00:21:23,479 Speaker 3: Do you share Cisco's concern. 396 00:21:25,880 --> 00:21:29,920 Speaker 1: We've seen some limited examples of people using AI based 397 00:21:30,000 --> 00:21:35,520 Speaker 1: tools to, for example, create profile pictures that of people 398 00:21:35,560 --> 00:21:38,560 Speaker 1: who don't exist but look realistic. We've seen some cases 399 00:21:38,560 --> 00:21:40,960 Speaker 1: of people trying to use generative AI tools to produce 400 00:21:40,960 --> 00:21:44,040 Speaker 1: deep fake videos. So these threats are certainly out there, 401 00:21:44,280 --> 00:21:47,120 Speaker 1: and it's important to be ready to counter them. That's 402 00:21:47,200 --> 00:21:49,440 Speaker 1: why you need to have analyst teams that hunt for 403 00:21:49,480 --> 00:21:51,880 Speaker 1: this type of thing. And the other side of generative 404 00:21:51,880 --> 00:21:54,440 Speaker 1: AI is that generative AI has a lot of incredibly 405 00:21:54,560 --> 00:21:58,840 Speaker 1: positive potentials, and one of them is helping defenders who 406 00:21:58,880 --> 00:22:01,679 Speaker 1: are looking to protect public debate improve their detection and 407 00:22:01,720 --> 00:22:04,000 Speaker 1: response systems so we can keep people safe. 408 00:22:04,760 --> 00:22:07,520 Speaker 2: Does it particularly matter where it's coming from, Nathaniel, you 409 00:22:07,680 --> 00:22:13,200 Speaker 2: identify Ductail in particular, that's a Vietnam based cyber criminal operation. 410 00:22:13,520 --> 00:22:14,520 Speaker 4: Is a lot coming from there. 411 00:22:14,520 --> 00:22:17,399 Speaker 2: Can you see where a lot of this attack is 412 00:22:17,600 --> 00:22:18,320 Speaker 2: stemming from? 413 00:22:20,119 --> 00:22:22,560 Speaker 1: Malware development happens all over the world, and it can 414 00:22:22,600 --> 00:22:24,840 Speaker 1: be hard to tell exactly where the actors are where 415 00:22:24,840 --> 00:22:28,760 Speaker 1: they're coming from. When we are able to identify a 416 00:22:28,800 --> 00:22:31,920 Speaker 1: particular company or a particular actor who's behind it, we'll 417 00:22:31,920 --> 00:22:35,119 Speaker 1: take additional steps to deter them and impose cost on them. So, 418 00:22:35,160 --> 00:22:37,600 Speaker 1: for example, for some of the people developing this malware, 419 00:22:37,880 --> 00:22:40,440 Speaker 1: we referred them to law enforcement for further investigation, and 420 00:22:40,480 --> 00:22:43,320 Speaker 1: we send cease and assist letters. We often see or 421 00:22:43,359 --> 00:22:46,720 Speaker 1: sometimes see companies that try to seem legitimate but also 422 00:22:46,880 --> 00:22:48,879 Speaker 1: operate on the dark side of all of this and 423 00:22:48,920 --> 00:22:51,480 Speaker 1: try to share malware. And for companies like that, you 424 00:22:51,520 --> 00:22:54,000 Speaker 1: can impose a lot of cost on them by highlighting this, 425 00:22:54,040 --> 00:22:57,000 Speaker 1: by exposing what they're doing, and by bringing law enforcement 426 00:22:57,000 --> 00:22:58,320 Speaker 1: and government to bear against them. 427 00:22:58,920 --> 00:23:02,360 Speaker 2: Of course, what's thing is, we're wondering about a very 428 00:23:02,400 --> 00:23:07,000 Speaker 2: global cyber attack threat. We're thinking and talking to a 429 00:23:07,080 --> 00:23:10,280 Speaker 2: global company that of course has