WEBVTT - Anthropic Unveils Updated Opus 4.7 Model 0:00:02.520 --> 0:00:13.400 Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is live 0:00:13.440 --> 0:00:17.239 from coast to coast with Caroline Hide in New York 0:00:17.520 --> 0:00:20.640 and Eva Low in San Francisco. 0:00:22.000 --> 0:00:25.599 This is Bloomberg Tech Coming Up. Anthropic introduces an updated 0:00:25.680 --> 0:00:29.880 version of its widely available AI model Opus four point seven, 0:00:30.000 --> 0:00:33.559 just a week after its limited release of Mythos fus. 0:00:33.400 --> 0:00:36.800 Eli Muscus, kicking his terrifab plan off into high gear 0:00:37.040 --> 0:00:40.080 even if skepticism grows from the semiconductor industry. 0:00:39.800 --> 0:00:43.080 And TSMC reports a big surge in profit and raises 0:00:43.120 --> 0:00:46.240 its revenue outlook for twenty twenty six, driven by strong 0:00:46.280 --> 0:00:47.600 demand for AI chips. 0:00:47.880 --> 0:00:50.840 There is so much AI news out today, but there's 0:00:50.840 --> 0:00:54.760 still the geopolitical global context that investors are currently trying 0:00:54.800 --> 0:00:57.279 to navigate, and it's a tough one. Stocks as we 0:00:57.360 --> 0:01:00.680 speak at new record highs there we are come closer. 0:01:00.720 --> 0:01:02.520 We're up almost two tens percent and as that one 0:01:02.600 --> 0:01:04.959 hundred and therefore up for a twelfth straight day, the 0:01:05.040 --> 0:01:08.920 longest winning streak since twenty seventeen. But longer term peace 0:01:08.959 --> 0:01:11.920 process is needed. In the Middle East. All remains elevated 0:01:12.000 --> 0:01:14.959 concerns about the Hormons straight, of course, but at the moment, 0:01:15.240 --> 0:01:17.160 tech has been leading the charge, and it continues to 0:01:17.520 --> 0:01:19.280 as we see it on the day higher and then 0:01:19.319 --> 0:01:21.559 it's looking at it on the twelve days higher, we're 0:01:21.600 --> 0:01:25.760 now seeing that extraordinary run fourteen percent higher. Ed more broadly, 0:01:25.800 --> 0:01:29.200 but we've got to really talking as well about Anthropic introducing 0:01:29.240 --> 0:01:33.480 an updated version of its widely available AI model Opus, 0:01:33.520 --> 0:01:37.039 of course, barely a week after its limited release of 0:01:37.080 --> 0:01:41.600 a more advanced offering called Mythos. Now, Bluemberg's Rachel Metz 0:01:41.680 --> 0:01:43.280 joins us for a little bit more, and we want 0:01:43.280 --> 0:01:45.199 to get into a broader market context in a minute. 0:01:45.400 --> 0:01:48.360 But Rachel, for now, there hasn't been much market reaction 0:01:48.640 --> 0:01:51.280 to Opus. But what are you seeing at this current 0:01:51.320 --> 0:01:53.760 moment of really how it's improved four point seven versus 0:01:53.800 --> 0:01:54.440 four point six? 0:01:55.360 --> 0:01:58.840 Sure, I mean it sounds like it's improved at a 0:01:59.200 --> 0:02:01.000 broad range of keep abilities, at least that's what the 0:02:01.000 --> 0:02:07.040 Anthropic has said so far, Particularly at coding and taking 0:02:07.400 --> 0:02:10.320 complicated tasks that people might in the past cut up 0:02:10.360 --> 0:02:13.519 into a few pieces. You might be able to give 0:02:13.560 --> 0:02:16.840 it an entire more complicated task, especially as pertains to 0:02:16.880 --> 0:02:20.880 software engineering. It also sounds like it is better at 0:02:21.320 --> 0:02:24.320 computer vision, which is the task where a computer is 0:02:24.320 --> 0:02:26.880 trying to figure out what is in a picture. It 0:02:26.919 --> 0:02:30.240 can take on much higher resolution images, they say, and 0:02:30.400 --> 0:02:35.280 it can also better extract information from these images, which 0:02:35.680 --> 0:02:38.960 could lead to a whole range of different kinds of applications, 0:02:38.960 --> 0:02:42.160 like analyzing different kinds of data or complicated data, that 0:02:42.200 --> 0:02:42.680 sort of thing. 