WEBVTT 00:00:00.240 --> 00:00:04.557 So, just to back up a bit, I've been working on AI for uh for 10 years, and 00:00:04.640 --> 00:00:08.397 probably the thing I've noticed most about it is is how fast it's making 00:00:08.480 --> 00:00:11.997 progress. Two years ago, it was at the level of a smart high school student. 00:00:12.080 --> 00:00:14.877 Now, it's probably at the level of a smart college student and reaching 00:00:14.960 --> 00:00:19.277 beyond that. And on one hand, I think there's a number of very positive things 00:00:19.360 --> 00:00:23.517 that are going to happen. I used to be a biologist. I think AI has exactly the 00:00:23.600 --> 00:00:27.837 kind of skills that are needed to cure important diseases like cancer and 00:00:27.920 --> 00:00:30.477 Alzheimer's to provide cheaper energy 00:00:30.560 --> 00:00:33.037 many positive things but exactly those 00:00:33.120 --> 00:00:35.597 same kind of skills things like 00:00:35.680 --> 00:00:38.237 summarizing a document brainstorming 00:00:38.320 --> 00:00:40.317 putting together a financial report 00:00:40.400 --> 00:00:43.037 makes me worry a lot that entry-level 00:00:43.120 --> 00:00:47.277 jobs in areas like finance consulting um 00:00:47.360 --> 00:00:50.477 tech uh many many other areas like that 00:00:50.560 --> 00:00:53.277 entry level white collar work. I worry 00:00:53.360 --> 00:00:54.797 that those things are going to be first 00:00:54.880 --> 00:00:57.997 augmented but before long replaced by AI 00:00:58.080 --> 00:01:02.477 systems and that we we we may indeed it's hard to predict the future but we 00:01:02.560 --> 00:01:06.317 may indeed have a have a serious employment crisis on our hands as the 00:01:06.400 --> 00:01:11.437 pipeline for for these the early stage white collar work starts to uh starts to 00:01:11.520 --> 00:01:15.917 contract and dry up. >> What's the timeline here? 00:01:16.000 --> 00:01:17.997 >> Yes. So it's very hard to predict these 00:01:18.080 --> 00:01:23.357 things given how fast um you know given how fast AI is making progress. Um but I 00:01:23.440 --> 00:01:27.917 would I would not be surprised if somewhere between one and five years um 00:01:28.000 --> 00:01:31.837 we started to see big effects here. You know, I've heard a number of people talk 00:01:31.920 --> 00:01:34.077 about this in private. AI CEOs talk 00:01:34.160 --> 00:01:36.477 about this in private. CEOs of other 00:01:36.560 --> 00:01:39.277 companies talk about this in private. Um 00:01:39.360 --> 00:01:43.677 uh and I I really felt that uh you know the message that this is happening 00:01:43.760 --> 00:01:48.157 hasn't been getting out to to ordinary people, hasn't been getting out to our 00:01:48.240 --> 00:01:49.997 legislators, our Congress people either. 00:01:50.080 --> 00:01:52.397 And so I felt I needed to speak up on 00:01:52.480 --> 00:01:54.317 the record. I do think that this is 00:01:54.400 --> 00:01:57.357 something we can prevent this uh but but 00:01:57.440 --> 00:02:02.077 we need to act now. I don't think we can stop the AI bus. Um, you know, there are 00:02:02.160 --> 00:02:06.877 there are six or seven companies just in the US working in this air working in 00:02:06.960 --> 00:02:10.877 this area. I just run one of them. Even if our company stopped doing what it was 00:02:10.960 --> 00:02:14.557 doing today, all the other companies would would continue. Even if all six 00:02:14.640 --> 00:02:18.317 companies started, then China would beat us. I think that's a big and important 00:02:18.400 --> 00:02:22.397 threat. So, we can't stop the bus. But I do think we may have an opportunity to 00:02:22.480 --> 00:02:26.637 steer it. And and I think you know we need to get congress people, 00:02:26.720 --> 00:02:30.717 legislators, the public to to think about this. Now one of the first steps 00:02:30.800 --> 00:02:34.237 we're taking is just measuring the effects. Enthropic has put out something 00:02:34.320 --> 00:02:38.957 that we call our economic index. And using the economic index, we try and 00:02:39.040 --> 00:02:42.637 track the rate at which these things are happening. You know, at some point we 00:02:42.720 --> 00:02:46.557 have to go beyond measuring this to actually taking action to trying to 00:02:46.640 --> 00:02:51.677 build our AI systems in a way that augments rather than replaces. But that 00:02:51.760 --> 00:02:55.037 won't be a long-term solution either. We need to think about what what the 00:02:55.120 --> 00:02:58.157 government is going to do. And you know, there's a wide range of actions that 00:02:58.240 --> 00:03:02.557 that the government uh uh the government could take. And you know, it's it's not 00:03:02.640 --> 00:03:06.317 not my place to decide which of those are which of those are correct. But but 00:03:06.400 --> 00:03:09.517 I think the first step is people need to think about this and start taking it 00:03:09.600 --> 00:03:12.157 very seriously. >> Yeah. And you're talking about AI 00:03:12.240 --> 00:03:15.197 possibly wiping out all entrylevel white 00:03:15.280 --> 00:03:17.917 collar jobs and spiking unemployment to 00:03:18.000 --> 00:03:20.957 10 to 20%. So dive into the mitigation 00:03:21.040 --> 00:03:25.277 of this a little bit more. How could we mitigate the worst possible scenario 00:03:25.360 --> 00:03:28.877 here? >> Yeah. So it's hard to know the exact 00:03:28.960 --> 00:03:32.077 numbers. I mean, you know, I don't have a crystal ball. I can't I can't predict 00:03:32.160 --> 00:03:35.997 the future. But what I will say is I think what AI can do, you know, there 00:03:36.080 --> 00:03:40.077 have been these technological changes in the past, but I think what's different 00:03:40.160 --> 00:03:42.877 about AI is it's broader, it's deeper, 00:03:42.960 --> 00:03:47.277 and it's moving faster than than previous uh such technologies. And so in 00:03:47.360 --> 00:03:51.757 terms of mitigations, you know, I think I think one is just educating people in 00:03:51.840 --> 00:03:55.517 order in order to you in order to use AI. There will always be people who are 00:03:55.600 --> 00:03:57.917 who are who are using AI and helping our 00:03:58.000 --> 00:04:00.637 workers to adapt faster I think I think 00:04:00.720 --> 00:04:03.197 is an important move. A second move is 00:04:03.280 --> 00:04:08.077 that we may want we may want government to you know to to to kind of to kind of 00:04:08.160 --> 00:04:12.157 find a way to level the economic playing field and you know this may be a 00:04:12.240 --> 00:04:16.317 controversial proposal but you know something like taxing AI companies like 00:04:16.400 --> 00:04:18.957 us we we may need to get >> and that's a very interesting message 00:04:19.040 --> 00:04:21.517 coming from you who runs an AI company. 00:04:21.600 --> 00:04:24.957 Uh but you're doing a public service announcement for all of us uh in terms 00:04:25.040 --> 00:04:28.480 of this future concern. andio, thank you.