rechanged its name, talking 430 00:23:10,359 --> 00:23:13,040 Speaker 2: itself about the metaphors, but also deeply ingrained in investing 431 00:23:13,080 --> 00:23:13,920 Speaker 2: in AI in some way. 432 00:23:14,080 --> 00:23:14,119 Speaker 8: Ed. 433 00:23:14,400 --> 00:23:17,560 Speaker 3: Yeah, I mean if Aaniel I covered Meta earnings, Mark 434 00:23:18,280 --> 00:23:23,240 Speaker 3: Zuckerberg really emphasizing Meta's competence in the field of AI. Right, So, 435 00:23:23,280 --> 00:23:27,159 Speaker 3: how do you and the security team use artificial intelligence 436 00:23:27,200 --> 00:23:31,240 Speaker 3: as a tool? How can it help you? 437 00:23:31,400 --> 00:23:35,280 Speaker 1: GENERAI is a novel and fast development technology. There are 438 00:23:35,320 --> 00:23:39,040 Speaker 1: a lot of places where I can have real impact medicine, education, 439 00:23:39,440 --> 00:23:42,080 Speaker 1: But I do think one of the interesting opportunities is 440 00:23:42,119 --> 00:23:44,840 Speaker 1: in what we would call integrity, trust and safety, keeping 441 00:23:44,840 --> 00:23:48,600 Speaker 1: people safe. You can quickly be able to identify and 442 00:23:48,720 --> 00:23:52,480 Speaker 1: track and counter threats around the world. When we're moving 443 00:23:52,520 --> 00:23:55,200 Speaker 1: in a global environment where there are many many countries, 444 00:23:55,240 --> 00:23:57,440 Speaker 1: and there are billions of users, and there are many 445 00:23:57,480 --> 00:24:00,159 Speaker 1: threat actors. Everything we can do to help move more 446 00:24:00,240 --> 00:24:03,919 Speaker 1: quickly encountering these threats is effective. We work to pair 447 00:24:04,040 --> 00:24:07,880 Speaker 1: those types of automated tools with expert human investigators who 448 00:24:07,880 --> 00:24:10,600 Speaker 1: can track and counter the most sophisticated threat actors and 449 00:24:10,680 --> 00:24:13,640 Speaker 1: tinys together. Pairing them up has been most effective encountering 450 00:24:13,680 --> 00:24:15,280 Speaker 1: the bad Guess all. 451 00:24:15,280 --> 00:24:17,679 Speaker 3: Right, Nathaniel Gleischer and Meta, thank you for bringing us 452 00:24:17,720 --> 00:24:21,800 Speaker 3: your latest report. Another story that we're following TikTok's head 453 00:24:21,840 --> 00:24:24,800 Speaker 3: of trust and safety for the US is leaving the company. 454 00:24:24,840 --> 00:24:27,919 Speaker 3: The senior official in charge of ensuring user safety is 455 00:24:27,960 --> 00:24:31,640 Speaker 3: departing amid increased pressure for the US government to ban 456 00:24:31,840 --> 00:24:34,080 Speaker 3: the app. Eric Hahn had been one of the most 457 00:24:34,119 --> 00:24:38,040 Speaker 3: prominent officials in TikTok's efforts to convince lawmakers that the 458 00:24:38,080 --> 00:24:40,199 Speaker 3: app is safe for US users. 459 00:24:40,480 --> 00:24:41,000 Speaker 6: Carolyne I. 460 00:24:41,040 --> 00:24:44,119 Speaker 2: Meanwhile, look coming up, we're more on tech, more on 461 00:24:44,160 --> 00:24:46,000 Speaker 2: competition between China and the US. 462 00:24:46,040 --> 00:24:47,480 Speaker 4: With someone who calls it the. 463 00:24:47,440 --> 00:24:50,320 Speaker 2: Most important war in the next two decades, the techno 464 00:24:50,359 --> 00:24:52,720 Speaker 2: economic war would of course be joined one another than 465 