WEBVTT - Tech Gains Again Ahead of Nvidia Results and OpenAI Fires Back on Regulation 0:00:01.680 --> 0:00:05.360 From the heart where innovation, money and power Collie in 0:00:05.440 --> 0:00:06.920 Silicon Valley NBN. 0:00:07.280 --> 0:00:10.640 This is Bloomberg Technology with Caroline Hyde and. 0:00:10.720 --> 0:00:11.800 Ed Ludlow. 0:00:25.079 --> 0:00:27.960 From New York and San Francisco. This is Bloomberg Technology 0:00:28.360 --> 0:00:31.160 and tech Talks are on a roll for another weekly gain. 0:00:31.200 --> 0:00:33.959 As fedshare pal says the time has come for raycuts 0:00:34.240 --> 0:00:35.559 and read from Jackson Hole. 0:00:36.440 --> 0:00:39.600 Harris backs America in AI and the space race. At 0:00:39.640 --> 0:00:43.440 the close of the DNC, we have the reaction from Chicago, an. 0:00:43.280 --> 0:00:46.400 Open AI opposition. We are joined by California State Senator 0:00:46.440 --> 0:00:51.040 Scott Wiener as industry leaders push against his proposed AI legislation. 0:00:51.920 --> 0:00:54.520 First as check in on these markets, and it's all 0:00:54.560 --> 0:00:57.600 about fed share pow. It's all about an increase of 0:00:57.600 --> 0:01:00.000 eight ten percent now notable than in fact, the NASA 0:01:00.320 --> 0:01:02.400 one hundred is actually coming off of its highs of 0:01:02.520 --> 0:01:05.400 earlier trading. This really is just showing that there is 0:01:05.440 --> 0:01:07.360 a rebound. I hope for these rate cuts to kick 0:01:07.360 --> 0:01:10.160 in as soon as September, but we suddenly saw actually 0:01:10.840 --> 0:01:12.840 we're joining a light on some of the French companies 0:01:12.840 --> 0:01:14.640 here and the pound overall. I was going to show 0:01:14.640 --> 0:01:15.920 you what the bomb market was up to when the 0:01:15.959 --> 0:01:18.840 fact that crypto has also been on the higher side. Edd, 0:01:18.840 --> 0:01:20.600 But what are you watching on the micro basis? 0:01:21.400 --> 0:01:23.920 I want to go straight to Nvidia because we've talked 0:01:23.959 --> 0:01:26.960 about how in Video's earnings next Wednesday, August twenty eight 0:01:27.040 --> 0:01:30.480 will be a macro level event and we'll talk a 0:01:30.520 --> 0:01:32.400 bit more about the markets. Right, Cara, were on the 0:01:32.440 --> 0:01:35.440 heels of that Powell statement, which is highly anticipated. 0:01:36.319 --> 0:01:39.440 It certainly was. We were all focusing in on what 0:01:39.560 --> 0:01:42.040 was occurring with Jackson Holme more broadly, we're all thinking 0:01:42.040 --> 0:01:44.760 about what this meant for the tech sector. Look, it's 0:01:44.760 --> 0:01:46.759 no surprises once again, when you're looking at the NASAK 0:01:46.760 --> 0:01:48.800 and the NaSTA one hundred. It is the key players 0:01:48.840 --> 0:01:50.560 that we're so used to leading us high. It is 0:01:50.560 --> 0:01:52.440 the in videos of this world. And maybe that's the 0:01:52.440 --> 0:01:54.480 macro event for your and my market at the moment 0:01:54.520 --> 0:01:57.000 where all eyes on August of twenty eighth, in Videos 0:01:57.040 --> 0:02:00.880 earnings coming and indeed what that means for the picture 0:02:01.040 --> 0:02:03.920 for AI. Not much of that coming from Jackson Hole today, 0:02:03.920 --> 0:02:06.240 of course, there is much more a signal from fed 0:02:06.320 --> 0:02:08.440 chair pal about the fact that the time has come 0:02:08.440 --> 0:02:10.840 for rate cuts. The time has been assessed that really 0:02:10.880 --> 0:02:12.920 we are seeing that weakness in the labor market that 0:02:13.000 --> 0:02:15.640 he doesn't want to accentuate any further. I'm pleased to 0:02:15.639 --> 0:02:17.840 say that to get a weaving of the narrative of 0:02:17.919 --> 0:02:20.760 what comes from a macro policy perspective and how it 0:02:20.840 --> 0:02:23.639 hits the micro in terms of individual movers, we can 0:02:23.639 --> 0:02:26.320 bring in one jess menton look, I mean, we're coming 0:02:26.320 --> 0:02:29.120 off of our highs, but certainly what fed chair pal 0:02:29.240 --> 0:02:32.240 said fueled the market of setting some of yesterday's