AI agents got so good that this top Meta exec quit her job

She spent decades building software for businesses at Salesforce and Meta. Then she watched her own agents beat her best employees, and she says it radicalized her. Now, Clara Shai runs a nonprofit for the two million recent graduates who may not be able to find a job. On the final episode of this show, what we learned and what happens next. That's this week on Platformer. This episode is brought to you by Jira by Atlassian, where teams and coding agents get the context they need to do the right work. Try it free at jira.dev. That's J-I-R-A dot D-E-V. Welcome to Platformer. I'm Casey Newton. This is the final episode of the Platformer podcast, for now at least, and I wanted to end where we started. Fourteen episodes ago, I asked whether the AI jobs apocalypse was coming. Since then, we've talked to people who are building the technology, economists who are measuring it, and this season, people who are betting their companies on new AI tools. Today's guest has been all three. Clara Shai was the CEO of Salesforce AI, where she helped build Agentforce. She then ran business AI at Meta, building the agents that now answer customer messages for businesses on WhatsApp and Instagram. And this spring, she left to start a nonprofit called the New Work Foundation and a media platform called Dear CC for recent graduates who are worried that the first rung of the career ladder disappeared while they were in school. She calls this the biggest reorganization of human labor ever. We're going to ask her what she saw from the inside, what she's doing about it, and what entry-level workers should be doing today. But before we do that, here to give us some news about the state of AI and jobs one last time is Platformer fellow and Gen Z AI correspondent Ella Marquianos. Ella, how are you? I'm doing well, yeah. I'm excited to be, you know, coming into the podcast one last time before, you know, the robots kill us all. And therefore, like, we—the, like, greatest loss of this comes, which is, of course, us being unable to continue bringing you news stories in the world of AI jobs. That's—I mean, are you at least happy that you were able to have a complete podcasting career before the total automation of labor? That was always number one on my bucket list, is to be a podcast guy, so I can die fulfilled. I actually—sorry, this is kind of random. I was actually recently at a play last night called The Last Musical Written by Humans, which is what had this topic on my mind. It was like a musical, an amateur musical, where I appreciate the efforts. I can't speak for the singing quality, about an AI company that is building an increasingly super intelligent AI, which they make— So even in your downtime, you're finding stories about the automation of labor. Yeah, it's all day, every day, running through my head. Well, why don't you tell us what news caught your attention this week? Yeah, so if you've been on the internet, you may have been, like, seeing basically these videos of, like, workers in various jobs in India, whether that's, like, shoemakers, embroiderers, recycling plant sorters, basically, like, strapping these big VR—like, sometimes it's a VR headset, a lot of the time it's just, like, strapping a camera onto their head and basically, like, recording footage of them doing these tasks. And now these tasks, these, like, videos are kind of being contracted in order to send to robotics companies so that they can use these while, like, training their AIs. Yeah, so tell us a little bit more about how this works. Like, what sort of worker might have a camera strapped to their head, and what do the robotics companies want from this data? Yeah, so like right now the variety is like sort of crazy. Like as I was reading through this recent Bloomberg report on this, I saw a list that was like environments including Botox injectors, gas station employees, diamond cutting centers, hospitals, like sort of any task that a person like does with their body and hands right now robotics manufacturers are interested in. You know, like there are, I think partially because there are like a few industries, like for example auto, which we've talked about in previous podcasts, where they're like directly they want robots to do these things specifically. But then also they just really want robots to like learn to use their hands in a human-like way in general. And so like much like we've used like a very wide variety of text data, some of which is like totally unrelated to what most people are asking ChatGPT for, people are looking for like a really wide variety of human body data. And how is the pay for these workers? Well, like according to our terms, like basically like a lot of this stuff we cover, like you've covered so much like content moderation work where like as a Westerner, where you look at the payment numbers, you're like, this is crazy. They're making these companies so much money. This is such little money. In this case, like there's one example this Bloomberg report used that I think was like pretty representative, where there's a worker at a recycling plant. She normally earns 20,000 rupees, which is about $211 every month sorting recycling. And then this headset thing basically gives her another 150 rupees, or $2 an hour. And now for her, like that is like a 50% pay increase. Like she basically told Bloomberg she was able to like start setting aside money for her kids' education, like because of this. Yeah. So workers have a strong incentive to do this. I wonder if they know very much about what happens to the data that they are capturing. The answer is, like, I think basically no. Like, at least according to Bloomberg, like, a lot of people just, like, don't really understand what the data is going to be used for. A lot of the time, there's, like, no written contract or, like, a few lines of contract that, or it's not even, like, clear that it's enforceable. So, yeah, there's also, like, that, like, very clear ethical issue for a lot of this work. When you think about all of these human beings being drafted into this massive data collection project, like, how do you feel about it? Are there things that you worry about? Are there things that I worry about? Yeah, I guess I will say one thing is, like, I