0:02:42.760 --> 0:02:44.800 I think the point that in Forbig would make is 0:02:44.840 --> 0:02:48.160 that out in the real world everything is not plain text, 0:02:48.600 --> 0:02:53.560 so that vision component was really important. Generally speaking, they 0:02:53.680 --> 0:02:57.639 are at pains to make distinctions between Opus four point 0:02:57.720 --> 0:03:01.959 seven being an updated, improved, existing model from four point 0:03:02.080 --> 0:03:05.200 six and how it's different from Mythos, right. I think 0:03:05.240 --> 0:03:08.520 they even talk about in the blog about how they 0:03:09.080 --> 0:03:13.919 experimented with ways to quote differentially reduced four point seven's 0:03:14.000 --> 0:03:17.000 cyber capabilities. Could you just explain a little bit about 0:03:17.000 --> 0:03:21.320 the relationship between four point seven and Mythos that we've 0:03:21.360 --> 0:03:22.880 all been talking about for like a week now. 0:03:24.160 --> 0:03:24.440 Sure. 0:03:24.560 --> 0:03:28.519 So, Mythos is a model that the company has said 0:03:28.639 --> 0:03:31.440 is so powerful that it doesn't want to generally release it, 0:03:31.520 --> 0:03:35.000 and it is handing it over to a number of 0:03:35.240 --> 0:03:38.920 companies who it wants to be like using it in 0:03:38.920 --> 0:03:41.560 the earlier stages. This is a model that's meant especially 0:03:41.600 --> 0:03:46.160 for spotting cybersecurity vulnerability so that people can then patch them. 0:03:46.600 --> 0:03:48.640 But of course it could also be used for the opposite, 0:03:49.600 --> 0:03:51.400 since if it is as good as they say it 0:03:51.400 --> 0:03:54.760 is at finding those vulnerabilities, So this model is not 0:03:54.840 --> 0:03:57.240 going to do all of those things. It is purposely 0:03:57.280 --> 0:03:59.480 not doing all of those things. For that reason, some 0:03:59.520 --> 0:04:02.360 people not be happy with it, but the company says 0:04:02.400 --> 0:04:04.640 that it is also better at a wide range of 0:04:04.680 --> 0:04:08.920 things than its previous Opus model. So we'll see how 0:04:09.000 --> 0:04:10.560 it is used. For a lot of people. It will 0:04:10.600 --> 0:04:14.560 probably be considered an improvement step up from the past one. 0:04:14.200 --> 0:04:17.480 Maybe not the step change that everyone feels that mythos 0:04:17.520 --> 0:04:19.240 is Rachel, but just give us the context. This is 0:04:19.240 --> 0:04:22.200 a company that we understand. According to Bloomberg reporting, it's 0:04:22.240 --> 0:04:25.599 pushing back on an extraordinary valuation yet more funding because 0:04:25.640 --> 0:04:27.960 the venture community wants to get in on it so much, 0:04:28.120 --> 0:04:30.200 and we are seeing this global roll out a little 0:04:30.240 --> 0:04:33.400 bit more. Aphobic's just so hot right now. It's four 0:04:33.400 --> 0:04:36.720 point seven going to be enough to maintain that momentum. 0:04:37.279 --> 0:04:39.440 Oh, that's a good question. I guess We're going to 0:04:39.520 --> 0:04:42.760 see how people are using it. I mean, I think 0:04:42.800 --> 0:04:45.479 the company is really on a tier right now due 0:04:45.480 --> 0:04:49.120 to a different bunch of different factors. And some of 0:04:49.160 --> 0:04:53.039 that is its technology and the capabilities of that technology. 0:04:53.720 --> 0:04:57.720 Some of it I think is vibes right, but we'll see. 0:04:57.760 --> 0:05:01.560 I mean this is like as we no in AI, 0:05:01.680 --> 0:05:04.360 as in any other market, there isn't usually like one winner, 0:05:05.120 --> 0:05:08.360 So I think that it's still anybody's game, and there 0:05:08.360 --> 0:05:10.560 will probably be a bunch of winners in the end. 0:05:11.640 --> 0:05:13.919 We're showing this chart, which is on the Bloomberg terminal. 0:05:13.960 --> 0:05:15.640 It's data that I think we keep our on. It's 0:05:15.640 --> 0:05:19.200 actually coming from Ramp shows open Ai ahead in terms 0:05:19.200 --> 0:05:19.760 of adoption. 