00:24:52,800 --> 00:25:11,360 Speaker 2: Koslaventure's founder Venov Kosla. Time now for our VC round 466 00:25:11,440 --> 00:25:13,800 Speaker 2: up and starting with LinkedIn co founder and Read Hoffman 467 00:25:13,920 --> 00:25:18,000 Speaker 2: and DeepMind co founder Mastuffa Sulliman, who have their own 468 00:25:18,040 --> 00:25:20,280 Speaker 2: AI startup. Who would have thought, And they're rolling out 469 00:25:20,280 --> 00:25:23,639 Speaker 2: a chatbot called PIE, which stands for Personal Intelligence and 470 00:25:23,680 --> 00:25:26,760 Speaker 2: it's intended to serve as a supportive personal companion that 471 00:25:26,840 --> 00:25:28,080 Speaker 2: gives friendly advice. 472 00:25:28,520 --> 00:25:30,359 Speaker 4: Meanwhile, let's look at the VC firm Accel. 473 00:25:30,520 --> 00:25:33,800 Speaker 2: It's refocusing its India Accelerator program on who. 474 00:25:33,640 --> 00:25:35,000 Speaker 4: Guessed at AI startups? 475 00:25:35,080 --> 00:25:36,919 Speaker 2: Now the firm will choose as many as half a 476 00:25:36,960 --> 00:25:39,800 Speaker 2: dozen early stage startups to fund, to mentor in that 477 00:25:39,920 --> 00:25:43,320 Speaker 2: starting this month, is also targeting companies that use technology 478 00:25:43,359 --> 00:25:46,840 Speaker 2: to make conventional industrial processes all the more efficient. 479 00:25:47,240 --> 00:25:50,879 Speaker 3: Ed let's stick with all things bench capital, things AI. 480 00:25:50,960 --> 00:25:54,040 Speaker 3: With Coast the ventures founder Vinode Coaster, who joins us 481 00:25:54,320 --> 00:25:58,119 Speaker 3: for today's VC Spotlight here in San Francisco, Let's go 482 00:25:58,160 --> 00:26:01,320 Speaker 3: to AI. It's all that anyone wants to talk about. 483 00:26:01,320 --> 00:26:05,520 Speaker 3: I mean open ai was the first big check that 484 00:26:05,560 --> 00:26:07,080 Speaker 3: he wrote, right, and. 485 00:26:08,200 --> 00:26:11,720 Speaker 9: We wrote the first very large check in open ai 486 00:26:11,920 --> 00:26:16,359 Speaker 9: that we've done, Yes, first venture capital firm, along with 487 00:26:16,400 --> 00:26:17,440 Speaker 9: some individuals. 488 00:26:17,520 --> 00:26:20,160 Speaker 3: So fast forward to present day. There is a hype 489 00:26:20,160 --> 00:26:24,920 Speaker 3: cycle around US Fisher Intelligence and half of your field 490 00:26:25,280 --> 00:26:27,480 Speaker 3: of vcpaers talk about, well, we've been doing this for 491 00:26:27,520 --> 00:26:30,200 Speaker 3: a decade, and then half getting into it, and you 492 00:26:30,840 --> 00:26:32,880 Speaker 3: how do you assess what's happening? 493 00:26:33,760 --> 00:26:36,639 Speaker 9: Well, everybody wants to get on the bandwagon because it 494 00:26:36,760 --> 00:26:39,400 Speaker 9: is a major trend, and it is a major trend, 495 00:26:39,560 --> 00:26:42,879 Speaker 9: so not surprising people claiming they've been doing it for 496 00:26:42,960 --> 00:26:46,840 Speaker 9: a decade. I wrote about it extensively about ten years ago, 497 00:26:47,119 --> 00:26:49,720 Speaker 9: and then four years ago we invested in open ai. 498 00:26:50,480 --> 00:26:52,919 Speaker 9: Five six years ago, we had a whole bunch of 499 00:26:53,040 --> 00:26:56,600 Speaker 9: other startups we invested in in the same area. 500 00:26:57,760 --> 00:26:59,719 Speaker 7: And so it is an exciting trend. 