don't necessarily see this. On the one hand, it is dystopian. Like, it is both, like, first of all, inherently sort of like a cyberpunk image to, like, hold in your mind. And also, like, there's the very obvious thing of, like, people are being paid to do this stuff, which, like, later could be used to automate the work that they're doing. On the other hand, like, I don't necessarily see it, like, it's not guaranteed to me that it will go that way. Like, you know, like, economists, like, have different points of view on, like, how automation from AI will go. In the short term, this money means a lot to a lot of people. And, like, I can imagine a world where, like, nobody has to scrub a toilet again, and, like, people, in fact, transition to doing jobs they're happier with. But, like— I don't know. That, like, really depends on, like, what economic transitions look like, which is, like, something that we have, like, really no understanding of right now. Right. And, you know, to me, the question looming in the background is like, well, what, like, is there any natural cap on the sorts of jobs that robots will be able to do in 10 or 15 or 20 years? And if that comes to pass, how will we feel about the fact that it was, you know, a bunch of human beings who created their own replacements, you know, for an extra two dollars an hour? Yeah. I think there's, like, there's this one anecdote from the Bloomberg piece that really stuck out to me, which is they're talking to one guy who's a factory worker, and they're like, What will you do if the AI takes your job? And he's like, Well, I'll go back to, like, farming on my family's field. No AI can do that. Where, you know, it's like he thinks about it because, like, farming is such a, like, human activity in some way as, like, irreplaceable. But, like, there's no actual theoretical, like, limit on this. I was in Nashville recently and was being driven to the airport, and they'd recently introduced Waymo to Nashville, and I was asking my driver how he felt about it, and he said he wasn't threatened because he mostly drives these really big, like, motor coaches, like basically, like, big buses. And he was like, you know, No, no Waymo can do that. And I was kind of like, Well, yeah, give it a couple of years. Well, as always, a lot to think about, and I do want to thank you in advance for wearing the camera I'm sending you so that you can record all of your actions as you write in Platformer, because I think I can make some pretty decent money off of selling that. I'll require at least $30 per hour. American data is expensive. That seems fair. All right, well, thank you, Ella, for great work on this and last season. And when we come back, my conversation with Clara Shih. Jira by Atlassian is where agent speed meets team intelligence. Assign any work item to your favorite coding agent and see exactly what it's doing without leaving your flow. When something stalls, you're not digging through logs. You can see what's running, unblock what's stuck, and stay in control at scale. Don't let your agents start cold. The Teamwork graph feeds them context from across your entire stack, delivering 44% more accurate results with 48% less token usage. So the work they pick up is the right work done right the first time. Try it free at jira.dev. That's J-I-R-A dot D-E-V. My guest today is Clara Shih, and she has a really great perspective on AI and jobs. She was a founding product marketer on Salesforce's AppExchange in 2006, wrote a Facebook app for salespeople as a side project, and turned it into HereSay Systems, a social media for financial advisors company she ran for a decade and which Yext bought for $125 million in 2024. Along the way, she wrote a book called The Facebook Era and, at 29, joined the board of Starbucks. Then she went back to Salesforce, first to run Service Cloud, its customer service business, and in 2023, as CEO of Salesforce AI, where she led the launch of Agentforce, the autonomous agent platform that Mark Benioff credits with enabling him to cut his support staff from 9,000 people to about 5,000. In November 2024, she left for Meta, where she built and ran a new business AI group, the agents that answer a business's customers on WhatsApp, Messenger, and Instagram, which Meta is now rolling out to businesses of all sizes. And then something interesting happened. She said that last fall, she watched Meta's agents match and even surpass some of her top employees across multiple tasks, and it radicalized her. In April, at the Time 100 Summit, she announced the new work foundation, a 501(c)(3) with Andrew Yang as a founding advisor, and its consumer brand, Dear CC. That's CC as in Clara and also as in career coach. It publishes a podcast where hiring managers explain what they actually want now, a data tool called Field Report that shows you your major's AI exposure, an open-source job search agent called JobClaw, and as of about 10 days ago, a free mentoring and community app called Game Plan. All of it is free, and she remains a senior advisor to Meta. You know, most of the builders that we've had on this show, Aaron Levy, Boris Terni, Matt Garman, told us that jobs would be fine, but that they would just change a lot. The Economist, like Catherine Ann Edwards, Molly Kinder, told us that those changes were the problem. Clara is a builder who has come around to the economist's point of view, and she's trying to do something about it. But how much can a nonprofit really accomplish here? And who else needs to be part of the solution? These are my questions. Let's find out. Here's my conversation with Clara Shai. Clara Shai, welcome to Platformer. Thanks for having me. So you've said that last fall, when you were still at Meta, you watched AI agents match and then beat some of your best people on real tasks, and that you felt radicalized in that moment when you saw it working. Can you take us into that room with you? What was the task, and what did you see? Sure. I mean, it feels like just yesterday, but this is happening. You know, it started off in our product design and our product development was just this traditional process of coming up with an idea, doing user research, surveying, interviewing users, and then coming up with mockups, and then translating that into a product requirements