Stanford AI Expert: 71% of People Won't Survive the AI Shift — Here's the 30-Minute Fix

1. 0:00 – 0:44

   Intro

   1. KKKian Katanforoosh0:00

      How often do you use AI? If it's not daily, I think you're generally behind.

   2. MMMarina Mogilko0:04

      That's Kian Katanforoosh, Stanford AI professor who built one of the world's top AI education platforms with Andrew Ng. Now through his company, he's tested over a million people on their AI skills, and today he has a step-by-step plan so you don't fall behind. What are the three moves everyone should make in 2026?

   3. KKKian Katanforoosh0:24

      Learn the foundations of AI. Assess yourself to make sure you're ready. Build the habits of learning. If you focus on one thing for a day, you probably are already in the top X percent of the world. For a week nonstop, you're in the top 10%. Focus on a month, you're in the top 1%, but to be in the top 0.1%, you will have to.

   4. MMMarina Mogilko0:43

      Hello,

2. 0:44 – 1:47

   The big 2026 question

   1. MMMarina Mogilko0:44

      everyone. Welcome back to Silicon Valley Girl and Davos. Kian, let's start with your big idea. 2026 is the year of humans, but also we're getting a completely different narrative, you know, a new model every week, uh, replacing jobs. How do you think we should focus on humans right now, and what is the shift?

   2. KKKian Katanforoosh1:04

      I think the shift that is happening is, uh, broadly due to the fact that people generally overestimate the impact of technology on the short term-

   3. MMMarina Mogilko1:14

      Mm-hmm

   4. KKKian Katanforoosh1:15

      ... and, uh, underestimate what the technology can do on the long term. Um, if you look at all the reports from found- foundation model labs, uh, you know, OpenAI, Anthropic, and others, uh, there's a lot of task-level reports, like AI is good at task A, task B, task C is getting automated. And actually, going from a task to some human's job changing with a job usually being made of hundreds of tasks is not that simple.

   5. MMMarina Mogilko1:43

      Mm-hmm.

   6. KKKian Katanforoosh1:44

      It can take decades, and every... almost every

3. 1:47 – 3:03

   When will drivers disappear?

   1. KKKian Katanforoosh1:47

      prediction that I've seen since the launch of ChatGPT of X, Y, Z job is going away has not happened.

   2. MMMarina Mogilko1:53

      Mm-hmm.

   3. KKKian Katanforoosh1:54

      You know, the famous one is the radiologists will go away and the drivers will go away, and then you see this meme of radiologists driving to work in their, [laughs] in their car.

   4. MMMarina Mogilko2:04

      You mentioned drivers. Uh, do you have an estimate, for example? 'Cause if you go in San Francisco, it's almost... Waymos are almost everywhere.

   5. KKKian Katanforoosh2:11

      Yeah.

   6. MMMarina Mogilko2:11

      I don't really see [laughs] old, older taxis. Uh, so we see the replacement happening, but how soon do you think it's gonna happen for drivers, for example?

   7. KKKian Katanforoosh2:20

      Yeah. Well, you look, like, you know, the, the rise of, you know, Waymo, Cruise, um, all these companies in the self-driving space, you know, really started in 2014, 2015. So we're already 11 years-

   8. MMMarina Mogilko2:35

      Yeah

   9. KKKian Katanforoosh2:35

      ... into them having hired tons of engineers to build that problem. So even autonomous driving has been a decade of full-on research with people working so hard. So, you know, why wouldn't it be the same for the rest? I, I think, like, maybe in the next decade we're gonna start, you know, seeing less voice actors, less, uh, translators, uh, maybe customer support is going to completely change. I agree fully with that. I just think people thought it will happen within six months-

   10. MMMarina Mogilko3:00

       Yeah

   11. KKKian Katanforoosh3:00

       ... and it hasn't.

   12. MMMarina Mogilko3:01

       Yeah. So we're safe

4. 3:03 – 4:33

   You use AI the wrong way

   1. MMMarina Mogilko3:03

      for now, at least, like, for the next five years.

   2. KKKian Katanforoosh3:05

      Yeah. Generally, I think s- safe in, in a career comes down to learning velocity. It turns down to can you reinvent yourself?

   3. MMMarina Mogilko3:12

      Yeah.

   4. KKKian Katanforoosh3:13

      You know, the, the WEF has this, uh, metric called the half-life of skill that is going down, meaning on average a skill is not useful that long. It's two years in tech or AI, uh, and so you have to refresh yourself, and that's what makes you safe ultimately.

   5. MMMarina Mogilko3:26

      Absolutely. And, and your data shows 71% of people misjudge their AI skill level. Can you give us some benchmarks? So what is an AI proficient person? What does it... his day-to-day look like? Does he start with, like, chatting with, uh, his AI, or is it, you know, not writing emails by yourself [laughs] or having AI manage your schedule?

   6. KKKian Katanforoosh3:46

      Yeah. I try to separate adoption of AI and proficiency. Uh, so I'll give you an example. Adoption is, is like, uh, you use AI every day, and I use it every week. You're a better adopter than I am.

   7. MMMarina Mogilko3:58

      Mm.