501 00:27:00,080 --> 00:27:03,520 Speaker 9: It is hard to predict exactly how fast it is 502 00:27:03,560 --> 00:27:06,960 Speaker 9: developing or how fast it will develop. Every day you 503 00:27:07,080 --> 00:27:12,000 Speaker 9: see new changes, but it is exciting, and it's exciting. 504 00:27:11,520 --> 00:27:13,560 Speaker 7: From both a business point of view. 505 00:27:13,600 --> 00:27:16,359 Speaker 9: As well as the technology capability point of view. 506 00:27:16,680 --> 00:27:19,080 Speaker 3: Is open AI the only game in town? Is it 507 00:27:19,160 --> 00:27:21,160 Speaker 3: just a clear leader in this field right now? 508 00:27:21,560 --> 00:27:25,560 Speaker 9: No, in America, there's a lot of interesting startups in 509 00:27:25,600 --> 00:27:28,920 Speaker 9: this area. Google of course is a very credible player, 510 00:27:29,040 --> 00:27:32,639 Speaker 9: and then there's other startups that are credible too, trying 511 00:27:32,680 --> 00:27:35,240 Speaker 9: to make a go at it. I think the leader 512 00:27:35,359 --> 00:27:37,480 Speaker 9: always has a huge advantage. 513 00:27:38,320 --> 00:27:41,600 Speaker 2: What about leadership coming from China, because I know you 514 00:27:41,720 --> 00:27:44,000 Speaker 2: keep a keen eye on what you really see as 515 00:27:44,680 --> 00:27:47,440 Speaker 2: a competitive threat, the new war in the next couple 516 00:27:47,480 --> 00:27:49,440 Speaker 2: of decades, techno economic. 517 00:27:50,400 --> 00:27:55,439 Speaker 9: Oh absolutely, I probably worry more about a Chinese AI 518 00:27:55,720 --> 00:28:00,560 Speaker 9: doing bad things than anything to do with AI itself 519 00:28:00,680 --> 00:28:03,280 Speaker 9: going sentient or the kinds of things a lot must 520 00:28:03,359 --> 00:28:10,240 Speaker 9: talks about. Not that that's not a concern. China is 521 00:28:10,320 --> 00:28:12,800 Speaker 9: going to be a big enemy in this. I believe 522 00:28:12,840 --> 00:28:15,760 Speaker 9: we are in a technoeconomic war with China for economic 523 00:28:15,800 --> 00:28:19,440 Speaker 9: dominance and hence dominance of political systems to go with it. 524 00:28:19,720 --> 00:28:22,760 Speaker 2: Okay, So do the risk reward for us here at 525 00:28:22,760 --> 00:28:25,439 Speaker 2: the moment or balance the risks when you're talking to 526 00:28:25,520 --> 00:28:28,240 Speaker 2: regulators as you have done if you talk to people 527 00:28:28,240 --> 00:28:30,879 Speaker 2: in power. You've been over there in Washington. You had 528 00:28:31,119 --> 00:28:33,720 Speaker 2: coming together of course the Hell and Valley Forum, where 529 00:28:33,720 --> 00:28:37,520 Speaker 2: you gathered vcs and tech community to talk to Well, government, 530 00:28:38,000 --> 00:28:41,160 Speaker 2: do you want them to regulate how we use AI 531 00:28:41,320 --> 00:28:43,120 Speaker 2: in the wild in the US at the moment? Or 532 00:28:43,120 --> 00:28:45,280 Speaker 2: are you more worried about the regulation of China and 533 00:28:45,360 --> 00:28:47,240 Speaker 2: money flowing there to build their own AI. 534 00:28:47,680 --> 00:28:50,760 Speaker 9: Well, I'm much more worried about the competitive picture with 535 00:28:50,960 --> 00:28:55,040 Speaker 9: China than about regulation here. I think it's too early 536 00:28:55,120 --> 00:28:59,720 Speaker 9: to regulate AI. We don't know enough about it. I 537 