weakness, 0:02:32.240 --> 0:02:32.960 but only some of it. 0:02:33.040 --> 0:02:34.760 That's right. And even though we've seen some of the 0:02:34.800 --> 0:02:36.720 indexes come off some of those highs, still the S 0:02:36.720 --> 0:02:38.680 and P five hundred is a whisker away from that 0:02:38.720 --> 0:02:41.800 all time high on July tenth, and the Nasdaq one 0:02:41.840 --> 0:02:43.679 hundred's still about four percent away from that. But we 0:02:43.720 --> 0:02:45.760 want to see underneath the hood when it comes to 0:02:45.800 --> 0:02:48.960 tech and what's rallying, it's actually the unprofitable tech companies. 0:02:48.960 --> 0:02:51.480 So Goldman Sacher actually has a basket that tracks that 0:02:51.520 --> 0:02:53.680 it's actually at more than two percent at this point. 0:02:53.720 --> 0:02:55.760 So then you're seeing other corners of the stock market 0:02:55.800 --> 0:02:58.799 that obviously are going to benefit from easing policy. Here, 0:02:58.840 --> 0:03:01.400 see like the Russell tooth and small caps taking off 0:03:01.600 --> 0:03:04.399 over two percent, but then also companies with weaker type 0:03:04.440 --> 0:03:07.640 balance sheets because there's different baskets that city, and then 0:03:07.720 --> 0:03:10.160 others track as well, and you're seeing those outperforming right 0:03:10.160 --> 0:03:10.760 now too. 0:03:11.440 --> 0:03:13.760 Jess, I always go straight to niglobe wrap on the 0:03:13.800 --> 0:03:16.120 Bloomberg terminal. I just look at the top line and 0:03:16.160 --> 0:03:18.320 I say, how does the Bloomberg USERM feel about this? 0:03:18.360 --> 0:03:20.680 And they call it the pal pivot. Maybe it's more 0:03:20.720 --> 0:03:22.560 of like a fed victory lap. I think is some 0:03:22.639 --> 0:03:25.400 of the narrative from Jackson Hole. But let's please go 0:03:25.520 --> 0:03:28.120 back to Nvidia. We kind of knew this would happen 0:03:28.120 --> 0:03:30.640 at some point, a rate cut being on the table. 0:03:30.919 --> 0:03:33.080 August twenty eighth in video is our super Bowl here 0:03:33.120 --> 0:03:34.800 on Bloomberg Technology. 0:03:34.560 --> 0:03:37.400 Of course, and as you know, Esha Day on our 0:03:37.440 --> 0:03:40.000 equities team actually did the taking stot Colum this morning. 0:03:40.160 --> 0:03:42.320 But if you look at what city's pricing in stuke 0:03:42.360 --> 0:03:44.600 Kaiser when it comes to the potential swing here in 0:03:44.640 --> 0:03:47.480 the options market, betting on what we see because to 0:03:47.520 --> 0:03:49.800 your point, in video reports after the bell on Wednesday, 0:03:49.840 --> 0:03:52.880 so the following day during intra day trading, the swing 0:03:52.920 --> 0:03:55.400 could be around ten percent in either direction for Nvidia 0:03:55.440 --> 0:03:58.360 in particular, so that's actually above its longer term average 0:03:58.360 --> 0:04:00.720 of about seven and a half percent. What that means 0:04:00.800 --> 0:04:03.280 is what does it mean for other potential chip stike sucks, 0:04:03.400 --> 0:04:05.440 which obviously a lot of them have come down because 0:04:05.440 --> 0:04:08.240 we had that sell off obviously after things peaked out 0:04:08.240 --> 0:04:10.480 for the NASUAQ one hundred back in July. But of 0:04:10.480 --> 0:04:12.240 course we've seen a bounce back here, so it'll be 0:04:12.240 --> 0:04:14.600 interesting to see as far as even though valuations came 0:04:14.640 --> 0:04:17.000 down a bit, they came back up slightly because of 0:04:17.000 --> 0:04:18.800 the rebound we saw in the past couple of weeks, 0:04:18.800 --> 0:04:21.800 so that'll really shape the direction. Even though volatility is 0:04:21.880 --> 0:04:24.159 expected to fade a bit over the next couple of weeks, 0:04:24.200 --> 0:04:26.719 just because heading into obviously Labor Day, that's still going 0:04:26.760 --> 0:04:29.000 to be a big volatile move potentially next Thursday, after 0:04:29.000 --> 0:04:31.960 we hear from Nvidia after the bell on Wednesday, and in. 0:04:32.000 --> 0:04:34.479 Video up almost four percent, not at session highs, but 0:04:34.520 --> 0:04:38.000 