   8. KKKian Katanforoosh3:59

      But turns out that if we watch you prompt engineer and me, uh, maybe your prompts are just simple prompts, and when you look at what I'm doing, I'm doing, um, a variety of techniques. I'm doing zero-shot prompt. I'm doing few-shot prompts. I'm doing a chain of thoughts. I'm doing a prompt chain that is super complex that f- feeds one into another. I'm doing a retrieval augmented generation system that I built. My proficiency is higher than you.

   9. MMMarina Mogilko4:23

      Yeah.

   10. KKKian Katanforoosh4:24

       That's the difference between adoption and proficiency.

   11. MMMarina Mogilko4:26

       Okay. What you just said, uh, makes me feel like I'm an, a beginner [laughs]

   12. KKKian Katanforoosh4:30

       [laughs]

   13. MMMarina Mogilko4:30

       ... 'cause my prompts are really, really simple. Okay.

5. 4:33 – 5:10

   90 days to real proficiency

   1. MMMarina Mogilko4:33

      If I want to sound like you in 90 days, what should I be doing?

   2. KKKian Katanforoosh4:36

      So first, if you have 90 days, I would say, at first we need to establish the foundations. You take a few foundational classes. We can recommend some on deeplearning.ai, on other platforms. There's a lot of content out there, honestly, um, high quality. Establish the foundation. Uh, you will get to a point where what will matter the most in AI, because the market is moving so fast, is that you're plugged in the network. So what I recommend generally is you go to X, you go to Reddit, you go to some of the machine learning, uh, popular newsletters, and you register to all of these.

   3. MMMarina Mogilko5:09

      Can you recommend-

6. 5:10 – 6:01

   Who to follow in AI

   1. KKKian Katanforoosh5:10

      What's-

   2. MMMarina Mogilko5:10

      Like, who do you, who do you follow on X for best advice?

   3. KKKian Katanforoosh5:13

      Uh, well, a- actually if you go on my X and you look at who I follow, you can-

   4. MMMarina Mogilko5:17

      Mm. Okay. Best. Yeah

   5. KKKian Katanforoosh5:17

      ... follow the same people. But you know, some of them are here, like Andrew Ng is a great person to follow.

   6. MMMarina Mogilko5:22

      Yes.

   7. KKKian Katanforoosh5:22

      Great newsletter called The Batch. Uh, Richard Socher, you know, Yoshua Bengio, a lot of great AI scientists that, you know, people trust.

   8. MMMarina Mogilko5:30

      Yeah.

   9. KKKian Katanforoosh5:30

      Um, and actually it allows you to cut through the noise when there's so much noise coming. Like I tell you, when I was in grad school, um, we would read a lot of people that come up in archive, the website that, uh, you know, where papers are often published. Today there's just so much that you have to find ways to differentiate signal from noise.

   10. MMMarina Mogilko5:47

       Yeah. Every time I scroll through my feed on Instagram, there's this new app, and th- this company just changed the game in this market, and, like, happens every day.

   11. KKKian Katanforoosh5:56

       Yeah.

   12. MMMarina Mogilko5:57

       Uh, so okay, we established this. I follow the right people. What is the

7. 6:01 – 7:20

   Top AI tools + the bottleneck

   1. MMMarina Mogilko6:01

      next step? Are there, like, top three AI apps that I should be using?

   2. KKKian Katanforoosh6:05

      Yeah, I mean, you know, I recommend, uh, obviously Workera for testing yourself, uh, although it's mostly used in corporations. Other than that, you know, deeplearning.ai has a lot of free content out there. It's really good. You also find, uh, that the LLMs can help you learn, like you can actually prompt the LLMs, but the bottleneck is people don't know what to ask the LLM, and that's where the assessment is so important because at some point you're gonna be pretty good at AI, and you're gonna sort of have a wall in front of yourself, like, "What do I do next? Am I actually that good? Do I know?" You know, like, to give you an example, at Stanford we have, as you said, uh, the class on campus with a lot of students, and we have the class on YouTube ver- same content published on YouTube with a lot of view- like a lot of views. And those students would tell you that the difference between them and the Stanford, uh, uh, uh, kids is that, um, it's not the material, it's that they don't know if... how good they are. The Stanford, uh, students, they have friends at OpenAI, they have friends at Meta, they have friends at Google, they have friends... They know how good they are compared to the bar, how, how much does it take to get a job there? But if you're somewhere in the world with no ecosystem, you're not plugged in, it's really hard. And so that's where the assessment is so important. It can tell you, "Hey, actually, you, you thought you, you were very pretty good, but that's not the bar. The bar is actually higher."

   3. MMMarina Mogilko7:19

      Top

8. 7:20 – 9:50

   3 questions reveal your AI level

   1. MMMarina Mogilko7:20

      three questions that you should ask yourself to kind of understand your level.

   2. KKKian Katanforoosh7:23

      How often do you use AI?

   3. MMMarina Mogilko7:25

      Mm-hmm.

   4. KKKian Katanforoosh7:25

      You know, if it's not daily, I think you're generally behind-

   5. MMMarina Mogilko7:28

      Mm-hmm

   6. KKKian Katanforoosh7:28

      ... right now.

   7. MMMarina Mogilko7:29

      Wow.

   8. KKKian Katanforoosh7:29

      Uh, that's the simple one. Uh, the other one is, you know, think about 10 products that use AI that, that, that you encounter in your daily life. Can you come up with 10 products, you know? And some people would realize, like, "Actually, I don't realize where is AI. Is it here? Is it there? I, I don't know. I, I don't have this ability to, like, identify AI."