00:28:59,840 --> 00:29:04,080 Speaker 9: do think we should heavily fund safety research, especially in 538 00:29:04,120 --> 00:29:09,280 Speaker 9: the universities and third parties, but regulating it would be 539 00:29:09,320 --> 00:29:13,200 Speaker 9: a real mistake and put us behind China in the race. 540 00:29:14,000 --> 00:29:17,240 Speaker 2: And the race it does feel is something that vcs 541 00:29:17,280 --> 00:29:19,640 Speaker 2: that Peter til that Venola you're talking about ed and 542 00:29:20,000 --> 00:29:23,080 Speaker 2: also a relationship that's being built a little bit more 543 00:29:23,120 --> 00:29:25,800 Speaker 2: firmly between DC and Silicon Valley. 544 00:29:25,800 --> 00:29:28,080 Speaker 4: Where you sit, Yeah, I think for me. 545 00:29:28,440 --> 00:29:30,680 Speaker 3: You know, the question Vinode is what was the upshot 546 00:29:31,120 --> 00:29:34,680 Speaker 3: of the Hill and Valley forum. You know that we 547 00:29:35,040 --> 00:29:38,560 Speaker 3: reported on it. We discussed about the relationship between DC 548 00:29:39,160 --> 00:29:40,200 Speaker 3: and Silicon Valley. 549 00:29:40,640 --> 00:29:41,880 Speaker 6: What has happened as a result. 550 00:29:42,560 --> 00:29:45,640 Speaker 9: Well, there's clearly a lot of interest in this question, 551 00:29:45,840 --> 00:29:50,960 Speaker 9: not only AI and AI regulation, but equally importantly on 552 00:29:51,040 --> 00:29:55,520 Speaker 9: our AI race with China. My interest is highlighting how 553 00:29:55,560 --> 00:29:59,440 Speaker 9: important that issuers and how much we should focus on 554 00:29:59,560 --> 00:30:05,040 Speaker 9: the race with China for technology supremacy here. I do 555 00:30:05,080 --> 00:30:07,480 Speaker 9: think there'll be multiple players, but I do think there 556 00:30:07,520 --> 00:30:11,840 Speaker 9: will be winners that take a disproportionate part of the 557 00:30:11,880 --> 00:30:15,480 Speaker 9: economic pie. And I think the US model is well 558 00:30:15,520 --> 00:30:18,560 Speaker 9: set up for it. But it's not obvious that we 559 00:30:18,680 --> 00:30:21,400 Speaker 9: can win or we will win in twenty years from now. 560 00:30:22,760 --> 00:30:24,560 Speaker 3: He said economic pie. So I'm going to do a 561 00:30:24,600 --> 00:30:27,440 Speaker 3: small pivot. There is a FED meeting today. We expect 562 00:30:27,440 --> 00:30:31,120 Speaker 3: a twenty five basis point hike. We talk daily about 563 00:30:31,200 --> 00:30:36,400 Speaker 3: tighter financial conditions. VC of your experience in your scale, 564 00:30:36,800 --> 00:30:40,400 Speaker 3: how does that impact the ecosystem that you are operating? 565 00:30:40,440 --> 00:30:41,920 Speaker 7: In well fed. 566 00:30:42,000 --> 00:30:46,640 Speaker 9: Rate hikes affect short term things. You know what we 567 00:30:46,720 --> 00:30:49,920 Speaker 9: invest in open AI four years ago. We don't expect 568 00:30:49,960 --> 00:30:53,480 Speaker 9: liquidity for six, seven, eight years, So I'd have to 569 00:30:53,520 --> 00:30:56,320 Speaker 9: predict the market in twenty thirty and the federate in 570 00:30:56,400 --> 00:30:59,680 Speaker 9: twenty thirty to really have it impact our. 571 00:30:59,680 --> 00:31:02,160 Speaker 7: Very popular within this organization, if you could. 