still op significantly. Bloomberg's just men and Happy Friday, okay. 0:04:38.040 --> 0:04:42.560 With the US market heavily concentrated in tech hardware in particular, 0:04:42.880 --> 0:04:46.120 especially in Vidia, investors are looking ahead to the company's 0:04:46.160 --> 0:04:49.760 quarterly results. As we said August twenty eighth, Wednesday, after market, 0:04:49.839 --> 0:04:53.000 let's bring in Alison Porter. She's a portfolio manager at 0:04:53.040 --> 0:04:57.200 Janis Henderson Investors covering their global technology leaders and sustainable 0:04:57.200 --> 0:05:01.279 future technology funds. And that's the date of play. Jackson 0:05:01.320 --> 0:05:04.160 Hole has been and gone in videos next Wednesday, which 0:05:04.200 --> 0:05:05.520 is top of mind for you, Allison. 0:05:07.120 --> 0:05:09.560 I think you know we very much focus on the 0:05:09.640 --> 0:05:13.560 bottom up and focus on where the technology, sayclers, and 0:05:13.640 --> 0:05:17.200 the key focus is going to be on earnings next week, 0:05:17.760 --> 0:05:21.440 and you know, seeing if the enterprise demand pick up 0:05:21.480 --> 0:05:23.920 that we saw some signs of that from both Cesesco 0:05:24.080 --> 0:05:28.080 and the novel in recent weeks. If that carries through, 0:05:28.160 --> 0:05:31.240 not just send video, but we have Dale and network 0:05:31.240 --> 0:05:35.080 Applianes and Upure storage or report next week as well. 0:05:36.000 --> 0:05:38.719 Okay, I love how you're going bottom up here, Allison. 0:05:39.320 --> 0:05:43.719 Does any conversation around the dinner table or conversation at 0:05:43.760 --> 0:05:45.880 big conferences that you're going to at the moment end 0:05:45.920 --> 0:05:48.440 up with people saying, boy, your industry is just sucking 0:05:48.520 --> 0:05:51.640 up all the oxygen but also all the passive money. 0:05:51.720 --> 0:05:54.240 And is there a worry about indexes having to come 0:05:54.279 --> 0:05:55.400 in and rebalance here in some. 0:05:55.400 --> 0:06:00.279 Way, I think we always come back to fundamental and 0:06:00.320 --> 0:06:02.920 there's always a focus on you know, will interest rates 0:06:03.200 --> 0:06:08.080 you know, result in this rotation into smaller caps, is 0:06:08.080 --> 0:06:10.040 they're necessarily going to be bread And I think we 0:06:10.160 --> 0:06:13.440 look back if you look back even to nineteen ninety 0:06:14.520 --> 0:06:18.080 when you see you know, trillion dollar companies, the pace 0:06:18.120 --> 0:06:21.159 at which they've reached two trillion has been faster and 0:06:21.200 --> 0:06:24.400 more likely than companies at one hundred billion going to 0:06:24.440 --> 0:06:27.440 two hundred billions. So technology always tends to be quite 0:06:27.760 --> 0:06:31.560 size agnostic, and those large cap winners, when they achieve 0:06:31.680 --> 0:06:36.559 scale and a key competitive advantages, can continue to grow 0:06:36.680 --> 0:06:42.040 on that in spite of the interest rate environment through 0:06:42.240 --> 0:06:45.520 rising and falling rates. So we're quite agnostic about the 0:06:45.520 --> 0:06:49.280 size of companies, and you know the indexes, you know 0:06:49.520 --> 0:06:52.200 just now there are winners and losers in those large 0:06:52.200 --> 0:06:55.039 cap names. When you look back to prior cycles, not 0:06:55.120 --> 0:06:58.200 all of the past leaders become the leaders in the 0:06:58.279 --> 0:07:00.120 next cycle. So we think it's a great time to 0:07:00.200 --> 0:07:03.400 be an active investor and actually choose which of those 0:07:03.440 --> 0:07:05.360 companies are going to lead in this next week. 0:07:05.680 --> 0:07:09.560 So does Inveria just say the main company that benefits 0:07:09.600 --> 0:07:12.880 from a picks and shovel perspective, Do others win and indeed, 0:07:12.880 --> 0:07:15.240 when does the investment coming from the hypher scale is 0:07:15.240 --> 0:07:16.160 payoff for them too? 0:07:17.400 --> 0:07:19.840 Yeah, I mean I think, you know, for Nvidia, we 0:07:19.880 --> 0:07:22.480 had many of the same questions about Apple during the 0:07:22.520 --> 