0:05:19.880 --> 0:05:21.360 It's a sort of data set. 0:05:21.480 --> 0:05:23.360 But we remember that in the context of opu's four 0:05:23.400 --> 0:05:26.920 point seven caro and with Methos, the story of Anthropic 0:05:26.960 --> 0:05:29.960 from day one was enterprise. That is a more recent 0:05:30.000 --> 0:05:32.800 pivot for open Ai. Bloomberg's Rachel Metz, you have done 0:05:32.880 --> 0:05:36.120 terrific reporting on the inner workings of Anthropic, what's going 0:05:36.120 --> 0:05:36.840 on with these models. 0:05:37.040 --> 0:05:38.000 Really appreciate it. 0:05:38.120 --> 0:05:41.680 Another big story from Bloomberg, Elon Musk is racing ahead 0:05:41.680 --> 0:05:44.480 with his chip making terra fab project. Sources say that 0:05:44.520 --> 0:05:48.080 billionaires the tenants have already reached out to industry supplies, 0:05:48.120 --> 0:05:52.360 including Tokyo Electron, Applied Materials and LAMB, and they're seeking 0:05:52.400 --> 0:05:56.320 price quotes delivery times for an array of chip making gear. 0:05:56.400 --> 0:05:59.000 Let's get out to Bloomberg's Global Tech editor Peter Elstrom. 0:05:59.080 --> 0:06:01.599 This was a big team effort around the world to 0:06:01.640 --> 0:06:04.719 try and work out how real is this with terror 0:06:04.800 --> 0:06:07.240 Fab and we got a lot of detail Peter on 0:06:07.320 --> 0:06:10.359 where they're at and the pressure that Elon Inc. Is 0:06:10.400 --> 0:06:12.919 putting on the supply chain to get moving, give it 0:06:12.960 --> 0:06:14.000 what it wants. What do we need to know? 0:06:15.360 --> 0:06:17.080 Yeah, just to set the stage here. 0:06:17.160 --> 0:06:19.760 When Elon Must first laid out this vision of a 0:06:19.920 --> 0:06:24.120 terrafab a few weeks ago, nobody really knew what to 0:06:24.200 --> 0:06:26.599 make of it. He was talking about this getting into 0:06:26.640 --> 0:06:29.400 chip manufacturing for the very first time, something he's never 0:06:29.440 --> 0:06:32.400 done before. It's a very competitive market, a very difficult market. 0:06:32.560 --> 0:06:35.479 He was talking about building the equivalent of hundreds of 0:06:35.560 --> 0:06:37.360 new fabs that he was going to roll out, and 0:06:37.360 --> 0:06:40.800 he was going to invest. One analyst estimated that it 0:06:40.800 --> 0:06:44.000 would cost something like five trillion to thirteen trillion dollars 0:06:44.040 --> 0:06:45.799 to be able to do what he was actually talking 0:06:45.800 --> 0:06:47.880 about so nobody knew how seriously. 0:06:47.600 --> 0:06:48.000 To take him. 0:06:48.000 --> 0:06:49.960 There were a couple of analysts who who put it 0:06:50.000 --> 0:06:51.960 from the beginning and said, he's not really going to 0:06:52.320 --> 0:06:55.920 go ahead and manufacture chips or even try to manufacture chips. 0:06:56.040 --> 0:06:58.280 He's just trying to get the existing companies to move 0:06:58.320 --> 0:07:00.960 ahead more quickly. But I want to turn the tables 0:07:01.000 --> 0:07:03.000 around a little bit here, Ed, because you were part 0:07:03.040 --> 0:07:05.120 of the people on the byline in this story. It 0:07:05.160 --> 0:07:07.120 sounds like now we need to take them a little 0:07:07.120 --> 0:07:09.400 bit more seriously. What do you make of this? How 0:07:09.440 --> 0:07:12.040 seriously do you think we should take this effort now? 0:07:12.120 --> 0:07:14.120 Well, what I would say is the two data points 0:07:14.120 --> 0:07:16.720 that we had in the story. Three thousand wafers per 0:07:16.760 --> 0:07:20.600 month initially is nothing in the context of current capacity 0:07:20.640 --> 0:07:23.600 for lead edge fab and then what we understand from 0:07:23.600 --> 0:07:26.360 sources carrot, by the way, twenty twenty nine is when 0:07:26.360 --> 0:07:28.440 they start. You know, I don't know how far ahead 0:07:28.440 --> 0:07:30.160 you're looking at your calendar, but I don't even know 0:07:30.160 --> 0:07:33.360 what's happening next week. Twenty twenty nine is quite far away. 