doc, and then passing that off to a front-end engineer, and a back-end engineer, and an ML engineer. We just saw all those steps collapse into one or two people being able to ideate in a room, generate the prototype with vibe coding, test it with real users as well as simulated users, and then have a much leaner team of people be able to build that into production. And just seeing is believing, and in that moment, I just knew, I just imagined this amplifying across the economy, and just we're in for a big ride. I mean, that must have been a crazy moment for you. You've worked in this industry for many years now, and you're describing the traditional product design process, which does have a lot of different hands on it. And what you're saying is, seemingly overnight, it was as if that entire stack could be handled by a person or two. Where did your mind go from there? What did you start to think this would mean both for the company you were at and the broader economy? Well, as you start seeing this pattern, and of course, you know, at a place like Meta, you're under extreme pressure to deliver, you start thinking about how can I apply this to other areas, other bottlenecks, other business processes to help us go faster. So marketing, you know, distribution, privacy policy. Of course, you've got humans in the loop, experts reviewing the final output, but you're just able to collapse what previously took 10 steps and 10 days into a matter of minutes. And as we started doing this, it just, it couldn't stop thinking about what this would mean for if other companies do this and you multiply this across an entire economy with many companies, many industries. It's not that this isn't going to be a great place that we land, but what does that transition look like for the people whose jobs are now radically transformed? Totally. And this is really the question that we've been trying to get at in this series is like, what is the nature of this transition? What is the nature of this transition that we are about to see, and what should people do about it? Interestingly, I have found that a number of people that I've spoken to think that the transition is going to be fairly gentle. They'll say things like, Well, you know, your job is going to change, but, you know, it's still going to be the job that you know and love. You saw something much more radical. I imagine you also talk to people who are a little bit more skeptical. What do you feel like they're not seeing, and what do you tell them? I guess there's nothing like actually seeing AI agents deployed, because otherwise it's all theory. And I think, you know, academics—I love academics. My brother is a professor. We know that these economic reports, every transition, whether it was the Industrial Revolution or it was internet e-commerce, economic reports lag the actual on-the-ground innovation that's taking place. So, I mean, take for example, if you were, you know, if you saw the first factory machine being put in place, or you were part of—you saw the first thousand e-commerce transactions, maybe you were an early Amazon employee, those numbers wouldn't have shown up in the economic reports for years, if not decades. But if you see that happening and you see those transactions coming, you can think about, okay, well, what if maybe not all transactions move online, but a certain percentage of them, that would be very transformational. And so I think the same thing is happening here, is it depends on who you ask, and it's very hard to understand the early part of these exponential curves when you're on the outside of companies and when you're not actually working with people performing actual tasks. So that's why it's really important. I always encourage people like my brother and other economists to go and watch how people work and watch how they deploy these AI agents, because that'll tell you much more than any type of aggregate statistic you might be able to pull at this point. Yeah, the data lag is very real. I get frustrated when I will read these reports about, like, the current state of AI automation, and it was a study that was, like, just published but was conducted, like, a year ago using GPT-3.5, you know? And it's like, this may not actually tell us everything that we need to know about what is going to happen. Another question in this vein is some people I've talked to have said a variation of, Well, Casey, it's easy to automate a task, but it's hard to automate a job. So maybe you can use AI to generate that slide deck now, or you can generate it to send some emails, but you're still going to need that person to navigate the organization, use their critical thinking skills. They can sort of figure out how to route a request. What do you make of that argument as a reason that people don't need to take AI job disruption that seriously? The challenge with any type of new disruption, whether it's AI or it's previous paradigm shifts, is that the impact is never uniform. And so there's actually been great research done on this specific topic by David Autor and his lab at MIT, and they basically proved that the impact on a job depends on what percentage of the tasks in a job get automated and also which specific tasks. So there's basically three cases of what can happen. Case one is if a significant percentage of the tasks get automated, like in translation work, like in customer service, then you're going to see some direct substitution, just displacement happening. Not to 100% of the jobs, but to enough where it starts to no longer be a great job to go into or to stay in. Case two, and this is the happy case, is that what gets automated are the routine tasks. So this is what optimists and, you know, I want to believe this, and I think that this will happen to certain jobs where previous bottleneck tasks like, I don't know, like code snippet generation and test case generation, those get automated. And then what ends up happening is that the people who are in the role, they really do get freed to do higher order tasks. And then you've got, you know, Jevons paradox as well. And then you can see some really great outcomes, such as what happened to software engineers the last two decades. I think it'll continue, actually, for software engineers as well as radiologists and other types of