   9. MMMarina Mogilko7:51

      Mm-hmm.

   10. KKKian Katanforoosh7:51

       You're probably behind.

   11. MMMarina Mogilko7:53

       When Kian says that, a lot of people have the same reaction. "Okay, by that definition, I'm definitely not using AI enough yet." And honestly, for most teams, it's not a motivation problem, it's that there is no simple visual way to plug AI into the work they're already doing. Right now, AI usually lives in fragments, Cursor in one tab, Cloud Code in another, maybe a copilot or an OpenAI model somewhere else. That's exactly how my team used to work too before we changed our setup, ideas in one place, guidelines in another, code somewhere else, video editing on another platform, and almost no visibility into how all of it connects. That's why I got excited about partnering with Miro and their MCP Server. MCP lets you connect your Miro canvas directly to the AI coding tools you already use. So instead of Miro being notes on the side, it becomes the central hub where your specs, diagrams, and context actually feed your agentic coding workflows. Practically, it changes two big things. First, you can take shared context, diagrams, docs, notes, system maps, and send that straight into your AI assistants to build better code because now Claude or Cursor has actual context from your diagrams and specs, not just a prompt. Second, you can instantly visualize that code as diagrams in Miro in a collaborative environment where the whole team can understand, comment on, and iterate together without digging through a repo. If you've been wanting to use AI more seriously at work but it's always felt abstract, fragmented, or messy, Miro's MCP makes it concrete and collaborative in a way that finally clicks. If you wanna try it yourself, check out the link in the description and the MCP tutorials on Miro's YouTube channel. And now back to my conversation with Kian. If I

9. 9:50 – 10:52

   Make AI 10x more useful

   1. MMMarina Mogilko9:50

      want to start using AI for work, what questions should I be asking myself?

   2. KKKian Katanforoosh9:54

      I think when it comes to work, a lot of the value of, uh, language models is in the context. So for example, on ChatGPT, there is, um, this, uh, feature that allows you to give custom instructions to the model. So, "Hi, my name is Kian. I'm XYZ. I like to speak in English or in whatever language, and I like to be concise," or, "I like to..." You know, whatever your style is. Um, that's example of context that you give to the LLM.

   3. MMMarina Mogilko10:23

      It's like memory, right?

   4. KKKian Katanforoosh10:24

      Yeah, memory that you give to the LLM. Um, although, yeah, me- memory is slightly different than context. I can, I can explain after.

   5. MMMarina Mogilko10:30

      Mm.

   6. KKKian Katanforoosh10:30

      But the, um... You know, and at work, you sort of want your documents to be accessible to your LLM if possible. You want your custom instructions to be accessible. You even want the custom instruction of your coworkers so that when you talk about your coworkers or you're trying to send an email to XYZ, it will figure it out. So the value of the LLM increases with the amount of context it has access to at work.

10. 10:52 – 14:16

    How Workera cut human approvals

    1. MMMarina Mogilko10:52

       Is that how proficient organizations use AI?

    2. KKKian Katanforoosh10:55

       Yeah. I give you a concrete example.

    3. MMMarina Mogilko10:56

       Mm-hmm.

    4. KKKian Katanforoosh10:57

       So, uh, at, at Workera, we, we are a big, uh, Anthropic shop internally. We use a lot of clouds. Um, all our engineers are on this version of Claude called Claude Code Max, uh, which, um, is very powerful to code. And across the company, we have s- things that we call skills, Anthropic calls skills, where, where you can think of them as files that define a certain way of doing a certain thing. Like, here is how we recruit at Workera, or here is our brand guidelines. This is the font we use. This is how we speech. These are the color palettes that you can use. Before, if an engineer wanted to build a website, they would have to call the marketing team at the end and say, "Can you review the font? Can you review the alignment? Can you review XYZ?" Today, because it's all coded, you don't need anymore to talk to a human. The engineer just asks the LLM, "Can you just verify that the copywriting is correct, the color palette is right?" And they know that the marketing team has maintained that code.

    5. MMMarina Mogilko11:55

       I love that.

    6. KKKian Katanforoosh11:55

       And so it cuts communication, and it's very powerful. You gain actually so much speed and, uh, create so much more time for the marketing team to think about, uh, do we need to change our fonts? Do we need... Rather than like-... every day talk to an engineer and say, "No, change that font. Change that font."

    7. MMMarina Mogilko12:12

       Yeah

    8. KKKian Katanforoosh12:12

       So.

    9. MMMarina Mogilko12:13

       Do you check the result afterwards? Like, oh-

    10. KKKian Katanforoosh12:15

        Yeah, the engineer does.

    11. MMMarina Mogilko12:16

        Yeah. Okay.

    12. KKKian Katanforoosh12:16

        Yeah, the engineers do.

    13. MMMarina Mogilko12:17

        Wow. So now I'm very curious about your day-to-day as a founder. What has changed in the past three years in how you just deal with your coworkers? So you mentioned using Claude, uh, that cuts communication. What else?

    14. KKKian Katanforoosh12:31

        I would say one thing that has changed is, uh, we're, we are getting flatter as an organization, which means we have, uh, for example, our head of AI decided to become an IC, an individual contributor, from a manager role, and that didn't used to happen before. And he's doing great as an individual contributor, and he feels more productive, and he feels like he's back close to the machine. Um, and I think that's a trend that we're gonna see a lot. Um, the second aspect is, so in tech you have this ratio of within a perfect team, do you... how many engineers do you have? How many product managers do you have? How many product designers? Historically, you would have, I don't know, some... Jeff Bezos calls it the two-pizza team. The team has to be able to eat two pizzas.