572 00:31:03,800 --> 00:31:06,440 Speaker 9: So, my goal is really to focus on what adds 573 00:31:06,480 --> 00:31:11,360 Speaker 9: substantial societal value and because of that economic value and 574 00:31:11,440 --> 00:31:15,480 Speaker 9: what will be true in terms of real value active 575 00:31:15,560 --> 00:31:20,000 Speaker 9: society in twenty thirty if I'm going to make good investments. 576 00:31:20,400 --> 00:31:23,680 Speaker 9: So our focus is much longer term than short term 577 00:31:23,720 --> 00:31:26,760 Speaker 9: interest rate types and not it's not that enable more 578 00:31:26,880 --> 00:31:29,320 Speaker 9: or less financing to happen in the interim, So we 579 00:31:29,360 --> 00:31:32,440 Speaker 9: do pay attention to it, but mostly we ignore it. 580 00:31:32,560 --> 00:31:35,520 Speaker 3: Yeah, financing an interim, I mean caroline for all the 581 00:31:35,600 --> 00:31:40,520 Speaker 3: volatility in public markets, and from a policy perspective, we 582 00:31:40,560 --> 00:31:44,640 Speaker 3: see checks getting written by vcs to venture founded startups. 583 00:31:44,360 --> 00:31:46,560 Speaker 2: And largely in the realm of AI at the moment 584 00:31:46,560 --> 00:31:49,000 Speaker 2: of VINOD To that extent, are you writing a lot 585 00:31:49,040 --> 00:31:51,560 Speaker 2: of checks? What sort of size of company you're most 586 00:31:51,640 --> 00:31:53,040 Speaker 2: looking for at the moment? In this new. 587 00:31:53,000 --> 00:31:57,360 Speaker 7: Environment, we are writing a lot of checks. We are 588 00:31:57,360 --> 00:31:58,240 Speaker 7: being aggressive. 589 00:31:58,440 --> 00:32:02,600 Speaker 9: But I would say there's many more bad AI startups 590 00:32:03,240 --> 00:32:07,320 Speaker 9: than good startups, and it's very hard to differentiate if 591 00:32:07,360 --> 00:32:11,000 Speaker 9: you're not experienced with AI. So I do think lots 592 00:32:11,000 --> 00:32:14,640 Speaker 9: of bad investments will be made, but overall more money 593 00:32:15,000 --> 00:32:19,640 Speaker 9: will be made than lost. What even if ninety percent 594 00:32:19,680 --> 00:32:21,160 Speaker 9: of the startups failed, which they. 595 00:32:21,160 --> 00:32:24,440 Speaker 2: Will interesting and of course that's often the bet, no, 596 00:32:24,760 --> 00:32:27,160 Speaker 2: that you will have failures, but you'll then have from 597 00:32:27,240 --> 00:32:31,560 Speaker 2: those absolutely extraordinary success that gives the money and the 598 00:32:31,640 --> 00:32:34,680 Speaker 2: returns to vcs. But you know, in this current environment, 599 00:32:34,680 --> 00:32:37,480 Speaker 2: we're talking a lot about so called zombie companies, the 600 00:32:37,480 --> 00:32:40,160 Speaker 2: companies that have massive valuations that can't vindicate the growth 601 00:32:40,200 --> 00:32:43,120 Speaker 2: at the moment, having to make cutbacks, having to think 602 00:32:43,200 --> 00:32:46,800 Speaker 2: more about revenue driving, profit driving. What are you doing 603 00:32:46,840 --> 00:32:49,479 Speaker 2: with those companies at the moment that just had huge 604 00:32:49,600 --> 00:32:52,120 Speaker 2: valuations that perhaps don't have the run rate that they 605 00:32:52,200 --> 00:32:52,440 Speaker 2: used to. 606 00:32:54,120 --> 00:32:59,160 Speaker 9: Well, there are clearly founder choices to be made. Good 607 00:32:59,240 --> 00:33:03,120 Speaker 9: founders are responding to the environment by saying, let's