0:07:27.560 mobile era. Could Apple continue to win? And we do 0:07:27.640 --> 0:07:31.440 think that the majority of the value in parallel processing 0:07:31.960 --> 0:07:35.040 is accruing to in Nvidia, but that is the majority 0:07:35.080 --> 0:07:37.360 of it and not all. And we will through time, 0:07:37.480 --> 0:07:40.640 just as we've seen previously, see a broadening out of 0:07:40.840 --> 0:07:45.720 beneficiaries and on that return on investment spend on AI. 0:07:46.160 --> 0:07:48.280 You know, just as recently as it a few days ago, 0:07:48.320 --> 0:07:51.600 we held Walmart actually talk about the huge benefits to 0:07:51.640 --> 0:07:55.560 their product catalog being able to you know, take eight 0:07:55.640 --> 0:07:58.720 hundred and fifty million pieces of data that would have 0:07:58.720 --> 0:08:01.200 taken them a hundred times ahead count before, so real 0:08:01.280 --> 0:08:06.000 productivity gains that is coming from that AI spend. We've 0:08:06.040 --> 0:08:11.320 heard Andy Jazz's blogged today talking similarly about the number 0:08:11.360 --> 0:08:14.880 of developers that they can save with faster coding. So 0:08:14.920 --> 0:08:18.200 we think that will continue to encourage spend and that 0:08:18.280 --> 0:08:21.600 broadening of spend not just to the hyperscalers, but eventually 0:08:21.960 --> 0:08:24.520 to enterprises as well. 0:08:24.640 --> 0:08:26.480 To that point, Jackie and the team in New York, 0:08:26.560 --> 0:08:28.440 let's bring back that chart you had a second ago, 0:08:28.560 --> 0:08:31.080 some of the megacaps and their performance in the session. 0:08:31.080 --> 0:08:33.760 Because Alison, in your biggest fund these are all names 0:08:33.800 --> 0:08:37.800 you hold Microsoft, Amazon, Alphabet and Meta not in that 0:08:37.960 --> 0:08:41.040 ranking or order in terms of allocation. When all is 0:08:41.080 --> 0:08:44.280 said and done August twenty ninth, what would need to 0:08:44.320 --> 0:08:47.880 happen within Video's earnings for you to rethink your allocation 0:08:48.040 --> 0:08:49.800 to each of those names going forward? 0:08:51.800 --> 0:08:53.560 You know, I think even if you look back at 0:08:53.679 --> 0:08:58.240 last year, we don't react just to one quarter of 0:08:58.320 --> 0:09:00.960 earnings because a lot of possessions and come into that. 0:09:01.520 --> 0:09:04.600 But we think we've already seen evidence from all of 0:09:04.640 --> 0:09:07.559 the reports of those companies that you just saw that 0:09:07.600 --> 0:09:11.320 they are committed to continuing to spend where they see 0:09:11.360 --> 0:09:14.840 a return, and we think in Video's results will just 0:09:14.920 --> 0:09:19.240 continue to show that they are seeing strong visibility and 0:09:19.240 --> 0:09:21.760 seeing that broad and so, you know, we don't really 0:09:21.760 --> 0:09:24.400 see that as being an event from Nvidia that's going 0:09:24.440 --> 0:09:27.760 to materially change your view on any of those megacaps. 0:09:28.920 --> 0:09:31.640 Allison, is there anything that excites you outside of the 0:09:31.760 --> 0:09:34.800 US listed megacat tech names anywhere else in the world. 0:09:36.080 --> 0:09:40.560 Yeah, we see lots of exciting new companies. Particularly we 0:09:40.640 --> 0:09:44.800 see some you know, exciting emerging technology leaders in India. 0:09:45.760 --> 0:09:49.560 We also see, you know, through the US market, there 0:09:49.600 --> 0:09:53.000 are several areas that are just emerging in terms of 0:09:53.080 --> 0:09:57.559 the productivity gains that they can be had from artificial 0:09:57.600 --> 0:10:00.880 intelligence that are being underappreciated even in some of the 0:10:01.080 --> 0:10:05.600 contract manufacturers, where we see trends to what's still towards 0:10:05.640 --> 0:10:10.160 de globalization, towards focus on you know, higher standards and 0:10:10.240 --> 0:10:13.240 supply chains, that some of the this AI spent is 0:10:13.640 --> 0:10:16.160 actually trickled down to some of the companies that will 0:10:16.200 --> 0:10:21.240 actually have to manufacture. And this not just TSMC for Nvidia, 0:10:21.360 --> 0:10:24.079 but the servers, the storage that we'll have that will 0:10:24.120 --> 0:10:27.720 be made for many of these hyper scalers. We think 0:10:28.000 --> 0:10:30.400 you'll start to see that accrue to some some different 0:10:30.440 --> 0:10:32.719 names much further down the market CAPCR. 