    15. MMMarina Mogilko13:15

        [laughs]

    16. KKKian Katanforoosh13:15

        Uh, if it's more than two pizzas, the team is too big, basically.

    17. MMMarina Mogilko13:18

        Oh, this has grown beyond that. [laughs]

    18. KKKian Katanforoosh13:19

        Uh, yeah. And, and so right now, uh, I think historically we've had, I don't know, eight engineers, one product manager, one product designer. I think now it's getting way more efficient on the engineering side, where you can actually probably put a team together with two engineers, one product manager, one product designer, and the engineers are very empowered to perform, um, to build everything on their own almost with some input from, you know, the other parties. And so we are seeing at Workera a lot of smaller teams, a lot of, you know, instead of having three big teams, we might have six, seven smaller teams that have more ownership of their surface area. We have, uh, you know, transcriptions of meetings, which is really helpful because I can remember, you know, what was the context. You know, we use our own product in our interviewing, so there's an AI interviewer.

    19. MMMarina Mogilko14:07

        Oh, wow.

    20. KKKian Katanforoosh14:07

        I think we just make all these tools accessible to our workforce, uh, and we make sure they adopt it very frequently.

    21. MMMarina Mogilko14:15

        Who does your

11. 14:16 – 14:45

    His AI briefs him every morning

    1. MMMarina Mogilko14:16

       calendar? Is it AI now?

    2. KKKian Katanforoosh14:18

       Every morning I have a briefing that my... So my assistant builds, uh, AI systems herself. Uh, and d- um, she has a mo- a little, you know, agent call it or workflow that, uh, tracks my calendar and, uh, tracks what I know or what past conversations I've had. And every morning I get a briefing automatically in Slack that tells me that this is where you need to be-

    3. MMMarina Mogilko14:40

       Mm

    4. KKKian Katanforoosh14:40

       ... and this is what you need to know-

    5. MMMarina Mogilko14:41

       Nice

    6. KKKian Katanforoosh14:42

       ... pretty much, which is really helpful, you know? Um-

    7. MMMarina Mogilko14:44

       Yeah. Everything

12. 14:45 – 15:25

    Hire AI talent or start yourself?

    1. MMMarina Mogilko14:45

       that you described, if I want the same in my company, do you think I need to hire someone who's more AI native, or my team can just handle it?

    2. KKKian Katanforoosh14:51

       No, you should not.

    3. MMMarina Mogilko14:51

       And we're all, like, creatives and-

    4. KKKian Katanforoosh14:53

       Yeah, I think you should start yourself.

    5. MMMarina Mogilko14:55

       Mm.

    6. KKKian Katanforoosh14:55

       It all starts, uh, by yourself.

    7. MMMarina Mogilko14:58

       Mm-hmm.

    8. KKKian Katanforoosh14:58

       So I think you should try it yourself, and, uh, you will actually figure out that you can get a lot done by yourself. Uh, and you're already very proficient, so it will be easier probably for you. If you wanna get in the technical realm, yeah, you will need someone more technical. You need someone who has coded in the past, who knows, like, you can get a lot more done with those people.

    9. MMMarina Mogilko15:16

       But the, but the basics like connecting documents and... We should have done that-

    10. KKKian Katanforoosh15:20

        Yeah

    11. MMMarina Mogilko15:20

        ... months ago

    12. KKKian Katanforoosh15:20

        I think it's more about agency

    13. MMMarina Mogilko15:21

        Now you're like-

    14. KKKian Katanforoosh15:22

        It's, it's having agency to do that.

    15. MMMarina Mogilko15:24

        And that's

13. 15:25 – 19:24

    The skill AI can't replace

    1. MMMarina Mogilko15:25

       agency. I'm glad that you mentioned it, 'cause I was thinking a lot being here in Davos, everyone's talking about AI. I was thinking about top three skills that everybody should be developing, and I think you mentioned that in one of your talks, that there's some skills that die out really fast and some skills that just stay with you.

    2. KKKian Katanforoosh15:40

       Yeah.

    3. MMMarina Mogilko15:41

       Uh, they have more longevity. And I think agency is something that, you know, if we imagine AI as this bar, it's already telling some people what to do. Like, they're kind of below AI. Like, if you work in customer support, right? You just prompt something and you read it out loud. Most of us are still beyond this, um, this line because we're s- we're using AI as helper, but this bar is rising. What do you think? And, like, the way to stay beyond it and make AI work for you, not control you, is to have agency, or maybe something else. What do you think?

    4. KKKian Katanforoosh16:13

       Yeah. I mean, I'd say 100% agency is a durable skill. We feel it. Durable as in it will be useful even 10 years from now. It's very important. There's a lot more durable skill. Critical thinking, problem-solving, effective communication. Um, I think AI literacy is a durable skill. People will need it for a long time. Coding I think is a very important durable skill. I-

    5. MMMarina Mogilko16:33

       Mm. Still?

    6. KKKian Katanforoosh16:34

       Yeah, I think so.