ignore 608 00:33:03,120 --> 00:33:06,440 Speaker 9: what happened in the past, let's ignore our old valuation 609 00:33:06,760 --> 00:33:10,239 Speaker 9: and focus in on how we create value for the 610 00:33:10,280 --> 00:33:14,280 Speaker 9: next few years. And not worry about what the valuation was. 611 00:33:14,360 --> 00:33:18,200 Speaker 9: So founders that fundamentally decide whether to cater to investors 612 00:33:18,760 --> 00:33:21,400 Speaker 9: or to the reality of the business they're in. And 613 00:33:21,680 --> 00:33:24,440 Speaker 9: most of the good founders I know are really focusing 614 00:33:24,480 --> 00:33:26,920 Speaker 9: in it in on what they need to do to 615 00:33:26,920 --> 00:33:30,640 Speaker 9: build a great business and hope that investors will get 616 00:33:30,680 --> 00:33:34,760 Speaker 9: on board later and not worry about short term perception 617 00:33:34,960 --> 00:33:35,640 Speaker 9: of valuation. 618 00:33:36,840 --> 00:33:40,000 Speaker 3: There are other deeply technical areas that you are passionate 619 00:33:40,040 --> 00:33:43,200 Speaker 3: about and that you are investing in climate or energy 620 00:33:43,240 --> 00:33:46,840 Speaker 3: related and one of them is fusion. You've written op 621 00:33:47,000 --> 00:33:51,480 Speaker 3: eds about this because AI is dominated headlines. We talked 622 00:33:51,480 --> 00:33:53,720 Speaker 3: about it less, but can you just explain why you're 623 00:33:53,720 --> 00:33:55,040 Speaker 3: so focused on that area. 624 00:33:55,640 --> 00:33:58,080 Speaker 9: Well, if you look at Commonwealth Fusion, which is an 625 00:33:58,120 --> 00:34:02,120 Speaker 9: investment we made about four years ago, five years ago 626 00:34:02,760 --> 00:34:06,280 Speaker 9: before there was a company. Actually we started working with 627 00:34:06,360 --> 00:34:08,960 Speaker 9: Bob Mumguard when he was a senior fellow at the 628 00:34:09,120 --> 00:34:10,360 Speaker 9: MIT Fusion Lab. 629 00:34:11,920 --> 00:34:14,439 Speaker 7: It was very clear if you cracked. 630 00:34:14,120 --> 00:34:17,040 Speaker 9: Fusion as a technology, it was one of the largest 631 00:34:17,200 --> 00:34:19,800 Speaker 9: markets in the world, much much larger than for example, 632 00:34:19,880 --> 00:34:24,480 Speaker 9: Google's market. Because energy markets are very large, and so 633 00:34:25,000 --> 00:34:27,680 Speaker 9: that is exciting. You can lose one time, So your 634 00:34:27,719 --> 00:34:29,760 Speaker 9: money you can make a thousand times your money. 635 00:34:29,760 --> 00:34:31,080 Speaker 7: That's a pretty good trade off. 636 00:34:31,400 --> 00:34:34,560 Speaker 3: Then I brew up against commercial breaks very quickly. Is 637 00:34:34,600 --> 00:34:36,480 Speaker 3: it a good time to be a bench catalyst. 638 00:34:36,680 --> 00:34:38,719 Speaker 7: I think it's a great time to be an a 639 00:34:38,800 --> 00:34:39,759 Speaker 7: venture capitalist. 640 00:34:40,160 --> 00:34:44,320 Speaker 9: Most of the large companies are cutting off their advanced projects, 641 00:34:44,960 --> 00:34:48,920 Speaker 9: and so the best people leave start companies. That's the 642 00:34:49,000 --> 00:34:52,800 Speaker 9: raw material. We need great talent, and there's plenty of capital. 