0:10:33.440 --> 0:10:36.120 Alison, that's really interesting. But again we're sort of talking 0:10:36.160 --> 0:10:40.439 about the infrastructure layer. Where will the money go from 0:10:40.440 --> 0:10:42.640 the application layer? At some point do you. 0:10:42.640 --> 0:10:46.560 Think, well, it just takes time and we've seen that before, 0:10:46.600 --> 0:10:49.080 you know, when we had the mobile era, you know, 0:10:49.160 --> 0:10:52.080 between when Apple launched its iPhone and two thousand and 0:10:52.120 --> 0:10:56.080 eight and you actually had the IPO of Uber. You know, 0:10:56.160 --> 0:11:01.160 over ten years of developing the app store, develop applications, 0:11:01.200 --> 0:11:03.040 and you know, we do tend to see and this 0:11:03.120 --> 0:11:06.440 is followed a very similar pattern for AI. You tend 0:11:06.480 --> 0:11:11.079 to see advertising benefit first, and then we see e commerce. 0:11:11.120 --> 0:11:14.240 We're hearing that from Walmart, and then you start to 0:11:14.240 --> 0:11:19.120 see enterprise software, some productivity gains, and then eventually that 0:11:19.160 --> 0:11:22.520 filters into other new applications, many of which we won't 0:11:22.600 --> 0:11:26.320 have thought of first. But this happens over multiple years, 0:11:26.360 --> 0:11:30.800 you know, in technology proper waves rather than just type cycles. 0:11:30.880 --> 0:11:34.120 These take you know, five to seven years to play out. 0:11:34.240 --> 0:11:37.760 So investors shouldn't necessarily look and think that investments have 0:11:37.840 --> 0:11:40.880 to be in those applications today because many of those 0:11:40.920 --> 0:11:44.240 still have to emerge as we start to see you know, 0:11:44.320 --> 0:11:48.600 the interface for AI for consumers still involved somewhat over 0:11:48.640 --> 0:11:49.600 the next few years. 0:11:50.120 --> 0:11:52.320 Alison is always so great to have you on the show. 0:11:52.600 --> 0:11:55.280 Thank you for joining us, Happy weekend. Thank you in Edinburgh. 0:11:55.320 --> 0:11:59.120 Alison water portflio manager at Jannis Henderson Investors, Global Technology 0:11:59.200 --> 0:12:03.640 leaders Sstainable Future Technology funds any morehead stick with us 0:12:03.640 --> 0:12:04.720 the bluebot technology. 0:12:15.720 --> 0:12:17.880 Okay, it's time for talking tech and first up. Ali 0:12:17.920 --> 0:12:21.280 Barber shareholders have approved a plan to upgrade its Hong 0:12:21.400 --> 0:12:25.000 Kong listing to primary status, a move that's expected to 0:12:25.000 --> 0:12:28.480 attract billions of dollars of investment from mainland China. This 0:12:28.640 --> 0:12:31.920 is shares of the online retailer have trailed key rivals 0:12:32.240 --> 0:12:35.000 like ten Cent. The change goes into effect August twenty 0:12:35.040 --> 0:12:38.000 eighth plus Uber is set to start offering its self 0:12:38.080 --> 0:12:42.160 driving Cruise vehicles to customers beginning next year. The partnership 0:12:42.200 --> 0:12:45.200 comes as Cruise, which is owned by General Motors, has 0:12:45.240 --> 0:12:48.600 been trying to regain traction after losing ground and grounding 0:12:48.640 --> 0:12:52.280 its fleet last year following the mishandling of a collision 0:12:52.480 --> 0:12:55.800 with a pedestrian, and bank capitals forging ahead with revived 0:12:55.800 --> 0:12:59.800 plans for an IPO of chip maker Kayosia Holdings that 0:12:59.800 --> 0:13:02.840 can raise five hundred million. That's according to Bloomberg's sources. 0:13:02.920 --> 0:13:05.440 Bain said to be working with investment banks for the 0:13:05.440 --> 0:13:15.840 IPO kickoff that could begin in the coming weeks. 0:13:16.040 --> 0:13:22.480 I will make sure let me leave the world into 0:13:22.559 --> 0:13:23.200 the future. 