roles. Case three is kind of the not great case along the lines of case one. But case three is that there are more jobs that are created, but because what the AI automates is the expert tasks, the barrier to entry to getting that job drops very suddenly. And this is what happened to taxi drivers with the advent of GPS driving directions, which is the first AI disruption that took place 10 years ago, five or 10 years ago. And you see the labor market for driving, the combination of GPS driving directions plus ride share platforms like Uber taking off, they just flood the labor supply with lots of drivers. And you do have Jevons paradox, so you've got more demand for rides, but then at a certain point. Demand starts to plateau, but you still have labor supply coming in. So you do end up creating more jobs, but those jobs are no longer quality jobs. In fact, in metropolitan areas like New York and London, being a driver now, you're making less than a living wage. I think that's a really great example and something that a lot of people are thinking about. I want to get to what you're doing with the New Work Foundation and Dear CC, but I want to start by hearing a little bit about your own experience, because you've spoken about your own struggles getting entry-level work. And in this case that you just described, as the bar rises for expertise in the economy, this is something that we just expect more and more young people in particular are going to go through. So after Stanford, you've said that you were rejected by every bank and consulting firm you applied to. Happily, this did lead you to getting a job at Google, which you called one of the luckiest breaks of your career. But I'm curious, how did your own initial struggles to find work inform this new project that you've set out on? Well, I have a lot of passion for young people, and for me, I'm a self-made person, and I paid my own way through college. And not just that, I was expected, and I've always felt very proud about helping support my parents. I needed to make money for myself, but also to take care of my family. And so I tried to go after the jobs that a lot of people do when they graduate from a place like Stanford, and it was hard to get rejected. And so I think I have a lot of empathy for the millions of college grads in America alone today who are going through that same struggle. Well, so let's talk about those grads. Dear CC's homepage says that this is the worst entry-level job market in 37 years. I want to hear a little bit about why you think that is. I have talked to economists on this show that say actually things were worse during the Great Financial Crisis, for example. But the New York Fed has recent grads at 5.6% unemployed and 42% underemployed. That doesn't sound great. And Stanford and ADP say that employment for 22- to 25-year-olds in the most AI-exposed jobs is shrinking around 4% a year. So how much of what we're seeing so far do you think is AI, and how much might be attributable to other causes? Well, there's the macro numbers. And again, we talked about how those are lagging indicators, not leading indicators. I'll just say from my own lived experience, building and leading thousand-person teams at Salesforce and at Meta, and now being able to do a tremendous amount with a very lean team with the New York Foundation, that a lot of those tasks that traditionally you would farm out to a recent grad, it's just so much easier and faster. I mean, put the cost aside, it's easier and faster to set up an AI agent to do it. And so we have to consciously encourage, incentivize companies, and company leaders have to consciously decide, okay, they want to optimize for the long term of their talent pipeline to do something that in the short term today really doesn't make sense economically. You mentioned your work at Salesforce and Meta in this regard, and you helped to build, I think, a couple of products that are relevant here. Agentforce, you know, Salesforce CEO Mark Benioff recently said let the company go from around 9,000 customer support staff to 5,000 people. And Meta's Business AI is a customer service agent that is deployed across WhatsApp and other products. And I think the idea there is that it might reduce those businesses' need for staff. Customer service and support can be classic first jobs, entry-level jobs. When those products were getting built, was there an idea of, hey, like this will enable businesses to hire fewer people? Was that part of the conversation, or was that not on y'all's minds? I would say it was in the backs of our minds, but we were so excited about building this. And I think at the time, a lot of people—I'll just speak for myself—I believed that in this happy case that we'll build these products, and people who work in customer support, they'll use Agent Force, they'll use our AI products to automate the mundane, you know, the rote tasks, and then that'll free themselves for complex problem solving and relationship building. I think that has a little bit been true, but primarily not been true. I see. Well, tell us a little bit about what led you to start this new nonprofit and what you guys are up to. Well, so I had this radicalization moment last fall, and then on the personal front, I lost my dad. He spent the last few months of his life living in my home, and it was just a really intense time of personal reflection, which led to professional reflection. And I realized that, you know, we are living amidst this incredible societal change. I mean, just put companies aside. It's not just work life, not just how we spend time, not just our jobs, but every aspect of our personal lives and of society that's being changed. And I just started to ask the bigger question of, what is my purpose, and is there more I can do to help the large numbers of people who are already being affected by AI displacement and who will about to be? Yeah. Well, how do you want to help them? What do you think can be done for them right now? Well, the short answer is no one knows. No one has a crystal ball on exactly how this plays out. And of course, young people are not the only people who are being affected by AI. You asked me earlier whether I think this is a product of, you know, zero interest rates or macro factors. Certainly those are our non-zero