    7. MMMarina Mogilko16:34

       Even for someone like me who's a-

    8. KKKian Katanforoosh16:36

       Yeah

    9. MMMarina Mogilko16:37

       ... podcaster.

    10. KKKian Katanforoosh16:37

        I think, I think so. I, I don't think you'll have to learn syntax.

    11. MMMarina Mogilko16:40

        Mm.

    12. KKKian Katanforoosh16:40

        Like, you don't need to know how to code manually, but if you can tell if the coding agent is... what is, what is it doing, you have a significant advantage.

    13. MMMarina Mogilko16:49

        Mm.

    14. KKKian Katanforoosh16:49

        You can catch the errors faster. You can iterate faster. It is hard to negotiate that. And then to come to the top three skills, I think, like, I'd separate in three groups. So for technical folks, very technical folks, like foundational model level, um, right now companies are fighting for talent that can do, uh, reasoning, that can build reasoning loops and reasoning models. There's very few people in the world that can do it, and they're very, very valuable. The second one that's, uh, underrated, forgotten sometimes, is distributed computing. There's not that many people that can build clusters, that can train models on massive clusters. It is very complicated. It requires a combination of math skills, linear algebra, electrical engineering. It's very, very complicated, and those are, you know, hardcore engineers. Very valuable. And then the third one is reinforcement learning. So in AI, when, when you look at a model, it usually goes through different phases of training, like pre-training and post-training. People that have... And, and at some point in the sometimes pre-training, sometimes post-training, there are certain techniques from the world of reinforcement learning. Uh, that's why, uh, the, the idea is, uh, like AlphaGo or, uh, you know, chess, those games that you've seen AI play better, they're based on reinforcement learning methods.

    15. MMMarina Mogilko18:05

        When the machine learns by itself and-

    16. KKKian Katanforoosh18:07

        By itself

    17. MMMarina Mogilko18:07

        ... tries different things.

    18. KKKian Katanforoosh18:08

        It, it learns through experience, not through examples.

    19. MMMarina Mogilko18:10

        Yeah.

    20. KKKian Katanforoosh18:11

        And that skill is also very valuable. So that's the technical tier. In the tier, applied tier, I would say forward deployed engineering is very popular, meaning if you can also do business-And be technical at the same time. That combination is rarer. And then for day-to-day life, I think identifying AI, being able to use it natively is the most popular skill for general, uh, you know, awareness.

    21. MMMarina Mogilko18:33

        Kian just talked about how most people use AI every day, but their prompts are still super basic. Take my example. For months, I was struggling with AI writing. It just didn't sound like me, it used the wrong words, it used the wrong tone, it invented facts, and overall sounded like AI. So I decided to build a system, three files that teach AI your real voice, real facts about your background, and even your phrases you'd never say. And this transformed my entire workflow because I can now write better LinkedIn posts, I can now write better emails, and come up with better ideas. All of these files are free for my newsletter subscribers. There is a link in the description. Go ahead, download those files, and they come with an instruction on how to teach your AI to speak like you. The technology is amazing. Start using it in a proper way. The link is in the description. [whoosh] So do you

14. 19:24 – 21:52

    Real reason Gen Z can't hire

    1. MMMarina Mogilko19:24

       see jobs market going down at all, or w- what's your projection for the next five years?

    2. KKKian Katanforoosh19:30

       So a few things. I, I, I would say, one, people say Gen Zs... There's no job for Gen Zs. We've heard that over the last couple of years. I think last year was definitely the hardest I've seen for university grads.

    3. MMMarina Mogilko19:42

       But was it about AI? 'Cause a lot of people-

    4. KKKian Katanforoosh19:44

       I don't think so.

    5. MMMarina Mogilko19:44

       Yeah.

    6. KKKian Katanforoosh19:45

       I, I think, like-

    7. MMMarina Mogilko19:45

       Over-hiring during COVID or-

    8. KKKian Katanforoosh19:46

       I think companies have over-hired during COVID, and now they're saying AI is automating our stuff because it makes the stock go up. The truth is they're performance managing a lot. They're roster managing, they're exiting people. And, and maybe there's a little bit of that job is not as, uh, important as it used to be, but there's a lot of like we wanna keep our best people, and they hide it behind the AI lingo. You know, why would Meta exit people from their Metaverse team if it was AI, you know? No, it's because they wanted to make more out of that team, and he, he probably thinks they can get a lot more done keeping the best people and getting them to work hard. You know, otherwise, you wouldn't have heard about the Metaverse team exiting people. You would have heard of something else. So I think, uh, I think it's really performance management that is happening. Um, and I think they don't find enough AI native talent. The reason Gen Z has struggled to find jobs in the last year is that there's just not enough AI native talent in the markets.

    9. MMMarina Mogilko20:41

       Mm.

    10. KKKian Katanforoosh20:42

        There are still just pockets that are in hubs, and if you're in the hub, as a Gen Z, actually, you, you, you, you, you can f- do fairly well today. There's good offers, there's good opportunities. When you're outside of the hub, it's very hard. It's much diffi- more difficult. So long story short, what I think is gonna happen is over time, companies are going to figure out how to update their workflows. So yes, you will see productivity go up, and you will see a lot of movement internally. I think we're gonna see more internal mobility than we've ever seen in our life.

    11. MMMarina Mogilko21:11

        Interesting.

