Have the AI job losses started?

0:28Steve Kovach in the workplace. So what's the truth? Is AI already hitting the job market? And if not, when and whose jobs will be the first to go? Hello, and welcome to Beyond the Valley. I'm Arjun Karpal, and I'm so excited to introduce my new co-host, Steve Kovac. Steve, welcome to Beyond the Valley. Hey there, Arjun. It's good to be working with you. And I'm really excited. I'm CNBC's tech correspondent over here in the United States. I'm based in New York City, so I'm glad we're bringing a global element to Beyond the Valley. I'm really excited to join you. Steve, you've been covering tech for a long time.

1:04Just give us a quick background on you. Yeah. So I've been CNBC here. I just hit my seven-year mark, believe it or not. Congratulations. Thank you. Yeah. And I've been covering tech pretty much my entire professional career. Specifically here at CNBC, I focus on Apple, a little company maybe you've heard of. And Microsoft is my other big beat company. And then I also kind of dabble in some other things. When I was kind of putting together my thesis for my coverage here on CNBC, I thought, you know, video games, people don't realize how big video games are and what a big business that is. So I added gaming to my beat element, too.

1:44So I cover, you know, companies like Take-Two Interactive and the Xbox division over at Microsoft and all those cool companies doing interesting things in entertainment as well. I feel like any of our current listeners and any new listeners we pick up along the way, they're going to hear a lot about gaming. You and I both love our bits of gaming. I already have a Nintendo Switch 2 at home. That's my weekend plan. There you go. I just opened a new game yesterday, so we're on the right path. Just two grown men talking about toys. It's going to be great. And look, Steve, we're going to be talking about AI today.

2:16And I think that chat GPT moment we had a couple of years back now was such a big moment. Um, you cover, like you said, Apple, Microsoft, so many of these big companies as well. What's your sense at the moment of kind of what the big themes are amongst the big tech giants when it comes to AI? Yeah, I've been kind of toying around with this because I focus so much on Apple and they aren't exactly at the forefront of AI. But what I've been kind of thinking a lot about, Arjun, is like this build versus partner or build on top of. So, and this is kind of emerging out of Apple. Obviously, we saw this year they were going to put out that new super version of Siri.

2:55It was going to be personal. It's going to use the data on your iPhone. It's going to be chat GPT, but with an Apple twist, right? And that was what everyone got excited about. And they failed to deliver on it. And it was the first time, probably the worst time I've seen Apple go through as far as promising a major product and then failing to deliver. You can argue things like the air power, power charger thing, or, you know, little things like that. But as far as something coming out and saying, we're going to take part in this AI revolution and then failing to do it. And then now we hear these reports, Arjun, that they might put that to the side and partner, just like they do in search with Google.

3:32They might start partnering with OpenAI or Anthropic to kind of power the foundation of this new version of Siri and then move on from there. We're seeing that with Samsung. I was just at the Samsung event this week. We're recording this the week of July 10th. And it was so interesting to see them. they're not building their own AI. They're using Google Gemini. And Perplexity, one of the hottest companies right now in the startup world, they're not building their own LLMs. They're building a layer on top of the existing thing. So I'm kind of looking at it from that perspective now and seeing as these companies are spending enormous amounts of money building all those data centers, as they're spending so much money trying to achieve artificial general intelligence and the talent wars that we're going to get into in a little bit, where's all that money going?

4:17Who's building cool products on top of it. It's not necessarily the LLM makers right now. It's going to be, you know, there's this idea now it's going to be someone building on top of that. So that's the theme I've been watching. Yeah, those are some great themes there, Steve. And, you know, we're going to dig into those in some future episodes, no doubt. But let's get in to what we're talking about today on the podcast. We're talking about AI and whether the threat to the job market is real and happening right now, or maybe it has been overhyped, or maybe it's a bit both. To help us explore all of this and more, I have a very special guest joining us today, Michal Levram, who is a contributing editor at Fortune and also a CNBC contributor as well.

