00:00We, every week, month, and quarter, wait for the definitive data set that says AI is doing something here.
00:09Have you seen it yet? Will we see it soon?
00:14I really haven't seen it yet. I think everyone is so focused on what the technology can do because the
00:20technology is so amazing
00:21that they're jumping straight to, and therefore we're going to really quickly see these labor market impacts.
00:30I think it's really important to keep in mind there are a lot of things that affect deployment of technology,
00:36IT policies, economic pressures, demographic changes, liability concerns.
00:44And so the question isn't just what can the technology do, but how quickly is society going to,
00:51and companies going to rearrange itself around it?
00:54The news is that, you know, the unexpected cut to jobs in February, that's the news.
01:00That's what's driving markets today. Within the data set, did you see anything?
01:04Did you go deep and say, okay, what can I put from this into my own model?
01:10Yeah, and when you look at this data, I mean, strikes did play a role, so we should see those
01:15jobs coming back.
01:16But even if you take the strikes out, about 50 percent of the job loss was in goods-producing sectors
01:21as opposed to private service-providing sectors, so things like manufacturing, construction, mining, and logging.
01:29Those are not sectors that we expect to be affected by AI in the same way.
01:34And for context, right, they were about 50 percent of the cuts, but they're only 16 percent of employment.
01:39And so that doesn't suggest that this is AI.
01:42It suggests that, you know, you maybe have impacts of tariffs plus just some general cyclical weakening.
01:48Martha, yeah, we saw in information about 11,000 jobs gone out of those 92,000.
01:54Can you tell us what data points we should look at then if it is going to be about productivity
02:00gains,
02:00if it is going to be more about different areas that are starting to see the impact of an AI
02:06agent, for example?
02:08Yeah, one of the things that I've been looking at is how quickly the just composition of jobs in our
02:14economy is changing
02:15because we gain and lose jobs all the time.
02:17That's a pretty standard part of an economy.
02:20And so just because we lose jobs in one month doesn't mean that we should immediately assume that it's AI
02:26-related job loss.
02:27But if we're losing jobs predominantly in one sector or a couple of sectors that are unusual that look like
02:34AI,
02:35that's a signal that maybe this is starting to happen.
02:38And that's why, you know, at Budget Lab, we've been tracking this, you know, measure of occupational distribution every month.
02:44And it still isn't really flashing anything.
02:47I think it's really important to separate out the vibes on AI in the labor market from what's happening in
02:53the data.
02:54And, you know, I think you're just seeing broad agreement that it's not in the data yet.
02:58You just saw a report from Anthropik come out yesterday saying, yeah, we're really not seeing macroeconomic impacts here yet.
03:05Vibes are one thing.
03:06But Jack Dorsey saying he's cutting half of his staff is a data point.
03:11And, yes, we can say whether there's some AI washing within it.
03:14But still, there are executives saying we are not backfilling.
03:18We are using AI.
03:19We are cutting half of our staff because we think AI is at a cusp where we can do more
03:24with it, Martha.
03:25When does that start to leach in?
03:28I do think tracking CEO statements is not the ideal way to look at this for a couple of reasons,
03:35right?
03:35One is there's huge selection in who even makes a statement about layoffs.
03:39Your hardware store down the street is not issuing a statement every time they lay someone off.
03:44There's selection effects in whose statements get covered, right?
03:48The statement from Jack Dorsey was, you know, very eye-catching, and so it got picked up.
03:52And also, frankly, you know, CEOs have incentives to talk about things in a certain way.
03:58Shareholders want to hear that they're making AI investments and driving productivity.
04:03And they may not even have a perfect sense of what hiring would look like in the absence of AI.
04:09And so I do think it's a moment where, you know, yes, we should listen to people about what's going
04:13on.
04:13We should listen to CEOs, but it's really important to go back to the data and make sure we're taking
04:18a really data-driven approach here rather than, you know, following the latest company announcement.
Comments