    12. KKKian Katanforoosh21:11

        Um, it will be very common for you to start in the marketing team and go to the sales team, start in the sales team and go to the HRVP team. Whenever it, it... you know, you need to move. The- that's... the movement inside the company's gonna grow. The company's total head count I think is gonna decrease. I think on average, companies are gonna be slightly smaller, but it's not gonna be a massive cut. It's gonna be, you know, every year, maybe they don't backfill people who retire. They just don't hire more, you know? Or if someone leaves, they probably try to do a cultural refresh by bringing AI native talent that is, uh, coming out of universities, and at the same time, they invest in their talent to build AI native mindsets inside the company.

    13. MMMarina Mogilko21:51

        Do

15. 21:52 – 24:10

    Stanford lecturer: degrees are worthless

    1. MMMarina Mogilko21:52

       you think university loses its value in the next 10 years?

    2. KKKian Katanforoosh21:57

       Yeah. F- yeah, I think so. I think unless you're a top-tier university where you have a brand defensibility, people don't join for the content, they join for the network, the brands, the being surrounded by people that work hard, that are ambitious. Those will not lose their values. Um, so when you think about the university, you think about a bundle. Like universities have content, mentorship, uh, research, uh, blah, blah, blah, you know, and that bundle will for sure change.

    3. MMMarina Mogilko22:25

       Yeah.

    4. KKKian Katanforoosh22:26

       I think it's just... it's gonna be a different offer. Maybe it's not gonna be four-year bachelor's degree, two years masters. It's gonna chair. I think one of the weaknesses of universities today is the mismatch of the job market skills needed. Like, you have too many universities that s- still teach skills that you will- won't need. You know, I come from France, and I recall when I was a student, we had double the amount of physical educator being trained than the amount of jobs available after they graduate.

    5. MMMarina Mogilko22:52

       Oh.

    6. KKKian Katanforoosh22:52

       You don't want a society that has that. You want a society that has a zero skills gap. At all points, the people that are joining the job market have the exact skills that the market needs. It's not an easy problems, but I think universities could be better at it.

    7. MMMarina Mogilko23:04

       Yeah, and it's really hard for universities to do that, right?

    8. KKKian Katanforoosh23:07

       Super hard.

    9. MMMarina Mogilko23:07

       'Cause you have a program that's established.

    10. KKKian Katanforoosh23:09

        So one model is universities focus on durable skills, and then companies build the capabilities to teach perishable skills.

    11. MMMarina Mogilko23:16

        Mm.

    12. KKKian Katanforoosh23:16

        So for example, the problem is reasoning. Reasoning, the people who know reasoning, they're PhD students from the top AI labs in the world. That's where they come from. So it is coming from universities generally. Ideally, you would want all universities to give you AI native talents. They... Everyone who graduates has amazing AI skills. They're not specialized in a specific area, but they have great durable skills. Join the company, and the company has somehow a stack, an HR and learning stack that can take onboard an employee, and instead of them becoming a partner at a consulting firm in seven years, they become in six months.

    13. MMMarina Mogilko23:50

        Yeah.

    14. KKKian Katanforoosh23:50

        And that would be ideal, I think.

    15. MMMarina Mogilko23:52

        And that's what you do at Workera, right?

    16. KKKian Katanforoosh23:53

        Yeah. We help a lot of companies do that. We do part of this problem, but the general idea is, uh, durable skills taught at school, perishable skills taught at the company.

    17. MMMarina Mogilko24:02

        I love that. This is exactly how universities should be working, right? Not only now, but also like 20 years ago because skills keep changing.

16. 24:10 – 28:30

    Why 95% of AI agents fail

    1. MMMarina Mogilko24:10

       I think in Workera you have AI agents, right, that work in production, and a lot of companies are failing to build those AI agents. Also, we try. Like in my company, we have a media company. We're not that technical, but from what I see, agents sound great, but then in real world it's still like-... a set of steps that they're following, and you still need a lot of human work. Can you tell me why in your company [laughs] they're working and they're not working for a lot of other companies?

    2. KKKian Katanforoosh24:37

       Yeah. Yeah, for sure. I, I think, uh, it is very, very hard to put an agent in production. People don't realize that. A demo is not a production agent. You have-

    3. MMMarina Mogilko24:46

       Yeah

    4. KKKian Katanforoosh24:46

       ... demos are so easy to do now. You see so many of them. If you can tell the difference between a demo and a production system, then you know what... how hard it is. And that's why MIT's, uh, study said only 5% of agents work in production. So I'll give you some examples. The reason I think, uh, so we've done large deployments. One of the companies that is here, Bill McDermott, the CEO of ServiceNow, is here. Uh, ServiceNow uses Workera enterprise-wide, so everybody is being measured, mentored, skills gap identified, and they get sort of an AI driving license, essentially, a certificate for the year. That agent has been deployed very large scale. Um, for this to happen, you need... Y- there's so many things that can go wrong. OpenAI can fail. What do you do? We have a model routing layer that allows us to route immediately to an- the next best model.

    5. MMMarina Mogilko25:36

       Mm.

    6. KKKian Katanforoosh25:37

       Um, translation. People have different languages. It's not as easy as just saying, "Oh, tell... Do the assessment in Japanese." It's not at all as easy. If a Japanese person looks at that, they will say it has a lot of, uh, cultural gaps.

    7. MMMarina Mogilko25:51

       Yeah.