4:59Michal, thanks so much for joining us. Thank you guys for having me. It's great to be here. And we've got quite the global footprint between the three of us. Yeah, absolutely. We've got the East Coast, we've got the West Coast, and we're here in Europe. Michal, look, we introduced this podcast talking about the layoffs we saw at Microsoft. You know, the companies argued that it was nothing to do with AI. What's your view? Is it to do with what we're seeing in terms of the efficiency gains of AI, the productivity gains of AI, or is there something else going on here? Yeah, I think my sense is that it absolutely has to do with AI, but not in the way that everybody thinks or is scared of.

5:38I think that what we're seeing is not, you know, engineers or other workers losing their jobs as a direct consequence of AI overtaking jobs. I think we're pretty far from that at this point. I personally don't know a single software engineer who has lost their job to AI, been replaced by AI quite yet. Not saying it won't happen in the future. But what I do think is going on, and the reason it is directly related to AI, is that companies like Microsoft, obviously, we will see others as well, are sort of, you know, restructuring, reprioritizing, reorganizing, which, you know, is a nice way to say, laying people off, ultimately, in order to meet the demands of this new era of this technological transformation.

6:23So it is related, it is indirectly a cause of AI, but I don't think we're seeing people losing jobs because AI agents are taking them over. I don't know if you caught yesterday, again, we recorded this on July 10th, so for people who want to go back and look, Bloomberg had an interesting story yesterday talking kind of the fallout from those layoffs that we heard about last week. And one of them was the executives telling staff, we've been able to save$500 million through like AI, you know, productivity benefits or what have you. And, you know, more talk about how much code is being written by AI Microsoft.

6:56You know, I look at this with a much more cynical or skeptical view, probably than a lot of people. You know, my first question is, how do they even quantify that? I mean, you know, I talked to so many CTOs about using, you know, adopting these tools, and they have a hard time, you know, quantifying this kind of stuff. What was your take on that when you heard that$500 million number, which, by the way, is nothing to Microsoft? Like, yeah. True. Drop in the bucket. Totally. Yeah. I mean, look, I definitely hear from people out here in the Valley who are absolutely vibe coding. Like I think that that movement, that trend is real.

7:35But, you know, to what point? And also what's all of the other work that's involved? And so a CEO can get out there, a tech CEO, and say, you know, these are our savings or 20 to 30 percent of our lines of code are now being written by AI. But then what's all of the other work that's involved? And, you know, you've got to test it. You've got to check it. There's so much other, you know, there are other aspects of the job of an engineer. And so how good is it at what it's doing? You know, I don't think we're measuring that. And I agree with you. I think measuring just in general, the savings part, the dollars part especially, is tricky because maybe it's writing all this code, but then it's requiring more tests after the fact, right?

8:21I don't think that we're clear on exactly what the numbers are telling us and what the numbers are quite yet. But again, I do think that it's happening. I mean, I do hear from people that they're doing it. They're vibe coding, you know. So I think we're very early, though. I've not talked to a single programmer or coder who says they don't use an LLM of some sort to help them out. What do you think about entry-level jobs, though? Because I was talking to a colleague of mine here, and his son is a computer science major going into this world soon to graduate. And there's this kind of idea that his level of entry-level computer engineers are kind of being – it's tougher to get those jobs.

9:06Have you been hearing that out there? Not quite yet. I mean, honestly, what I hear more of is still concern. Maybe this is a surprise, but still concerns of losing jobs to offshoring. You know, that's much more prevalent. And I do think that that's, you know, that's much more prevalent among entry level jobs. Right. But again, where things are going, the trajectory, yes, I'd be very concerned. You know, we've got meta poaching people for hundreds of millions of dollars, which I know we'll get to. But then on the other end of the spectrum, like those people had to start somewhere, maybe not their PhDs.

9:43They came out of school with, you know, these accolades, but, but, but a lot of other people do have to start somewhere. So, so how do you get ahead? Because we're seeing this like, you know, huge market, huge demand for top talent, but top talent has to start somewhere. So I do think that that's a risk and that's something that, you know, pipeline wise companies are really going to have to think through because how do you cultivate that talent that you're going to need at some point? Mikhail, just before we get into that talent conversation, because it's a fascinating one, are there areas at the moment you are seeing more disruption to the job market in as a result of AI than others?

