WEBVTT - Single Best Idea with Tom Keene: Michael Chui

0:00:02.520 --> 0:00:16.560
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news, single best idea and

0:00:16.600 --> 0:00:18.840
<v Speaker 1>today we focus on one voice and it was a

0:00:18.880 --> 0:00:22.119
<v Speaker 1>phenomenal day, all sorts of good people to talk to.

0:00:22.280 --> 0:00:25.439
<v Speaker 1>Dan Ives was on with Webbush and he was just

0:00:26.000 --> 0:00:31.600
<v Speaker 1>blistering about the significance of this Oracle move. To summarize

0:00:31.600 --> 0:00:33.159
<v Speaker 1>here and not take a lot of time on it.

0:00:33.600 --> 0:00:37.599
<v Speaker 1>Oracle leapt over thirty percent today on the market opening.

0:00:38.040 --> 0:00:40.479
<v Speaker 1>I said this on air. I'll say it again. I

0:00:40.600 --> 0:00:45.239
<v Speaker 1>have never seen a blue chip stock leap have a

0:00:45.320 --> 0:00:52.600
<v Speaker 1>complete restructure reanalysis. After what we saw from Oracle last evening, profound,

0:00:52.640 --> 0:00:56.639
<v Speaker 1>profound performance. I should mention for clarity, Oracle is one

0:00:56.640 --> 0:00:59.400
<v Speaker 1>of the good sponsors of Bloomberg Surveillance. We thank them

0:00:59.600 --> 0:01:02.720
<v Speaker 1>for that support. But that's just a stunning, stunning moment,

0:01:02.720 --> 0:01:05.920
<v Speaker 1>and many others as well. All Senator on Annawan's important

0:01:05.959 --> 0:01:10.480
<v Speaker 1>Bloomberg Economics essay, saying, look, the employment revision yesterday, Guess

0:01:10.480 --> 0:01:13.240
<v Speaker 1>what it means. We've been in some form of recession

0:01:13.600 --> 0:01:16.319
<v Speaker 1>going back to the spring of twenty twenty four. She

0:01:16.360 --> 0:01:19.920
<v Speaker 1>finishes her note by saying, well, things look better now.

0:01:20.040 --> 0:01:23.959
<v Speaker 1>Maybe we're at the beginning of a more constructive business cycle.

0:01:24.480 --> 0:01:27.959
<v Speaker 1>What a treat today to have in from McKinsey Global Institute,

0:01:28.000 --> 0:01:32.360
<v Speaker 1>Michael Cheua, he's out of Stanford and at Indiana did

0:01:32.640 --> 0:01:37.160
<v Speaker 1>absolutely definitive academics on what we do every day, which

0:01:37.200 --> 0:01:41.160
<v Speaker 1>is search. Michael Chewa of McKenzie on search.

0:01:41.319 --> 0:01:44.600
<v Speaker 2>For several years, you know, AI usage, a regular usage

0:01:44.640 --> 0:01:47.720
<v Speaker 2>of AI had sort of plateaued around fifty percent, you know,

0:01:47.840 --> 0:01:53.080
<v Speaker 2>since the advent of jenerative AI chatchypt Claude Gemini, and

0:01:53.120 --> 0:01:55.840
<v Speaker 2>it's like, you know, that is really accelerated. And now

0:01:55.880 --> 0:01:58.440
<v Speaker 2>we see about ninety percent of the companies in our

0:01:58.480 --> 0:02:02.160
<v Speaker 2>survey saying that they're using A regularly. At the same time,

0:02:02.480 --> 0:02:06.080
<v Speaker 2>the majority are not saying it's having a significant impact

0:02:06.160 --> 0:02:08.640
<v Speaker 2>on its EBIT yet. But what we do see is

0:02:08.639 --> 0:02:14.000
<v Speaker 2>that the individual use case or business function level, it's

0:02:14.040 --> 0:02:16.480
<v Speaker 2>starting to create real value. And so what we say