contributors, but I think there's been enough studies now isolating those variables that show that AI is directly responsible for a lot of this job loss. And as I mentioned, I myself took down entry-level job postings because I was able to use AI and because I was under so much pressure to deliver fast. And so I can see this playing out. So what can be done? No one knows for sure, and that's why we have to bring humility and a beginner's mind and a desire and a willingness to experiment. And that's really what the New Work Foundation is about. Dear CC is our Gen Z platform that provides content, provides real-time data about what's happening in the market, and provides a few AI tools for people to use as part of their job search journey. And we treat everything like an experiment. So we'll put out certain content, long form, short form, see what works, invest more in what works, and set up that AI improvement loop. And then similarly on the product side. And so on the product side, we probably prototyped a dozen different things, and we've landed on two that we really want to invest in right now called Field Report, which is the market stats about what's happening to different college majors, what's happening to different jobs that typically people go into right out of school. And then separately, we have something called Game Plan, gameplan.dearcc.org, where anyone can go and sign up, put in their LinkedIn or their resume to get a snapshot on where they are today, put in a job posting that they recently got rejected from. And then we kind of do a gap analysis, and we try to explain to them, okay, here's your game plan on how to close the gap. And then we're going to match you into a crew of peers with a mentor so that you can walk through your eight- to 12-week game plan week by week with a community that's supporting you. And all of this is free for folks who are interested, is that right? That's right. We're a 501(c)(3) nonprofit. That's super cool. You know, here in Silicon Valley, I don't talk to a lot of people who are offering things for free unless there's some sort of, you know, advertising component to it. So a breath of fresh air here. You know, it strikes me that, like, we're early enough in this transition that a lot of the work to be done is just noticing the problem. And I saw that your first podcast episodes that you did as part of the nonprofit were about marketing and accounting and software engineering. I also saw this field report tool that you mentioned, which flags the legal occupation as having a very high automation risk, despite there being lots of open roles, which I can imagine coming as a surprise to entry-level workers or young people who are thinking about law school. But I bring all that up because the subject that everyone talks about is software engineering and the automation of coding. And I wonder, like, based on what you've seen in some of these other fields so far, if we might be too focused on coding. Like, are we at risk of missing maybe some other signs where we're already starting to see disruption? Yes, and. So I actually don't think that software engineers are at risk, because I think that the mindset that you get training in computer science—and I'm biased—but I really think that that mindset of thinking in algorithms, thinking in repeatable, you know, repeat-use modularity, I think that's exactly the skill set that is needed to be able to set up systems of agents and to evaluate their outputs. I actually think that the opposite, which is that there's going to be many more software engineers than even currently the rosiest predictions, and that software engineers will take over these other jobs. Legal, marketing, accounting. And so that's my hot take. Yeah, I've heard a version of this from Amjad Masad at Replit, who sort of said that the big growth job of the future is builder, right? Somebody who can just kind of come into an organization and make stuff. And I can imagine, you know, that being relevant to that thing that you just said. Well, and the reason is, there's a very important reason. The reason is because no task, no job is just a job anymore. You're no longer just building a marketing campaign for today or coming up with a legal brief for this specific case. Now you're bringing in this continuous improvement, self-reinforcement learning lens. reinforcement learning lens to these tasks to say, okay, I'm going to build a campaign today, and I have these near-term outcomes for this campaign. But what I'm really doing is building a knowledge base and a learning loop for all of my future campaigns. And so it's that kind of second-order optimization that people who have studied computer science or think that way or can learn to think that way have a real advantage. I'm curious how you think this winds up changing the shape of organizations and what they hire for. I've been asking a version of this to basically everyone that I've spoken to. And on one end, there are founders like Eugenia Kuyda at Wabi, who basically told me she's only hiring for star athletes now. Like, the people that she wants at her relatively small startup are, like, just absolutely elite. On the other hand, I talked to Matt Garman, CEO of AWS. He said replacing junior employees with AI is one of the dumbest things he's ever heard. He thinks there's so much advantage into bringing them into the workplace. You've run very large teams inside two of the biggest software companies in the world. What do you think is going to happen on this question? Yeah, so I have a few predictions. One is, I think that people who work in jobs—so about one in five roles today within companies are preparing some sort of artifact for someone else in the company to look at. It could be preparing a brief. It could be drafting a slide deck. It could be, you know, whatever it is, like coming up with, you know, an order form that a salesperson then delivers to the customer. And so I think those roles are especially Going to be challenged just because the end person who's ultimately accountable for that outcome and ultimately externally facing, whether it's someone dealing with a customer, so a salesperson, or a recruiter working with an external candidate, or a senior legal person who's interfacing with regulators, that person increasingly will find