10:21And by the way, vibe coding is probably one of the best jargony words to come out of tech in a while. It's quite fun. But yeah, are there other areas you're seeing because I know there's a lot of talk about, you know, customer services and call centers and these kind of areas. But yeah, just give us a sense of what you're seeing. Yeah, I haven't seen numbers on impact of customer service. I definitely think that that's going to be a big one that's impacted. And we've seen that kind of play out over time. It's not an entirely new phenomenon, of course. And again, with that one, there was a lot of offshoring.

10:54And I think that, you know, there still is, obviously. But at this point, yes, there's risk to those offshore jobs by American companies and other companies, you know, losing jobs to AI. I think that in a lot of areas, probably, the real risk longer term is have we reached kind of peak employment, you know? And that's a really interesting trend in my view. I just think it's a bit early to see, you know, there are certain jobs where it makes sense, absolutely, that we'll see more real losses. But the real shift that I want to be watching is, you know, it used to be the fact that, especially among tech companies, but other companies as well, when they grew their business, they grew their workforce.

11:45And I wonder if like the 2021, 2022, you know, crazy boom that we saw in hiring as these businesses had to expand because of demand, which was great for them, if that was a peak and like we'll never get back to it. Can I just clarify, peak employment you're talking about in the tech industry? So, yes, max employment. Wow. Yeah, such a fascinating idea. Yeah, I mean, if you look at the numbers over time and, you know, employment includes their, you know, I mean, some of them, of course, outsource this, but it's across the board, right? It is customer service. It's all sorts of areas. It's not just engineers.

12:25If you look at the trends, you know, over the decades for some of the bigger tech companies, every time the business grew, every time there were these inflection points, the workforce grew. They needed more people, right? And then there are, you know, these little tiny dips where there have been layoffs when, you know, the economy is bad, whatever is going on, right? Or they're restructuring. But ultimately, it's been like this crazy, crazy growth curve over the years. And in 2021, 2022, for some companies, we saw a significant peak. We saw it grow significantly. And then it's dipped off a bit since.

13:05And I guess my question is, you know, was that kind of like the last big, huge growth curve? Because if what some of these CEOs are saying is true, you know, maybe for future growth, including what we're seeing now, yeah, they need to add AI talent. But I don't know if they need to add in the same kind of numbers that we've seen in the past. That's all. I don't know if either of you two have interacted with an AI chatbot customer service thing, but it is a nightmare. Remember those old phone trees you call a customer service line and you have to slam on this zero button in order to get a human being to talk to you?

13:46There's no zero button with these. It's like you cannot get what you want. Long story short, I was stuck in a chatbot death loop of trying to get it to do what I want. Eventually, I had to, you know what, pick up a phone and call a human being to get done what I wanted to do. I had another hotel thing do the same thing, couldn't change my reservation. So I think we're a little bit of way off. I know we're going to get into this in a second, but like this is Mark Benioff's whole pitch, right, that we're going to be doing customer service via chatbots. But in my experience, it kind of stinks. Maybe you'll pay a premium to talk to a human.

14:20You know, I could totally see that happening. I would do it. supposedly people are nicer to when they're interacting with bots than when they're interacting with humans though which i think is really interesting it doesn't sound like you were but no no i was like i was swearing at this thing and when they become sentient and try to kill us i'll be the first on the mark uh to to get it yeah you're on the list steve you're on yeah i regularly find myself saying get me a human can you get me a human please can i speak to a human they don't know what's going on the digital version of pressing zero on your on your keypad yeah

15:21Beyond the Valor. wherever you're listening to this.

15:28Let's move on to another key part of this conversation. And we've spoken briefly about it earlier on. It's talent, right? We're hearing all these stories about Meta offering hundreds of million dollars for talent of top talent to come in to work on their super intelligence. There's a big almost talent land grab going on right now amongst big tech companies as well. how how do you see this playing out i mean this this is so intense and i've spoken to kind of people here in europe who are sort of seeing a similar thing as well uh big tech grabbing talent here in europe one because there's a lot of it uh two because it's cheaper uh as well but how do you see this kind of land grab for talent playing out michael yeah i mean i think uh it's not entirely new and we've seen speaking of of of uh you know overseas uh in in London, like we've seen a lot of people poached from DeepMind, you know, which is a huge, huge hub of incredible talent.