0:07:33.400 --> 0:07:36.320 It is. And look, we had Berenberg analyst on yesterday 0:07:36.360 --> 0:07:38.320 who I love that you quoted and for the video 0:07:38.360 --> 0:07:40.840 in the story of really saying like they're not factoring 0:07:40.840 --> 0:07:42.840 in these numbers yet. It's get it two three years 0:07:42.880 --> 0:07:46.080 really for them to understand when ASML gets involved in 0:07:46.200 --> 0:07:48.640 giving the equipment that's going to be necessary to make 0:07:48.680 --> 0:07:52.240 the chips. More broadly, Peter, but just talk about the 0:07:52.280 --> 0:07:55.880 reticence that the rest of the industry has towards this. 0:07:56.040 --> 0:07:58.240 I think one line that really caught my attention at 0:07:58.240 --> 0:08:01.560 court Ed's I know was the SAMSI like, no, thanks, 0:08:01.760 --> 0:08:03.360 we don't particularly want to help in this in the 0:08:03.360 --> 0:08:04.600 way that perhaps you'd like us to. 0:08:05.440 --> 0:08:07.440 Yeah, it's important to say that the leader in this 0:08:07.480 --> 0:08:10.320 market right now, the foundry business is TSMC. They're far 0:08:10.360 --> 0:08:12.600 ahead of everybody. We've just got their earnings today. We 0:08:12.640 --> 0:08:15.000 can talk about those a bit too. But Intel and 0:08:15.040 --> 0:08:17.520 Samsung are both trying to get into this foundry business 0:08:17.520 --> 0:08:20.160 because it has been so good for TSMC. They know 0:08:20.200 --> 0:08:22.800 that that business is growing. So if you have somebody 0:08:22.880 --> 0:08:26.480 like Samsung that's already expert in building these kinds of chips, 0:08:26.600 --> 0:08:28.520 they're going to be kind of reluctant to trade some 0:08:28.560 --> 0:08:31.160 of that expertise to somebody like Elon Musk who has 0:08:31.400 --> 0:08:34.480 totally no track record in this area whatsoever. The same 0:08:34.520 --> 0:08:38.400 applies to the chip equipment maker, so Tokyo Electron Applied 0:08:38.440 --> 0:08:41.560 Materials ASML. They need to decide who are we going 0:08:41.600 --> 0:08:43.439 to sell our equipment to. We're in this market right 0:08:43.480 --> 0:08:45.760 now where there's actually a limit to the amount that 0:08:45.800 --> 0:08:49.640 they can provide, and they're limited more by supply than 0:08:49.679 --> 0:08:51.599 actually demand at this point. So are they going to 0:08:51.640 --> 0:08:54.360 sell it to Elon Musk who's never done this before 0:08:54.400 --> 0:08:56.360 for the terror FIB project, or they're going to sell 0:08:56.400 --> 0:08:59.960 it to TSMC and Samsung and Intel who are pretty experienced. 0:09:00.040 --> 0:09:00.160 Yea. 0:09:00.840 --> 0:09:03.079 And when we broke the story in Asia, those names 0:09:03.080 --> 0:09:06.360 that you mentioned jumped pretty significantly in pre market. Some 0:09:06.400 --> 0:09:09.560 of the US suppliers also did. Then fell away on 0:09:09.720 --> 0:09:13.520 Samsung they basically counted his my understanding saying we're going 0:09:13.559 --> 0:09:16.160 to ramp up in Texas. But the story in the market, 0:09:16.160 --> 0:09:19.320 it's one of he is TSMC. Right on the call TSMCEO, 0:09:19.480 --> 0:09:22.280 it was like, this is the reality of building a fab. 0:09:22.360 --> 0:09:24.520 Three years to build it, two years to ramp it. 0:09:25.000 --> 0:09:28.240 But things for TSMC in particular are going well, Peter. 0:09:28.320 --> 0:09:30.120 The data tells us what you said. They are the 0:09:30.240 --> 0:09:32.280 leader and demand is accelerating. 