it easier and faster to use AI than coordinate across multiple kind of these input and output types of roles. And so I think that's going to be one area where we're challenged. Two is on the junior side, because in order to be effective at using AI, you need to be able to have enough domain experience and domain expertise to provide context to get the best response, but then also to be able to review and critique and refine what the AI comes back. What you don't want are really junior people who have no idea what looks good, what looks bad, and you're basically giving something to them that then they just delegate to AI, but then they're just passing the AI response back to you, in which case, you know, that's not really adding value. And so that's something we're really focused on with Dear CC, is helping young people start to get that hands-on experience working with AI, even in just one domain, so that they understand the art and science of both the prompting and agent setup, as well as on the eval and review side on the other end. It seems really important to me because the technology is still new enough in the workplace that even the middle managers and some of the senior executives are still struggling to understand what are effective ways to use this. If you're just out of college and you've been handed, I don't know, an agent or a box that you're supposed to type into to build something, I can imagine that being really overwhelming. I know it's relatively early in this new project, but I would love to hear about some of the early experiences that you've seen among young people who you're training to use these tools and then what they're doing with them. Yeah, I mean, there's really different stages that people are at. And I'd say that the first stage we want to tackle is there are a lot of very disillusioned Gen Z grads today. I mean, they were promised a bill of goods, and these people did everything that they thought they were supposed to, and now they're finding themselves, many with college debt to pay off, unable to find the kind of job that they went to college for. And so I think, I mean, we see this with all the booing at commencement speeches from Eric Schmidt and from others. And so the first thing we want to do is address this mindset and this information asymmetry. There are dynamics that those of us who work in AI understand that it's important, I believe, for every young person and every person in the country to understand what is happening. So name the problem. And then I think that a lot of people, once they understand, once they have a mental model of what's going on, they're going to make smart decisions. I think the second step after that is once people have the right mindset, it's the skill set. And of course, this is challenging because it's a moving target. But there are foundational skills that kind of build on each other, kind of like with math. Before you learn algebra and calculus, you got to learn subtraction, addition, multiplication, division. So same with AI. There's kind of foundational building blocks. And so helping people wade through the hundreds of thousands of AI courses that are out there, some paid, some unpaid, to figure out, okay, which makes sense for them, for their specific field, for their specific role or industry that they're interested in. And then what we've learned from that is, and this is why the community part is so important in what we do, is when we lean into AI. You know, the temptation is to say, okay, we don't need any humans. We can just create an AI agent as a mentor. No, that is not what people want. The best motivation, the best mentorship comes from conversations like this, in person, over video, building trust, building relationships. And that's what we're doing, is we're connecting people because these young people feel really lonely and isolated in their search. We found in our surveys that by month six of being unemployed, people start to question their own self-worth, their own sense of identity. You know, many of these young people, they've been high achievers their whole lives, and all of a sudden there's this rude awakening, and they have parents putting pressure on them, asking them why they're, you know, doing gig work, not understanding that there's been this broader shift in the job market. And so we want to put them with people like them who can both provide support but also be accountability partners as they go through each of the steps in their game plan: updating their LinkedIn profile, updating their resume, knowing which parts to use AI to spruce up versus what not to overuse AI and create spam for these job applications, teaching them how to network, teaching them how to have a conversation and reach out to a human hiring manager so that they can bypass the AI screening that so many companies have put in place. And so those are the steps that are so important that we're getting really positive feedback on. Yeah, I mean, I think that's great. And it's honestly striking to me how colleges don't seem to be doing a lot of this work. I mean, I'm sure, you know, all of them have career centers, and I'm sure there's a lot of good work that's being done in this regard. But you just described a lot of steps that I think now probably just should be part of the college education process and maybe art, at least in some places. I agree. And we're partnering with a handful of colleges. I hope we can work with more. I mean, here's the thing. There's still not that many people who are very deep into AI and agents. And the people who have that skill set, they're generally working at a startup or in mid-career Meta, Salesforce. And so I think that's also really unique, is that we've assembled a cast of really A-plus volunteers who work at those companies, and they see what's happening, but they're deciding to give their time to help address this issue in a way that scales. There's this dynamic here that I imagine you've thought about, which is that you're helping young people use AI tools to get into the workforce and do a good job, while at the same time, the managers, executives in that workforce are also using AI tools, and in part they're using them in an effort to have to not hire those people, right, or hire fewer of them. And so it can feel like there is