16:25Europe's got incredible talent overall. Lots of, you know, people coming out of universities there. What we're seeing now out here as well in the Valley, I think, is super interesting because it's a phenomenon that, and Steve, I don't know about you, I have not seen before, not in this kind of, you know, at this scale. these kinds of figures being thrown around. Oh, my God. It's insane. No, not for regular talent. Usually they buy the company for that money, but not this. Yeah. Exactly. Exactly. And so, you know, I think first of all, just the fact that it's so unprecedented and that like, you know, I've heard and seen memes about this, that like the nerds are finally as hot and as like in demand as like, you know, sports as athletes, right?

17:15That it's like open trading period. but with AI scientists. And so there's that just as a phenomenon, which is really unprecedented. But I think what's also interesting about what's going on, it's clear what the motivation is, a company like Meta especially is trying to figure out what its strategy is and trying to compete here. I think that part is obvious. But the other aspect here is that they're not just getting talent in-house in a way that they see as an open avenue for them. But they're also, at the same time, I think, denting their competitors, right? They're damaging their competitors.

17:54And that's also super fascinating as a strategy, as a tactic. And it's quite ruthless, but I think we're going to see more and more of that happening. It's just incredible to watch. Is your impression, so we have our CNBC events team, we hold these dinners with tech executives, And it's Chatham House rules, so I'm not allowed to say who said this, but it was an executive who said this in the know in Silicon Valley, et cetera. And he's a quantum guy and said, you know, when you look at quantum talent, there are only like a few hundred people in the world who are like really capable of doing this kind of stuff.

18:30When you move into AI, it gets bigger, but it's only like 8 ,000. And, you know, these top talent of it's literally the people capable of building this stuff and working on this stuff and who have researched this stuff. Is that your impression as well, that there might just the reason these enormous, you know, we hear these stories just this week that, you know, Meta poached Apple exec and for a package worth$200 million. That's insane. Usually you buy a company, not a person for that. So is it just a dearth of talent that that's causing these prices to go up? I absolutely think so. I also think that, you know, to your point, usually you buy a company for that amount of money because the AI valuations have been so insane and so hot and so over the top.

19:17You know, you can't go and buy, you can't easily, I should say, not very many companies can go out there and buy an open AI or an Anthropic or whatever, right? And so what you can do is spend several hundred million dollars and get this top talent to come walk through your door happily. I was really surprised about the Apple poaching. I think that's probably like – I don't think we're going to see a lot of people being poached from Apple because – They have no one. Exactly. I think it should be the other way around, but we'll see what Apple ends up doing here. But yeah, there was a LinkedIn survey from just a little while back that was looking at, and granted it's using LinkedIn data and so it's what people are putting on their profiles, but they basically looked at where are the top concentrations of AI talent.

20:08And the top concentration was in Israel actually, and then Singapore, and then it was kind of smattering across Europe, different European countries. You know, that said, these are really small numbers still. And so I do think that, I mean, the concentration I think for Israel was like 1.98 % of workers in the workforce, something like that. It was under 2 % have, you know, AI skills. I do think the concentration of really high level, like, you know, AI scientists, PhD level, the kind of people you really need in your door is still quite small and it takes time to build up that kind of talent.

20:54but there are also other jobs across companies where AI is a part of it and of course we all know that at this point it's like the most important thing you should do whatever job you're going for is to be AI fluent and all of that so I think also the way that we look at AI skills is probably going to change and it's very different if you look at different parts of the spectrum in a corporation but for the top talent yes there's there is we wouldn't see these kind of numbers if there was if there were a lot of people exactly supply and demand of the talent well it's a good thing uh here in america arjun uh the kids are more focused on becoming uh influencers on tiktok as opposed to uh ai phds or whatever so you know we got the same problem we got that going for us which is yeah whereas in china you know they're learning all this stuff at age four or whatever so yeah yeah Yeah, absolutely.