0:09:33.440 --> 0:09:35.920 Yeah, it was interesting to see cc way that CEO 0:09:36.320 --> 0:09:38.280 al was asked about terror fab and he said, there 0:09:38.280 --> 0:09:41.400 are no shortcuts. You can't do this faster than what 0:09:41.480 --> 0:09:43.600 the leaders in the industry have been able to. 0:09:43.559 --> 0:09:44.120 Do so far. 0:09:44.200 --> 0:09:47.560 But what TSMC's financial results show us is demand is 0:09:47.679 --> 0:09:50.200 very very strong. Their profit was up fifty eight percent, 0:09:50.400 --> 0:09:53.320 that was well above analyst estimates. They said they boosted 0:09:53.320 --> 0:09:56.200 the revenue growth forecast for the full year to more 0:09:56.240 --> 0:09:59.120 than thirty percent growth before they had had it below 0:09:59.240 --> 0:10:01.839 thirty percent. They also said that they're going to ramp 0:10:01.920 --> 0:10:03.520 up their capital spending. They're going to be at the 0:10:03.520 --> 0:10:05.720 top end of this range where they approach fifty six 0:10:05.760 --> 0:10:08.600 billion dollars, so they're moving very very quickly. They say 0:10:08.600 --> 0:10:11.439 the demand for AI is very very strong at this point. 0:10:11.600 --> 0:10:13.520 There were a couple of notes of caution in there. 0:10:13.559 --> 0:10:16.160 Of course, the Iran war. Nobody knows exactly what that's 0:10:16.200 --> 0:10:19.160 going to do to the global economy to trade. There's 0:10:19.200 --> 0:10:21.840 another issue that's kind of critical for the tech industry 0:10:21.920 --> 0:10:23.760 right now that we've talked about a couple of times, 0:10:24.000 --> 0:10:27.120 the crunch and memory chips, memory ships of course going 0:10:27.120 --> 0:10:30.120 to all sorts of things because the AI hyperscalers are 0:10:30.120 --> 0:10:32.440 building out so quickly right now, they're taking up a 0:10:32.480 --> 0:10:36.240 bunch of that capacity. That's limited the ability of smartphonemakers 0:10:36.240 --> 0:10:38.679 to be able to get supplied. That could taper off 0:10:38.679 --> 0:10:39.760 some of the demand in the future. 0:10:40.760 --> 0:10:43.520 Most ptails from who's been a busy man, editing the 0:10:43.559 --> 0:10:45.400 great story from Ed and the rest of the team, 0:10:45.760 --> 0:10:47.720 and keeping an eye on all things Asia. Let's keep 0:10:47.760 --> 0:10:49.880 an eye on all things broader markets, global investors that 0:10:49.960 --> 0:10:52.560 continuing to navigate precarious environment. We keep talking about it, 0:10:52.559 --> 0:10:55.840 but US stocks have recovered to new record highs has 0:10:55.840 --> 0:10:58.280 been led by the technology sector in the studio with 0:10:58.360 --> 0:11:00.920 us now. Scott Ladner is chief investment officer at Horizon. 0:11:01.200 --> 0:11:03.199 So do you buy in to check Chips even more 0:11:03.200 --> 0:11:05.880 in AMD than Intel A leading on the points perspective, 0:11:05.880 --> 0:11:08.120 on the dage you continue on the hardware AI theme. 0:11:08.520 --> 0:11:10.240 Absolutely, I mean, I think it's really the thing that's 0:11:10.320 --> 0:11:12.320 that's it was driving the market as we came into 0:11:12.360 --> 0:11:13.720 this conflict, and it's going to drive the market as 0:11:13.720 --> 0:11:15.319 we come out of this conflict. And I think the 0:11:15.360 --> 0:11:17.600 reason why we are seeing markets at all time highs 0:11:17.840 --> 0:11:19.320 is because you know, the markets side of this thing 0:11:19.400 --> 0:11:21.600 is going to be short. Were resuring in weeks not months, uh, 0:11:21.720 --> 0:11:23.760 and the tailwinds we had coming into it are just 0:11:23.760 --> 0:11:25.800 going to continue to kind of wash over the markets. 0:11:25.800 --> 0:11:27.240 And that's that's why we said where we are right. 0:11:27.120 --> 0:11:30.160 Now Scott's could be with you in person. I enjoyed 0:11:30.200 --> 0:11:32.760 our conversations of late. So if you look at the 0:11:32.800 --> 0:11:34.880 NATS like one hundred, just as there's the case study 0:11:34.880 --> 0:11:37.360 it's up for a twelve straight day in spite of 0:11:37.400 --> 0:11:38.160 the war in Iran. 0:11:38.600 --> 0:11:39.360 So what is it. 0:11:39.400 --> 0:11:42.800 What's the data point or the thesis that's driving the 0:11:42.840 --> 0:11:45.559 technology sector hire if people are willing to look past 0:11:45.800 --> 0:11:49.120 a major conflict and a huge impact in energy markets. 