0:02:16.520 --> 0:02:21.480
<v Speaker 2>is that's a process. To your point about industries, Yeah,

0:02:21.520 --> 0:02:25.000
<v Speaker 2>the tech industry, for instance, has moved ahead. But in fact,

0:02:25.160 --> 0:02:27.720
<v Speaker 2>rather than an industry having quote unquote figured it out,

0:02:27.720 --> 0:02:30.360
<v Speaker 2>what we're finding is lots of variation within industries. And

0:02:30.400 --> 0:02:32.520
<v Speaker 2>so again the companies that have figured out how to

0:02:32.560 --> 0:02:36.200
<v Speaker 2>rewire themselves right, whether it's travel and logistics or high tech,

0:02:36.520 --> 0:02:38.840
<v Speaker 2>they're just accelerating past their competitors.

0:02:39.040 --> 0:02:42.640
<v Speaker 1>We continue with Michael Cheua of Stanford, of Indiana, of

0:02:42.720 --> 0:02:46.680
<v Speaker 1>McKenzie and basically, if to a dummy like me, it's like, okay,

0:02:46.760 --> 0:02:50.760
<v Speaker 1>late nineteen ninety four, early nineteen ninety five, a revolution

0:02:51.280 --> 0:02:52.840
<v Speaker 1>Tua on the view forward.

0:02:53.040 --> 0:02:54.720
<v Speaker 2>I think one of the interesting things, and it's a

0:02:54.760 --> 0:02:57.080
<v Speaker 2>bit of a the you know, a debate within the

0:02:57.720 --> 0:03:01.320
<v Speaker 2>artificial intelligence community. A lot of the excitement now is

0:03:01.400 --> 0:03:04.600
<v Speaker 2>with these quote unquote neural network models, and that's underpinning

0:03:04.639 --> 0:03:07.040
<v Speaker 2>a lot of what you know, we currently often describe

0:03:07.080 --> 0:03:10.440
<v Speaker 2>as AI. But you know, AI started in the nineteen fifties,

0:03:10.520 --> 0:03:13.360
<v Speaker 2>or the term was invented in the nineteen fifty five,

0:03:14.120 --> 0:03:17.280
<v Speaker 2>and for a lot of that time, you know, the term,

0:03:17.680 --> 0:03:21.160
<v Speaker 2>you know, symbolic systems at Stanford comes from using logic,

0:03:21.320 --> 0:03:25.000
<v Speaker 2>using this way that we've managed to formalize reasoning. You know,

0:03:25.320 --> 0:03:27.160
<v Speaker 2>one of the things that people who use neural network

0:03:27.200 --> 0:03:28.880
<v Speaker 2>models are trying to do are trying to make it

0:03:29.080 --> 0:03:31.480
<v Speaker 2>reason better. But we also had all these other types

0:03:31.520 --> 0:03:34.320
<v Speaker 2>of symbolic ways of reasoning, and I think we're also

0:03:34.400 --> 0:03:37.400
<v Speaker 2>starting to see the hybrid models, and so I think

0:03:37.400 --> 0:03:40.720
<v Speaker 2>what we're seeing is, you know, history again, doesn't repeat itself,

0:03:40.760 --> 0:03:42.760
<v Speaker 2>but it does come back. We've had these debates about

0:03:42.800 --> 0:03:45.000
<v Speaker 2>what the best way to create AI has been, and

0:03:45.040 --> 0:03:45.880
<v Speaker 2>that's moving forward.

0:03:46.240 --> 0:03:49.360
<v Speaker 1>Mackenzie gives great access to those reports. Go to Mackenzie

0:03:49.360 --> 0:03:52.960
<v Speaker 1>and look for the report in March an Artificial Intelligence

0:03:53.040 --> 0:03:58.120
<v Speaker 1>AI with, among others, Michael Chew. We're on podcasts right Apple,

0:03:58.120 --> 0:04:01.960
<v Speaker 1>We're on spotif hand YouTube podcast. It's single best idea.

0:04:14.840 --> 0:04:14.960
<v Speaker 1>Hm