this kind of arms race dynamic. And I wonder if that ever feels completely exhausting to you. I would say anyone who works in AI today feels completely exhausted. Yeah. And yet there's also so much good that has happened and so much opportunity. And so I think that's why we need to have these honest conversations and saying, you know, extreme blanket statements like generalizations like, Oh, AI is going to replace all jobs, or, you know, AI is going to enable this unbelievable wealth for everyone, without going into the specifics. That just doesn't do anyone any favors. And so we're only going to get to this if we can align on what the issues are today, align on the importance of experimentation, measurement, taking a more agile approach, and that we're able to do this across stakeholders, not just big tech, not just startups, not just investors, but also city, state, local, federal government, bringing in academia, kind of weaving all of this together, because it's happening so fast and we're just not used to collaborating at this speed. Right. Your site's numbers say that 71% of hiring managers now prioritize AI skills over seniority. What do AI skills mean to the average manager in 2026? Is it a real thing, or is it just this year's equivalent of, like, proficient in Microsoft Office? Like, what are they actually testing for? What they're testing for, I think in this economic environment, everyone is feeling pressure, as I did, as I have for the last few years, to deliver more, better, faster. And so that's what it is, is can you use AI not just to run a pilot, not just to check the box, to say that you know how to do ChatGPT prompting 101, but can you actually change a business process? Take something that used to cross 15 different teams and take three months to ship your MVP to now doing that over a weekend with multiple agents running that stuff. Yeah, got it. This year Gallup put out a survey saying that 31% of Gen Z is angry about AI, including for some of the reasons we've been talking about today. You've said you hope that that flips to hopeful, maybe in as soon as a year. What do you think it would take to make that happen and make people feel hopeful about the outcomes here? I think if we start having more honest dialogue about where the challenges are and where also the opportunities are, and if we can start to evolve higher education, as you mentioned earlier, right now so many colleges and universities are taking, again, blanket approaches. They're banning AI or telling students that they can use it in only very limited ways because they're worried about cheating. So I would flip the question and say, how do we put guardrails to prevent or minimize cheating, but also make sure that people are graduating with the skills that they need to get hired in this economy? It's a question that I struggle with, I think just because it's moving so quickly. You know, like it strikes me that there are some nonprofits where, like, maybe you want to help people who have a certain kind of cancer, and it's like the problem is very well defined and your job is to raise money to fund research into cancer. When you're trying to help people get jobs in an environment where the requirements for jobs seem to be changing very rapidly, and also there's this end state, at least in the minds of some founders, where, like, they just don't want to maybe hire many people at all. That seems really tough to design around. So I know you've said that you're basically just approaching this with a lot of humility and trying to ask a lot of questions, which I think is the right place to start. I wonder if you've seen things that have made you hopeful that we will find an equilibrium here, that, like, even as AI continues to improve, there will still be a path forward for the college graduate. I do. I would say I am an optimist, and it doesn't sound like that based on some of the things I've shared, but I'm an optimist. But I'm a conditional optimist. I don't think that if we do nothing, it's not going to be great. It's not going to suddenly, the transition isn't going to suddenly be easy for people who find themselves laid off or unable to be hired in the first place. And so I think the ending is yet to be written. And what gives me optimism is people like you, others who are speaking truth, asking hard questions, not just representing a corporate line around what sounds good and making blanket statements, but really getting into the specifics of how could this be good? How could this be challenging? How do we make sure that people have the right mental models so that they can make smart decisions? And if we do that, I do feel very optimistic. Yeah. You're approaching this problem from sort of like the level of the individual person and like what they can do. And I love having those conversations because they're empowering. They give people real ideas for things that they can try. I also believe that we're probably going to need solutions at the government level here, right? And it can't just be left up to every individual to find a path through. So I'm curious if you've thought about that either just for your personal views, if this is a place where you might want to take the nonprofit, if there are policies out there that you like, wage insurance, universal basic income, or if there are things that you can imagine yourself lobbying for. Yeah, so, I mean, absolutely. Organizations like New Work Foundation are not going to solve this on their own, because we have to get every stakeholder group. Government has to take action, because there's policy actions that are required. There are things that employers have to do. Maybe government can incentivize employers to do certain things. There are things that higher ed and K-12 have to do. And then there's things that individuals have to do based on personal agency and personal action. So first, you know, New Work, one of my co-founders is Andrew Yang. He's our founding advisor, and so he, of course, has a lot of strong opinions about policy actions. I actually disagree with him. I don't know. I don't think that universal basic income on its own will address the gap that's widening. And the reason, and of course, it's Maslow's hierarchy of needs. So first, you have to make sure that people have a living wage. But I