0:11:49.240 --> 0:11:50.319 I mean, I think it's a couple of things that 0:11:50.480 --> 0:11:52.480 it's It's one, you know, the fact that this is 0:11:52.720 --> 0:11:54.760 sort of the next big thing, And when people talk 0:11:54.760 --> 0:11:56.679 about this as being like the next Internet, I think 0:11:56.679 --> 0:11:58.600 that's wrong. I mean, I think we're talking about this 0:11:58.600 --> 0:12:00.640 in terms of like in terms of scale and importance. 0:12:00.800 --> 0:12:02.400 We're talking about like the electrification of the. 0:12:02.400 --> 0:12:05.320 Country, like you know, the widespread availability like electricity al 0:12:05.360 --> 0:12:07.160 it's that scale of a productivity enhancement that I think 0:12:07.160 --> 0:12:09.360 we're talking about with AI. So so I just think 0:12:09.360 --> 0:12:12.200 people are actually still under sort of appreciating how important 0:12:12.200 --> 0:12:13.720 this thing's going to be. And then as well as 0:12:13.720 --> 0:12:15.240 the market kind of looks forward and kind of sort 0:12:15.320 --> 0:12:17.400 sort of figure out a price AI, like like you know, 0:12:17.400 --> 0:12:18.600 how do you how do you price the AI ins 0:12:18.640 --> 0:12:19.679 of a company? I think that's the thing that we're 0:12:19.679 --> 0:12:21.280 going to figure out next. It's gonna be through the 0:12:21.280 --> 0:12:23.640 margin channel, you know, like you know, ultimately like productivity, 0:12:23.679 --> 0:12:25.280 you know, it should should increase margins And that's and 0:12:25.280 --> 0:12:26.760 that is that is going to be where the rubber is. 0:12:26.760 --> 0:12:30.280 There's no point in exaggerating or hyperbole, but the NASAQ 0:12:30.360 --> 0:12:32.600 one hundred is up for twelve days. It hasn't done 0:12:32.600 --> 0:12:35.960 that since twenty seventeen. So just something clearly is happening underneath. 0:12:36.400 --> 0:12:40.920 Now we have earning season. How much could that derail that? 0:12:41.280 --> 0:12:43.240 Well, I mean, look, it could derail it, or it 0:12:43.240 --> 0:12:45.880 could sort of verify like why we are up for 0:12:45.880 --> 0:12:48.079 twelve day straight. You know, we're not up. We're not 0:12:48.200 --> 0:12:49.880 up twelve day straight in a vacuum though. I mean 0:12:49.880 --> 0:12:51.800 we're up twelve days straight coming off of some pretty 0:12:51.840 --> 0:12:54.000 nasty declines, like as a result of the war point, 0:12:54.520 --> 0:12:56.040 so you know that you know the context for that 0:12:56.080 --> 0:12:58.560 for that stuff matters. But you know, we actually think 0:12:58.600 --> 0:13:01.160 that the bar is is beautable for these tech companies. 0:13:01.360 --> 0:13:02.240 As as we come into this. 0:13:02.280 --> 0:13:04.640 Earning season, I mean, I think people are trying to 0:13:04.760 --> 0:13:07.319 understand like what is the next leg, Like what what's 0:13:07.320 --> 0:13:09.760 the next stage? And I do think like as folks 0:13:09.880 --> 0:13:11.920 start to figure out like the margin channel is really 0:13:11.920 --> 0:13:14.079 a thing that matters, that's going to make AI AAI 0:13:14.160 --> 0:13:15.480 even more important. 0:13:15.360 --> 0:13:16.560 To the business processes. 0:13:16.559 --> 0:13:18.000 And I think it's going to just like drive and 0:13:18.440 --> 0:13:20.800 just drive attention into the sector even higher. 0:13:20.840 --> 0:13:23.360 And I hate to say it, but margin often equates 0:13:23.400 --> 0:13:27.680 to fewer people. What's interesting is we're looking at Opus 0:13:27.679 --> 0:13:30.400 four point seven just getting announced in it it says 0:13:30.440 --> 0:13:33.560 as well, instead of the art score and finance agent evaluation, 0:13:33.840 --> 0:13:36.000 our internal testing showed Opus four point seven to be 0:13:36.000 --> 0:13:39.960 more effective finance analyst than Opus four point six. Now, 0:13:40.720 --> 0:13:42.840 how does that mean you feel about your role, about 0:13:42.880 --> 0:13:44.719 the people that you're hiring, and more broadly, where you 0:13:44.760 --> 0:13:46.600 want to invest in how it's going to help finance. 