    8. KKKian Katanforoosh25:52

       It is not culturally intelligent, so it's so much hard work in there. The agent has to be connected to the UI, and somehow the agent misses a button. It just doesn't see it, and then you're stuck. Uh, oh, the agent actually scored you very unfairly. Your score should have been 200, and you got 150, and you don't agree with it. Well, we have a feature that allows the person to say, "I think the agent was wrong." And then you send a human expert in the loop that we'd review within four business days and respond to the person, "We've upgr- upbr- upgraded your score, and we've corrected the agent." And when you do that across thousands and thousands of people, well, of course, the agent gets better over time.

    9. MMMarina Mogilko26:29

       Yeah.

    10. KKKian Katanforoosh26:29

        And yeah, the first deployment is a mess. The second one is a little bit less of a mess. And, you know, at some point, you just build that muscle of looking between the lines and in the details because that's what matters. In a lot of cases, we even removed AI. Like, we realized that, you know, we started, we were like, "Everything has to be stochastic," meaning on, you know, sort of non-non-deterministic. And then, you know, we got some feedback, and users said, "No, actually, I really like when part of the experience is deterministic, where I don't need to be real time talking to the AI interviewer, uh, because it stresses me out. I wanna take my pause, and I wanna be able to look at the multiple choice question and take my time to che- check A." That's not, you know, doesn't need, like, agentic AI.

    11. MMMarina Mogilko27:11

        Yeah.

    12. KKKian Katanforoosh27:12

        Um, and so we had to decide where do we do deterministic and where do we do stochastic? Because stochastic allows you to understand the reasoning of the person. You have a live conversation with an agent. You can dig deeper in their thoughts, but it's not always the right solution.

    13. MMMarina Mogilko27:26

        Wow. So from what you're describing, it feels like in order to deploy an AI agent in your company, you need a very technical person who can do the right reasoning and ask... And c- like, pave the right path for that agent.

    14. KKKian Katanforoosh27:37

        I think it's like... So companies now have these agent marketplaces. Like, you can go on, uh, their internal platform and create an agent with a prompt. That is very different than building an agentic company-

    15. MMMarina Mogilko27:49

        Mm

    16. KKKian Katanforoosh27:49

        ... where the bar is just super high. So for example, if you wanna create a bot on Slack that, uh, reads a channel and summarizes it for you every day, you don't need a team that is technical. You, you now can have someone go on the marketplace of agents, hook it, connect it to Slack, and tell it what to do. It will do it. Uh, but w- you know, we're building an AI agent that is supposed to be the best in the world at measuring someone's skills to give them feedback. That's a different problem. You can't get it wrong. The bar is extremely high. And there you need a research team. You need an applied team. You need a product team.

    17. MMMarina Mogilko28:22

        And talking about jobs, I feel like we need more and more people these days because of all of the tools, all of the opportunities that open up.

    18. KKKian Katanforoosh28:29

        Yeah.

    19. MMMarina Mogilko28:29

        But do you think there will

17. 28:30 – 32:04

    Will AI kill entrepreneurship?

    1. MMMarina Mogilko28:30

       be more companies? 'Cause it's easier to start a company now.

    2. KKKian Katanforoosh28:32

       Yeah, I think there will be more companies.

    3. MMMarina Mogilko28:33

       Is it gonna help even out the market?

    4. KKKian Katanforoosh28:36

       Uh, yes, I think so.

    5. MMMarina Mogilko28:37

       So-

    6. KKKian Katanforoosh28:37

       I think there will be more entrepreneurship. There will be more small businesses. You know, last year I saw on, on X some of these vibe coding tools, I'm not gonna say which one, uh, people would know, um, did a marketing campaign saying, "Oh, one of our users rebuilt Calendly and rebuilt DocuSign in six hours."

    7. MMMarina Mogilko28:55

       I saw. Yeah.

    8. KKKian Katanforoosh28:55

       In six hours. Where is that product? Who has used it? Nobody has ever used that product. Nobody has ever seen it. It's probably not even maintained anymore. Because what makes Calendly... And, and DocuSign, by the way, opened a new office in San Francisco, and they're growing, you know? So what's interesting is, um, if you don't have the best product, if you're not significantly better than DocuSign, why would I change to your product? The bar is high. Yes, it's easy to build a simple signature tool or calendar scheduling, but your Calendly is very actually powerful. It has so many features. And so the only way to replace that is if actually you build a product, the product is not only as good, but actually maybe 50% better for, for the cost of switching to be worth it for a user, 50% better, and on top of that, you will have to make sure it keeps being 50% better.

    9. MMMarina Mogilko29:44

       Yeah, you need the right marketing as well.

    10. KKKian Katanforoosh29:46

        And so, so I don't buy this idea that... of personal software. I don't buy that-

    11. MMMarina Mogilko29:49

        Mm

    12. KKKian Katanforoosh29:49

        ... people are gonna build their Calendly, and they're gonna build blah, blah, blah. I think some company will build a Calendly that is 50% better than Calendly that is AI native, and everybody will use that agent. And because you don't wanna... Y- you don't have the time. We don't have the time to-

    13. MMMarina Mogilko30:05

        Yeah

    14. KKKian Katanforoosh30:05

        ... build our personal software-

    15. MMMarina Mogilko30:06

        To build everything, yeah

    16. KKKian Katanforoosh30:06

        ... and maintain it, you know? So I don't know. I think it's just marketing campaigns.