643 00:34:52,920 --> 00:34:55,120 Speaker 9: So it's a great time. The next five years will 644 00:34:55,120 --> 00:34:56,640 Speaker 9: look really, really good. 645 00:34:57,080 --> 00:34:59,200 Speaker 4: You've got to be optimistic in this game. It's great. 646 00:34:59,200 --> 00:35:00,520 Speaker 4: Tows in time with you. Thank you. 647 00:35:00,640 --> 00:35:11,960 Speaker 2: Vernodekoes of Kosler Ventures, the founder there. Of course, it's 648 00:35:11,960 --> 00:35:15,520 Speaker 2: been going viral. Nordstrom plans to give up its store 649 00:35:15,560 --> 00:35:18,040 Speaker 2: in downtown San Francisco. The company will vacate more than 650 00:35:18,040 --> 00:35:20,680 Speaker 2: three hundred thousand square feet of space at Westfield San 651 00:35:20,680 --> 00:35:24,160 Speaker 2: Francisco Center that's on Market Street, shopping and tourist area 652 00:35:24,280 --> 00:35:26,839 Speaker 2: in the heart of downtown. It also plans to get 653 00:35:26,880 --> 00:35:30,400 Speaker 2: this shut it's nearby Nordstrom Rack Now. The move follows 654 00:35:30,400 --> 00:35:33,719 Speaker 2: others like Office Depot Anthropology also shutting down the Market 655 00:35:33,760 --> 00:35:37,040 Speaker 2: Street adjacent locations. It says the city just struggles with 656 00:35:37,080 --> 00:35:40,200 Speaker 2: empty office buildings and crime concerns, and ed I felt 657 00:35:40,200 --> 00:35:42,560 Speaker 2: that like having lived there briefly in twenty sixteen and 658 00:35:42,560 --> 00:35:44,440 Speaker 2: then coming back to see you as I love to do, 659 00:35:45,200 --> 00:35:48,040 Speaker 2: Market Street is just an extraordinary place. Nowadays, there is 660 00:35:48,120 --> 00:35:49,759 Speaker 2: nothing to go to, nothing to buy, and when you 661 00:35:49,800 --> 00:35:51,799 Speaker 2: do go into the Macy's there, for example. 662 00:35:51,440 --> 00:35:51,960 Speaker 4: It's empty. 663 00:35:52,719 --> 00:35:54,319 Speaker 3: Yeah, and you know it follows on from the Whole 664 00:35:54,320 --> 00:35:57,239 Speaker 3: Food's decision right to close that downtown store around a 665 00:35:57,320 --> 00:35:59,759 Speaker 3: year after it opened. There is a big discussion in 666 00:35:59,800 --> 00:36:02,960 Speaker 3: this city about what went wrong on Market Street. Remember 667 00:36:03,000 --> 00:36:06,040 Speaker 3: they tried to bring in tech Twitter and Uber's offices. 668 00:36:06,080 --> 00:36:09,200 Speaker 3: They gave them incentives to do so to regenerate the area, 669 00:36:09,280 --> 00:36:10,320 Speaker 3: but it hasn't worked. 670 00:36:11,440 --> 00:36:13,880 Speaker 4: What do you make though of the cynicism that you see. 671 00:36:13,800 --> 00:36:16,000 Speaker 2: In social media as it goes viral, everyone sort of 672 00:36:16,000 --> 00:36:18,160 Speaker 2: wringing their hands. Do we need to be as desperate 673 00:36:18,600 --> 00:36:19,960 Speaker 2: as some of the sentiments seems? 674 00:36:20,640 --> 00:36:22,239 Speaker 3: Yeah, Look, I think that there is a lot of 675 00:36:22,280 --> 00:36:24,680 Speaker 3: people calling on the mayor in the DA to take 676 00:36:24,760 --> 00:36:27,399 Speaker 3: action at street level to fix that part of town. 677 00:36:27,600 --> 00:36:30,280 Speaker 3: But there is definitely a tech and broader economic story 678 00:36:30,320 --> 00:36:32,840 Speaker 3: there that will continue to cover on this program. 679 00:36:33,000 --> 00:36:35,120 Speaker 4: And that does it for this edition of Bluebow Technology.