0:13:47.160 --> 0:13:50.240 I mean, frankly, it makes me kind of excited because 0:13:50.440 --> 0:13:52.760 the thing that we've seen out of AI, and so 0:13:52.880 --> 0:13:54.720 you know, we have some young people are at our company, 0:13:55.160 --> 0:13:57.720 and the thing that has been the most impressive to 0:13:57.760 --> 0:14:00.240 me is how quickly they are adopting sort of figuring 0:14:00.240 --> 0:14:01.679 out and getting up the learning curve. Frankly, I mean 0:14:01.720 --> 0:14:03.120 it's it's really tough for a twenty three year old 0:14:03.160 --> 0:14:04.920 to be very useful in this industry because it just 0:14:05.280 --> 0:14:06.200 you haven't seen very much. 0:14:06.440 --> 0:14:08.479 And they're using these tools to become. 0:14:08.480 --> 0:14:10.400 Much smarter, much faster, to be able to try more 0:14:10.440 --> 0:14:12.319 things and do a lot more of experimentation. And that's 0:14:12.360 --> 0:14:14.000 how you actually build knowledge and get to be a 0:14:14.040 --> 0:14:17.120 real good finance pro. And so I'm actually I'm actually 0:14:17.160 --> 0:14:19.440 kind of buffeted by by by by the use cases 0:14:19.440 --> 0:14:20.800 of this stuff. And it's not going to change the 0:14:20.800 --> 0:14:22.080 way we hired. It's going to change the type of 0:14:22.080 --> 0:14:24.400 people that we hired and how we train them. But 0:14:24.560 --> 0:14:26.160 in terms of you still need bodies to figure some 0:14:26.160 --> 0:14:26.800 of this stuff out. 0:14:27.000 --> 0:14:29.400 Scott briefly, should we buy software at this point? 0:14:31.440 --> 0:14:31.680 Yeah? 0:14:31.800 --> 0:14:35.280 Probably, but you have to be differentiated. I think I 0:14:35.320 --> 0:14:36.640 don't think you want to buy the ETF. You probly 0:14:36.680 --> 0:14:37.720 want to buy individual stocks. 0:14:38.480 --> 0:14:39.200 I was scrambling. 0:14:39.200 --> 0:14:41.320 I've got the I've got the benchmarks here on Opus 0:14:41.360 --> 0:14:42.840 four point seven, and I was waiting to see if 0:14:42.840 --> 0:14:45.720 Scott said he was worried about his job. Scott aland 0:14:45.760 --> 0:14:48.040 a chief investment officer at Horizon always great value on 0:14:48.080 --> 0:14:51.080 Bloomberg Tech, Thank you very much. A federal jury has 0:14:51.160 --> 0:14:55.600 ruled that Live Nation illegally monopolized the live events industry 0:14:55.920 --> 0:14:59.240 and overcharge fans for tickets to music performances. The landmark 0:14:59.320 --> 0:15:03.040 antitrust trial follows years of regulatory scrutiny of Live nations 0:15:03.040 --> 0:15:05.840 dominance in the live events business, and it now sets 0:15:05.840 --> 0:15:08.640 the stage for a possible breakup of the largest US 0:15:08.720 --> 0:15:11.760 concert promoter and ticket seller. Live Nation says it will 0:15:11.760 --> 0:15:12.960 appeal the verdict. 0:15:13.160 --> 0:15:16.560 Carroct now coming up ed. UK financial institutions will soon 0:15:16.600 --> 0:15:20.640 be getting access to anthropic mythos as concerns about cyberrists 0:15:20.680 --> 0:15:23.360 from the global regulators and enterprise continues to grow. That 0:15:23.440 --> 0:15:24.760 next this of bluebg. 0:15:24.400 --> 0:15:39.720 Tech, UK institutions will. 0:15:39.600 --> 0:15:42.840 Soon get access to anthropics powerful new model mythos on 0:15:42.960 --> 0:15:46.080 preview now. The company's head of EMEA North Pitt White, 0:15:46.120 --> 0:15:48.080 spoke the bluemost Tom McKenzie earlier taking. 