think that, you know, as humans, we need autonomy, mastery, and purpose. And for 250 years in our society, and for thousands of years before that in Western society, that has come from work. And so if work is going to start to change in such dramatic ways, bringing people along means giving them a living wage, but also giving them a sense of purpose. And so I don't know exactly what that looks like, but what gives me hope is a project that's been running in my hometown for the last couple of decades. It was put in place by a nonprofit, but also my friend's dad, who is the Also, my friend's dad, who is the longtime mayor of Arlington Heights, Illinois, and the group is called Connections to Care, and they basically connect recent retirees and other volunteers in the community with older people in their 80s and 90s to take them to doctor and dentist appointments and to the grocery store, these different tasks. And, you know, these are tasks that you could hire someone to do. You could Instacart or DoorDash, but it's just so much more meaningful when you connect people to each other and create that sense of shared community and purpose. And so, you know, in this future AGI world, whatever that looks like, that could be an answer. Maybe we can find inspiration from not just national service that so many people have talked about, but also local service. If you look around us, there's all kinds of unmet needs in local schools, nursing homes, and our neighbors that if we could play a coordination role, that could be very interesting. I like that a lot. So this is the last episode of this mini-series, Clara, and so I thought I wanted to try to tell you what I think I have learned about jobs and the economy over these past episodes that we've done. And then you can just tell me if you think I've got it basically right, what I'm missing, or if you disagree with me. How does that sound? I am happy to share my point of view. Okay. So this is what I think I have seen. The aggregate numbers about AI-related job disruption still look mostly okay. We are not living in a crisis yet. At the same time, you can look at studies like the Stanford, Canaries in the Coal Mine work, and it does seem like for some entry-level jobs and for jobs that are considered very exposed to AI, we are starting to see some early pain. And we do seem to have isolated AI as the variable for why we are seeing this. So if that is the case, and you assume that AI capabilities are going to continue to improve, which I do, my assumption is a year from now, we're going to see more people in more pain, and we're going to need to have developed, hopefully by then, an actual coordinated society-level response to a problem that I imagine is going to get worse. Clara, do I sound like a reasonable person, or have I completely lost it? You sound completely reasonable. I mean, we've already seen the trend data from last fall versus, you know, this summer. It is trending in a certain way. And in addition to it getting worse for junior-level employees, I think we're going to start to see, as AI, as agentic model capabilities continue to push the frontier, it'll go from, you know, one year out of school, two years out of school, to people who are doing three or four years out of school. And so it is a moving target. And that's why it's so important to teach and encourage not just a skill set, because a skill set can get you the job today, but the mindset, the first order, the first derivative of continuous learning, of hustle, of being more entrepreneurial, and not just thinking that you'll take a job and you'll just be there and, you know, grow steadily for the next 10 years, but that it's going to look, it's going to be volatile. And preparing yourself for that financially and psychologically, that's very important. For sure. Well, maybe let's end by asking you a question that I'm sure you're getting all the time as you talk to young people who are coming to your nonprofit for help, which is basically just the variation of, What do I do if I'm a sophomore in college right now and I have my eye on law school eventually, but I'm reading on your website, well, it seems like this is really highly exposed to AI. What are you telling that person? Is this a time to rethink their entire future career path? Is it a time to just get comfortable with AI skills and hope that those are enough to carry you into a good job? Or is it something else? I say three things. First, really find what you love. Don't choose a major because your mom told you to or because you see that it makes the most money today. Find something that you really love because then it won't feel like work. To go really deep in it, because this economy rewards deep expertise. Two is you do need to start learning AI skills, not basic ChatGPT 101, Claude 101, but actual serious agentic systems. Understand how context works. Understand how tool calls work. Start to learn how to build evals and actually ship something. Ship a legal agent for someone to—I get parking tickets a lot—to fight their parking tickets. It could be anything. And you'll know it's successful if you show it to someone who is not in your family and they decide to use it. That's the measure of success. And then three is make sure that you invest in people skills and have that community around you, because the next few years are going to be bumpy. We saw this with the manufacturing shock of the '80s and '90s, people suddenly seeing jobs shift and change even though the American economy in aggregate actually improved. So we just have to buckle up, and the best way to do that is to have a financial safety net, but also a social safety net of people who love and care and support us. Well, that is some great advice, and I think young people listening should take it to heart. The nonprofit is the New Work Foundation. The website for Gen Z and other interested parties is Dear CC. Clara, thank you so much. Thanks for having me. Platformer is produced by Lindsey Choo and edited by Fitz Harris at Story and Sound. You can watch this whole episode on YouTube at youtube.com slash Casey Newton. My email is casey@platformer.news, and we'll see you next week. Jira by Atlassian is where your team and your agents work from the same context. Try it free at jira.dev. That's J-I-R-A dot D-E-V.