0:13:23.520 --> 0:13:28.200 On space and artificial intelligence, that America, not China, wins 0:13:28.280 --> 0:13:32.720 the competition for the twenty first century, and that we strengthen, 0:13:33.080 --> 0:13:36.080 not abdicate our global leadership. 0:13:37.720 --> 0:13:41.640 Vice President Harris accepting the Democratic nomination for president last night. 0:13:41.880 --> 0:13:44.520 Bloomberg's Amory Hodden is on the ground in Chicago where 0:13:44.520 --> 0:13:47.560 the DNC is now wrapped up, and Amory, boy, it 0:13:47.600 --> 0:13:49.560 was an important speech. I have a feeling you might 0:13:49.600 --> 0:13:50.920 be a little tired after yesterday. 0:13:53.679 --> 0:13:57.360 Absolutely a very important speech and historic speech for Vice 0:13:57.400 --> 0:14:02.040 President Kamala Harris as she accepts the Democratic National Convention, 0:14:02.160 --> 0:14:05.640 she accepts the nomination to be the top of the ticket. 0:14:05.679 --> 0:14:08.800 We have to say, and cannot overstate this, what a 0:14:08.880 --> 0:14:11.280 one pint eighty we are standing here today than we 0:14:11.280 --> 0:14:14.760 were just a month ago. A tremendous amount of energy 0:14:14.840 --> 0:14:18.200 now has been infused into this campaign and now comes 0:14:18.240 --> 0:14:21.160 potentially the hard part. We heard from the Vice president 0:14:21.240 --> 0:14:24.600 last night. She talked broadly, a lot about issues, whether 0:14:24.640 --> 0:14:29.040 it was the economy, foreign policy, reproductive rights, really cherry 0:14:29.160 --> 0:14:32.320 picking part of the Biden plan that she was part of, 0:14:32.680 --> 0:14:34.720 and now she's going to potentially have to go into 0:14:34.760 --> 0:14:37.680 details of those policies. I spoke to Tobin Marcus this 0:14:37.720 --> 0:14:40.480 morning of Wolf Research, and he said that this campaign 0:14:40.560 --> 0:14:45.400 is using strategic ambiguity. What they want to basically connect 0:14:45.400 --> 0:14:49.040 with swing state voters, using populist rhetoric, but not having 0:14:49.080 --> 0:14:52.760 to go into the details of those policy proposals. They 0:14:52.760 --> 0:14:56.720 have about seventy four days now to get that job done. 0:14:57.040 --> 0:14:59.840 Will they continue to try to potentially put a little 0:14:59.840 --> 0:15:02.800 bit bit of a gap between them and the actual 0:15:02.840 --> 0:15:06.080 details of the policy and just continuously try to talk 0:15:06.200 --> 0:15:09.600 broadly about the issues. That's going to become much harder 0:15:09.800 --> 0:15:12.760 come September tenth, when you have Kamala Harris facing off 0:15:12.840 --> 0:15:14.479 Donald Trump in their first. 0:15:14.200 --> 0:15:19.760 Debate, Bloomberg's Emory horder and in Chicago, thank you very much, okay. 0:15:20.040 --> 0:15:23.480 Open Ai is opposing a state bill here in California 0:15:23.760 --> 0:15:27.000 that would place new safety requirements on AI companies, like 0:15:27.280 --> 0:15:31.320 a fail safe shutdown for AI systems or compliance disclosions. 0:15:31.400 --> 0:15:34.880 The startup joins a long list of academic, tech industry 0:15:34.960 --> 0:15:39.000 leaders and other politicians who argue the measures could stifle innovation. 0:15:39.160 --> 0:15:42.440 The author of bill SB one oh four seven, California 0:15:42.480 --> 0:15:45.280 State Senator Scott Wiener, joins us now in San Francisco, 0:15:45.280 --> 0:15:48.560 along with Bloomberg, Sharene Gafari, who actually broke the story 0:15:48.800 --> 0:15:51.360 that open AI had outlined its concerns in a letter 0:15:51.400 --> 0:15:55.000 to you State Senator Wiena. So let's start with this. 0:15:55.440 --> 0:15:58.520 Open ai is based in San Francisco, California. It is 0:15:58.640 --> 0:16:03.040 probably the world's leading AI name or startup, and they 0:16:03.040 --> 0:16:06.760 think that federal legislation is a better path than state. 