    17. MMMarina Mogilko30:11

        Yes, totally makes sense in the next five years. We just don't know [laughs] what's gonna happen in 10 years when AI is so good, and it just n- gets all the knowledge. Like, I don't know, I'm thinking about the lawyers who are using AI.

    18. KKKian Katanforoosh30:22

        Yeah.

    19. MMMarina Mogilko30:22

        Like, AI, an AI tool has all of the legal knowledge.

    20. KKKian Katanforoosh30:25

        Yeah.

    21. MMMarina Mogilko30:25

        It's just so much better than anything.

    22. KKKian Katanforoosh30:26

        For sure. I agree it will be an AI agentic tool. I just don't think there will be hundreds of them.

    23. MMMarina Mogilko30:31

        Mm.

    24. KKKian Katanforoosh30:31

        I think people will use the best.

    25. MMMarina Mogilko30:33

        Yeah.

    26. KKKian Katanforoosh30:33

        You know?

    27. MMMarina Mogilko30:33

        Yeah, yeah.

    28. KKKian Katanforoosh30:34

        So I don't buy that there will be-

    29. MMMarina Mogilko30:35

        But it, it will be one major company, don't you think?

    30. KKKian Katanforoosh30:38

        It will be, it will be. It, it will be
18. 32:04 – 35:18

    3 moves for 2026. Write down

    1. MMMarina Mogilko32:04

       our audience is 25 to 40 years old. They all want to become better in the age of AI, build something. What are the three moves that they should make in 2026?

    2. KKKian Katanforoosh32:14

       You know, learn the foundations of AI. Assess yourself to make sure you're ready. Build the habits of learning. Like, every f- every day when you wake up, take five minutes, read the X posts of the people that you trust in the space. And it turns out, you know, you, you won't feel better after a week, but you will feel a lot better after a year. You'll feel like you're at the... You- you're probably at the cutting edge. You know, some- some- someone said, I saw, like, you know, if you focus on one thing for a, a day, you probably are already in the top, uh, you know, X percent of the world in that thing. If you focus on it for a week nonstop, you're in the top 10%. Focus on a month, you're in the top 1%. F- But to be in the top 0.1%, you will have to build that habit and follow it for five, 10 years, and you might be the top 0.1% at what you're trying to do.

    3. MMMarina Mogilko33:03

       I love that. I also like your point about joining a hub, 'cause this helps you evaluate yourself-

    4. KKKian Katanforoosh33:09

       Yeah

    5. MMMarina Mogilko33:10

       ... uh, against other people, like compare notes and learn from each other.

    6. KKKian Katanforoosh33:13

       Yeah.

    7. MMMarina Mogilko33:14

       Maybe start locally, and then, you know, change groups.

    8. KKKian Katanforoosh33:17

       Yeah. I think especially if you're early in your career. Today hubs have, uh, significant advantages because, um... So, you know, AI started in the Silicon Valley, uh, pretty much the, the... I guess the, the new wave of AI agents. Um, so companies came, so there was more opportunities, so more people came. Because more people came, more companies came. And now if you're in San Francisco, you don't even need to put an effort to learn what's happening in AI. I go out at the dinner, uh, we talk about voice AI somehow. People talk about Tesla autonomous autopilots. You, you just learn constantly because you're in the hub. I think in the next few years it will be like that. The hubs are way advanced, uh, compared to the rest. But I think that in the next, you know, five, 10 years horizon, you know, people will get slightly older. They would wanna build families. They will leave the hubs, a lot of them. They will take that knowledge with them. They will probably start building somewhere else.

    9. MMMarina Mogilko34:08

       Local hubs. Yeah.

    10. KKKian Katanforoosh34:09

        And exactly. And there, uh, that's what happened in the dotcom when engine- software engineering was concentrated, and a few years later, uh, actually it became democratized because of online learning, because of access to information, but also because a lot of these experts moved elsewhere. And I think the same thing will happen. In 10 years, even outside the hubs you will find great AI native micro hubs or local communities.

    11. MMMarina Mogilko34:30

        Yeah. That's amazing. Thank you so much for this conversation. I love podcasts. When, uh... After the podcast, I'm gonna just text my team. We're gonna build the Claude thing. Uh, we're gonna make sure we have all of the documents, everything synced.

    12. KKKian Katanforoosh34:42

        Good.

    13. MMMarina Mogilko34:43

        Thank you so much.

    14. KKKian Katanforoosh34:43

        [laughs] Thank you for having me.

    15. MMMarina Mogilko34:44

        Love this feeling. [laughs]

    16. KKKian Katanforoosh34:45

        Yeah. Thank you.

    17. MMMarina Mogilko34:45

        Let's get to work.

    18. KKKian Katanforoosh34:46

        It was fun. Yeah.

    19. MMMarina Mogilko34:47

        Thank you. Kian is the reason why I came back home from Davos and started implementing Claude across everything that I do. We set up multiple Claude projects for all the social media that we're running, and honestly, it's been so transformational. Another conversation that's been really transformational was my conversation that I recorded at Davos with Ryan Roslansky. So if you're all about AI, if you're interested what's happening to jobs and how you can get a better job by posting on LinkedIn, watch that episode. It's live on my channel, and I'll see you very soon. Bye.

Episode duration: 35:18