21:48I mean, like China has been pretty, I think there was a report maybe last year or so, some of the schools in Beijing were adding sort of these AI courses into kind of lower school classes. So, you know, China is very focused on this, of course, necessarily because also the amount of restrictions there are on China in terms of access to technology, etc. They want that next generation to be able to come up with innovations that can put them back on the map when it comes to some of these foundational technologies. You know, there's loads of restrictions on chips and that's hampering perhaps some of the development that is possible out in China.

22:25And so they've laser focused on all of these and they play a long game in China. We know that. They're very long game. They think long term. I wanted to also maybe give the European perspective a little bit as well, because, Michal, you mentioned London as well. And some of these companies like DeepMind, of course, owned by Google. What we're seeing a lot of here in Europe, in the UK and elsewhere on the continent is a lot of these people who were working at either companies like DeepMind or other large US tech companies in particular. they're leaving and are beginning to start up their own AI stuff here.

23:00And there's been some success stories. You think of there's a foundational model company in France called Mistral. They were also like former deep mind execs, et cetera. There's H company, which raised a massive seed round last year. So you are starting to see a lot of that happening here in Europe. A lot of these European founders are coming from US big tech, which I think is a really interesting dynamic because Europe lost out on the last internet age, more or less. There were no major European players. And there's this real kind of fervor here in Europe that actually we need European champions.

23:36We need a European equivalent to open AI. We need the infrastructure players. So that's a big one. I think just naturally, these people coming out of these companies. But two, there is kind of this political will to some extent in certain countries like France, for example. Is that in terms of the big companies, the new startups coming out in the US. Is that a similar kind of thing? They're coming out from these established players and trying to set up their own stuff? I think we're starting to see that. I mean, I and I and I think we'll see a lot more of that. I mean, Anthropic is an example, right?

24:11We've we've certainly seen quite an exodus from open AI in particular. And so with Mira Marati and others going out and creating their own AI companies. I don't think that all of them are trying to create the kind of foundational models, you know, from scratch. There's all sorts of, you know, other products and layers to AI that people are looking at. But yeah, absolutely. And by the way, one of the interesting phenomenons, and I don't know how big this will get, I don't know what you guys think, but, you know, Sam Altman's whole vision for like we're going to see one person, billion dollar companies in the future because they're aided by AI, because you can grow a company to be quite, you know, successful, big, whatever.

24:57Doesn't he have a bet going on with his friends like when that's going to happen or something? Didn't he say that like when the first unicorn single human being company happens? I wonder how much money they're putting on this. I don't even want to know. Right. If they're betting on it, that's hilarious. But so we have not seen a billion-dollar company created, but there was just recently an acquisition of an Israeli company, Base44, which is basically a solo shop sold for$80 million to another company after just six months. And it's basically a vibe-coding company, and it's basically like one dude.

25:37And so I think that I don't think that that works for, you know, all kinds of products. You don't you don't build the next Microsoft with one person. That's, you know, even if we've reached peak employment, that's never going to happen. But for some products, it might be an option. And so that's also going to be interesting if we have people leave existing AI companies and go on to just, hey, I'm going to just grow this, you know, sell it, see what happens and not hire a whole mess of people. You also got to be aware, too, of like the charlatans out there. And this is like prime environment for such a thing where, you know, it's very easy to raise money.

26:19You mentioned Mira Mirati, the former CTO of OpenAI. She raised, what,$1.5 billion at a$10 billion valuation, something like that. No product. No team. Like literally nothing other than she used to work at OpenAI and be the CTO. and then I'm totally blanking. This was a European company, Arjun. I'm totally blanking on their name, but they were like a no-code AI company and they just went belly up. I'm completely blanking on that startup's name. But this just happened a couple of days ago. But there are a number of examples, like we saw this in Web 2.0 in the mobile era, you know, lofty promises and under-delivering too.

26:58So I guess my question, because, you know, we hear Mark Zuckerberg hiring all these people how much they're paying and so forth. And then Mark Zuckerberg comes out and announces this super intelligence group and they're going to do something. What is that something? Do you have any sense of what, it's just like, we're going to build AGI. Okay. And then what? What does that mean? There's no product there. So what do you think the product is? You know what? I'm with you, it is not clear to me what other than the need to jump on the bandwagon and, you know, be a part of this new technological transformation, what the real strategy is.