0:16:07.120 --> 0:16:08.200 Your reaction to that. 0:16:08.320 --> 0:16:12.440 Well, I will start by saying that another AI lab 0:16:12.480 --> 0:16:16.120 in San Francisco, Anthropic, which currently has the largest model 0:16:16.480 --> 0:16:21.200 in the world, issued a letter yesterday stating that on 0:16:21.440 --> 0:16:24.920 balance SP ten forty seven is good and that it 0:16:25.040 --> 0:16:30.000 is implementable, and that Congress is unlikely to add you're 0:16:30.000 --> 0:16:32.840 talking about entropic out. I'm sorry that I'd say the name. 0:16:33.160 --> 0:16:36.920 Anthropic came out and said that I'm balanced, the bill 0:16:37.000 --> 0:16:39.480 is good and that it's implementable, and that Congress is 0:16:39.560 --> 0:16:42.360 unlikely to act. And so I have a lot of 0:16:42.440 --> 0:16:45.680 respect for open AI. But just to be clear, the 0:16:45.680 --> 0:16:49.680 open AI letter does not actually criticize any provision of 0:16:49.760 --> 0:16:52.440 the bill. It does not criticize anything in the bill. 0:16:52.480 --> 0:16:55.000 All it says is that this should be handled at 0:16:55.040 --> 0:16:57.920 the federal level. And to be clear, I agree it 0:16:57.960 --> 0:17:00.640 should be handled at the federal level. But here we 0:17:00.680 --> 0:17:03.720 are in twenty twenty four and Congress still has not 0:17:03.800 --> 0:17:07.600 passed a data privacy law, has done nothing around social 0:17:07.640 --> 0:17:12.120 media other than banning TikTok, has not codified net neutrality. 0:17:12.440 --> 0:17:16.320 Congress has a very poor record in terms of regulating 0:17:16.400 --> 0:17:18.919 the tech sector, and I don't see that changing, and 0:17:18.960 --> 0:17:21.080 so we California should lead. 0:17:22.320 --> 0:17:24.480 Do you think that we should trust There's been a 0:17:24.480 --> 0:17:29.280 long list of tech companies, vcs, academics who have criticized 0:17:29.280 --> 0:17:32.000 this bill. But do you think that we should be 0:17:32.040 --> 0:17:34.399 trusting some of these voices who are against it? Are 0:17:34.440 --> 0:17:37.560 there other motivations that you think could be at play here? 0:17:38.280 --> 0:17:41.600 So to be clear, there are folks who are criticizing 0:17:41.600 --> 0:17:44.600 the bill in good faith and people I respect the lot. 0:17:45.119 --> 0:17:48.720 There are others who are making very inaccurate and extreme 0:17:48.800 --> 0:17:52.520 and exaggerated attacks on this bill. This is a very 0:17:52.600 --> 0:17:58.800 light touch, smart piece of regulation around safety, regulating only 0:17:58.880 --> 0:18:03.639 the largest model to just ask these developers to perform 0:18:03.680 --> 0:18:06.719 safety testing so that we can reduce the risk of 0:18:06.800 --> 0:18:10.240 things like shutting down the electric grid, or attacking the 0:18:10.240 --> 0:18:14.040 banking system, or creating a chemical or biological weapon. And 0:18:14.119 --> 0:18:17.880 some of the leading minds in AI, like Yoshua Benzio 0:18:18.119 --> 0:18:22.879 and Jeff Hinton formerly from Google, Lawrence Lessig, and Stuart 0:18:22.920 --> 0:18:25.639 Russell are all supporting this bill, in addition to a 0:18:25.680 --> 0:18:28.600 significant list of startups who are saying that this is 0:18:28.640 --> 0:18:31.840 a good idea. So, yes, there are some people in 0:18:31.880 --> 0:18:34.679 good faith who have criticisms of the bill, and we 0:18:34.840 --> 0:18:37.640 have been working with them to get their feedback. There 0:18:37.640 --> 0:18:40.200 are others who are just opposed to any and all 0:18:40.280 --> 0:18:43.840 regulation of tech, and I just don't agree with that. 0:18:44.560 --> 0:18:46.960 And I want to talk about whistleblower protections, which is 0:18:47.000 --> 0:18:49.080 also a part of the bill, and it's received less attention. 0:18:49.920 --> 0:18:54.040 Today two former open ai employees came out and supported 0:18:54.080 --> 0:18:57.000 the bill, and they have advocated for more whistleblower protections. 0:18:57.200 --> 0:18:58.480 Why is that a part of the bill? Why is 0:18:58.520 --> 0:19:00.760 it important that people who work for these AI companies 0:19:00.800 --> 0:19:02.800 you think should be able to speak out on safety. 