0:15:47.920 --> 0:15:53.000 Listen mythos found vulnerabilities in every single operating system and 0:15:53.480 --> 0:15:56.200 every single web browser, and so we wanted to build 0:15:56.200 --> 0:15:59.360 this small cohort first to basically learn from what they 0:15:59.440 --> 0:16:02.240 found in the contained in controlled manner. But you're right, 0:16:02.320 --> 0:16:05.360 our intention is definitely to expand this out in a 0:16:05.480 --> 0:16:06.520 very intentional way. 0:16:06.560 --> 0:16:09.680 Typically it hasn't been expanded to UK financial institutions yet. 0:16:09.560 --> 0:16:11.800 No, but I think that is in the very near term, 0:16:12.160 --> 0:16:14.200 like in the next week's or way in the next week, 0:16:14.520 --> 0:16:16.160 in the next next week, okay. 0:16:15.920 --> 0:16:17.600 And the top UK financial institution. 0:16:17.680 --> 0:16:19.560 Yeah, we've been I mean, as you would expect, the 0:16:19.920 --> 0:16:22.840 engagement that I've had from CEOs in the last week 0:16:23.560 --> 0:16:26.680 in the UK has been significant. Customers firstly want to 0:16:26.720 --> 0:16:31.160 say that they appreciate our very thoughtful approach to the 0:16:31.200 --> 0:16:35.480 way in which we have released mythos, but secondly, of 0:16:35.480 --> 0:16:38.680 course they want to understand the opportunity to be considered 0:16:39.320 --> 0:16:41.760 to gain access to that model in a controlled way, 0:16:42.120 --> 0:16:44.280 and so we've been having a lot of conversations with 0:16:44.360 --> 0:16:47.200 CEOs about the reason why firstly we did project Project 0:16:47.240 --> 0:16:50.160 glass Wing and the way that we thought about that intentionally, 0:16:50.600 --> 0:16:54.440 and then secondly being very clear and to understand exactly 0:16:54.880 --> 0:16:57.280 why they want to get access to mythos and then 0:16:57.360 --> 0:17:00.280 working in partnership with those customers to think of the 0:17:00.280 --> 0:17:02.760 next kind of phase two and three and four. 0:17:03.400 --> 0:17:07.680 Does a system of release from Frontier Labs need to change? 0:17:07.840 --> 0:17:10.800 Some will be looking at this and thinking Anthropic has 0:17:10.800 --> 0:17:13.720 made this decision to limit the release of a model 0:17:13.760 --> 0:17:17.240 that can prove very dangerous. There's the upside, but there's 0:17:17.280 --> 0:17:19.760 also clearly the dangers that you've articulated and that the 0:17:19.760 --> 0:17:22.320 team have articulated. It's come down to decision making it 0:17:22.359 --> 0:17:24.360 Anthropic as to whether or not how to release this 0:17:24.880 --> 0:17:26.199 That has to change, doesn't it. 0:17:26.920 --> 0:17:30.680 I think where I would position it is that this model, 0:17:30.680 --> 0:17:33.760 in particular, we saw a significant leap in cyber and 0:17:33.960 --> 0:17:36.040 as a result of that, we were very concerned about 0:17:36.040 --> 0:17:38.600 a mass release and made the decision not to do that. 0:17:39.119 --> 0:17:42.040 And obviously I've already articulated our kind of approach with 0:17:42.400 --> 0:17:45.199 Project Last Wing. I think labs have to take a 0:17:45.240 --> 0:17:50.080 responsibility for thinking about safeguards, for thinking about safe and 0:17:50.160 --> 0:17:53.080 responsible ais. It gets into the hands of the. 0:17:53.160 --> 0:17:58.840 Enterprise that was in Thropic's head of EMEA North, Pip White. 0:17:58.920 --> 0:18:02.880 Let's dive a little deeper into concerns Andthropics Mythos has 0:18:03.000 --> 0:18:08.240 raised even within the company. Bloomberg cybersecurity reporter Margie Murphy 0:18:08.560 --> 0:18:11.040 has been looking at exactly that in today's Big Take. 0:18:11.160 --> 0:18:14.880 This is an important body of work, timely and pack 0:18:14.960 --> 0:18:18.680 with detail about first the origins of Mythos, why they 0:18:19.160 --> 0:18:22.400 started with the cyber security capability, it's not the only capability. 0:18:22.720 --> 0:18:25.719 Take us inside that report and what we need to know. 0:18:27.200 --> 0:18:32.000 Yeah, so we were really keen to understand how Anthropic 0:18:32.119 --> 0:18:37.120 came to this decision, this really unusual, unprecedented decision to 0:18:37.160 --> 0:18:40.879 not only hold back, you know, what would be the 0:18:41.240 --> 0:18:44.400 latest update to claud and be you know, a great 0:18:44.640 --> 0:18:47.960