0:19:03.000 --> 0:19:06.080 So I am, just to be clear, a huge fan 0:19:06.320 --> 0:19:08.960 of AI. I think AI has so much potential to 0:19:09.040 --> 0:19:13.040 make the world a better place to solve what we 0:19:13.119 --> 0:19:16.560 have viewed as unsolvable problems, whether it's around various health 0:19:16.600 --> 0:19:20.600 conditions or climate change. And so I'm a fan but 0:19:20.720 --> 0:19:24.800 as with any extremely powerful technology, there are risks, and 0:19:24.840 --> 0:19:26.800 we should try to get ahead of those risks. And 0:19:26.840 --> 0:19:29.240 we know how to get ahead of those risks, not 0:19:29.320 --> 0:19:31.960 to eliminate risk, but to reduce the risk of something 0:19:31.960 --> 0:19:35.720 catastrophic happening. And so when you have people who are 0:19:35.720 --> 0:19:39.520 working in AI labs building some of the most powerful 0:19:39.560 --> 0:19:44.040 technology in the history of the world, they should be 0:19:44.119 --> 0:19:46.840 able to speak up and have legal protections to speak 0:19:46.920 --> 0:19:50.920 up if they see something inappropriate or dangerous happening. 0:19:52.040 --> 0:19:57.040 Are you yourself, Senator, can I ask nervous, even scared 0:19:57.080 --> 0:20:00.240 around of special intelligence having done this amount of research. 0:20:00.840 --> 0:20:04.520 I wouldn't. I wouldn't say scared, Like I am optimistic 0:20:05.040 --> 0:20:08.200 that AI can help us solve some of the hardest 0:20:08.240 --> 0:20:12.399 problems facing us. At the same time, it is it 0:20:12.440 --> 0:20:16.800 would be irresponsible to just allow it to spiral and 0:20:17.240 --> 0:20:21.240 not to in any way even think about the safety issues. 0:20:21.440 --> 0:20:25.280 And that's all this bill does. It requires safety protocols, 0:20:25.359 --> 0:20:29.280 a safety evaluation, the ability to shut down a model 0:20:29.560 --> 0:20:33.200 that is in your possession. These are all extremely reasonable things. 0:20:33.240 --> 0:20:36.119 And I want to be clear that every single large 0:20:36.160 --> 0:20:40.320 AI lab, including open AI and Meta and Google, they 0:20:40.359 --> 0:20:45.879 have already committed repeatedly and publicly to perform these safety evaluations. 0:20:45.880 --> 0:20:49.960 We're just asking them to do it in an obligatory way, 0:20:50.280 --> 0:20:52.680 and it should cause us all concern. Whenever you have 0:20:52.760 --> 0:20:56.240 an industry that says, we promise we'll do this, but 0:20:56.359 --> 0:20:59.600 don't require us to do it, just trust us. Voluntary 0:20:59.640 --> 0:21:03.040 commit ments in industry have a history of going awry. 0:21:03.600 --> 0:21:06.240 Senator, you have listened and listened hard to some of 0:21:06.240 --> 0:21:09.359 the concerns and made changes. What seems to be the 0:21:09.400 --> 0:21:13.399 ongoing narrative from academia and indeed smaller companies is this 0:21:13.520 --> 0:21:17.159 open source narrative and the way in which perhaps LAMA 0:21:17.520 --> 0:21:21.919 would in any way suddenly become less bit available and 0:21:22.000 --> 0:21:26.640 less innovative. How can you change to enable smaller companies 0:21:26.680 --> 0:21:29.080 an open source still to thrive under this regulation. 0:21:30.040 --> 0:21:33.439 Well, I'm a huge fan of open source, including in 0:21:33.480 --> 0:21:36.720 the AI space, and so many startups in particular. And 0:21:36.720 --> 0:21:39.040 by the way, startups are not covered by this bill. 0:21:39.320 --> 0:21:41.800 It applies only to labs that are spending more than 0:21:41.800 --> 0:21:45.800 one hundred million dollars to train their models. But AI, 0:21:46.040 --> 0:21:49.600 excuse me, open source AI plays an incredibly important role 0:21:49.840 --> 0:21:53.520 and we have worked intensively with GitHub and other open 0:21:53.520 --> 0:21:57.679 source advocates and made significant changes to the bill to 0:21:57.920 --> 0:22:00.800 make very very clear that this bill is not about 0:22:00.800 --> 0:22:04.520 being negative with open source. And yes, a lot of 0:22:04.560 --> 0:22:08.280 people rely on LAMA. But I know I sound like 0:22:08.280 --> 0:22:10.919 a broken record, but I'll say it again. META is 0:22:10.960 --> 0:22:14.760