Transcript: "Stop Thinking Like an Employee. You're a Startup." | Lior Frenkel (Adaptation)

Opening & introduction

Eyal David: [00:00] Hey Lior, how's it going? I don't think I need to introduce you too much — we'll ask a few questions and you'll tell us, but let's just say you mentioned up front that there are topics you know well and topics you'll put a disclaimer on. You've had a community running for years now about the "new world," you've been talking about this for ages, right? Now it's the new world on steroids. I wanted to start exactly where I know what's going through my head and through my friends' heads — around a changing world, around AI replacing all of us in maybe two years or less. We're in tech too, so we feel it even more. But tell me — in the community, in the conversations you have, in the companies you lecture at — what are you hearing?

Lior's community and the new world that became a commodity

Lior Frenkel: [00:54] So first, specifically about the community — I let it fade a few years ago, slowly. Partly because I got tired, but also because at some point I felt it wasn't needed anymore: the concept of the new, changing, fragmenting world of work, people wanting to move from being employees to being independent "slashers" — that whole topic became a commodity. Everyone understands it, everyone talks about it. I see what's happening on LinkedIn — the standard of understanding what's happening in the world of work, who has which interests and how things work, has jumped. Maybe it's because everyone just wants to post on LinkedIn, but I think it's simply a lot of podcasts, a lot of conversations. People understand a lot more today, especially in tech, than a decade ago about this machine called "the world of work." So the group isn't needed anymore — it became the common knowledge, not the cutting edge, but the mainstream. Now, in terms of the companies I work with — I always joke it's always "Apple, Google, all the way to the prison service" — that's my example of the spectrum's edge. Everyone feels the same confusion. The difference that bothers me the most today is the huge gap between what management in organizations — tech and non-tech — believes can be done with AI, versus what the people on the ground — developers, designers, product managers, ops — know can actually be done. CEOs and VPs are closer to LinkedIn posts, where the promise is huge — the actual result is maybe a tenth of that. Everyone's promised 10x, and meanwhile we're at about 1.5x, maybe 2x. And there's this oppressive feeling of expectations that are hard to meet. The standard CEO (forget Zuckerberg, who went back to coding) doesn't code, has no idea — he reads the newspaper, reads Globes, understands "guys," and then asks "guys, where are my tokens?" And when people say we need another person on the team, he says "we're not hiring anymore, do it with tokens" — that's not true everywhere, but in a lot of places it is. In tech there's another element: software is cheap to produce. In big traditional companies (Coca-Cola, Tnuva) it's much more about the CEO's and management's decision.

The gap: what management believes AI does vs. what actually happens to workers

Eyal David: [04:46] In tech there's another interesting element — actually on the personal side, which is maybe why I brought up the community: in your beer conversations, conversations with friends, coffee talks — what do you hear around you? What's going through your own head — where are you in all this?

Fear and anxiety in tech — what you hear in beer conversations

Lior Frenkel: [05:03] There are really two conversations. One — fear and anxiety are there, and I think, unfortunately, they're justified. It's better to talk about fear than anxiety, because anxiety isn't rational, and I think there's a rational reason to be afraid right now. The other conversation is with people who say "I'm broken," who are expected to do everything faster — meanwhile there's tons of slop, tons of tokens spilling everywhere, and in practice things move faster, but not that much faster. There's also a lot of talk about identity loss: "I'm not a developer anymore, so who am I?" I spoke with a guy over 50 at IBM. IBM got an order from above a month or two ago: you're not allowed to write a single line of code without "Bob" (that's their Claude). The first month it didn't work because Bob was bad, and then they folded Claude into Bob and it started working. But still — a 51-year-old guy, and most people that age in tech say "enough, I can coast" — they're in real trouble, because they've lost the patience to learn something new, and they feel rusty. There's another side too: a friend of mine, a team lead and developer, is tired in a different way — we're pushed to use AI at work, and then we come home and watch Netflix, and don't want another "transition" there either. It killed his motivation — the fun of solving a hard problem. There's a chapter in my course, "Management in the Age of AI," about how your employees have become managers — because AI is now the employee, and I, as the employee, am its manager, and need delegation skills, evaluation skills, all the Team Lead capabilities. Now, A — employees don't have those skills, and B — they didn't sign up for that, they didn't come here to be managers. And I'll say one more sentence about the IBM guy that kills me: they were told, unequivocally, that at the end of the month there'd be a check on who used the most tokens, and that's how they'd determine the top performer. In that sense we're back to the industrial revolution — measuring output, not efficiency. How many machines, not how long I sat there.

The IBM story and the order to write everything with "Bob"

Your employees became managers — of the AI

IBM measures by tokens — back to the industrial revolution

Eyal David: [10:00] We all know, speaking of Claude, that planning works better, right? What a senior knows to do is sit for an hour, plan, and then work for 30 hours — while the junior runs straight in and works 60 hours because they didn't plan. So clearly planning is better. So how can it be that they're counting tokens?

Lior Frenkel: [10:29] In the end it's a business, ROI — and also, all those tokens cost money, nobody talks about that, so I think soon it'll cost a lot more, it's subsidized right now. There's a potential that in half a year Claude will work well enough that someone could, with a clear conscience, lay off a lot of employees — but then what stops them from messing it up? It's like the revolution that happened with developers in India and then in Ukraine — the market corrects itself: the moment everyone wants only Ukrainian developers because they're more reliable, the Ukrainians raise their prices. Once adoption is big enough, it corrects itself.

Eyal David: [11:23] Let me jump for a second to another side — there's the fear issue, and there's this really deep entry point that maybe UX felt earliest, and these are hard conversations I've had with friends. Daniel Marcous from April (the US tax company) sat here last week — right before April there was a war, half his people were in reserve duty, and the question was how to even deliver features. He built AI teams — agents with QA, product, dev — and they started with small, measured features, and hit a one-week sprint. But it's important to separate a small company like April, where it's easy and fast to adopt AI, from a company like Monday with 3,000 employees, or Bank Leumi — where it's much harder to adopt, and we actually had an episode about that with Galit. From the career angle, anyone listening needs to keep asking: am I suited to the small-startup story, where everything moves fast, for better and worse, or to the big, stable company — where, on the other hand, the first thing they try to do is become more efficient, and robotics and automation are already inside too. So okay — what do you do when the middle disappears, and now there's inflation, in products and in people, and you don't want to go generic?

Daniel from April: building AI teams under wartime pressure

Small startup vs. big company — which story fits you

"The middle disappeared" — and the big frame: your career is a startup

Lior Frenkel: [15:36] I'll split this into two lines of thought. One — something I've believed for a decade, from having spent a lot of time in startups as a founder: I look at my career as if I am the startup. That means I ask myself the questions I'd ask about a business — who are my customers (my colleagues, my managers, the teams I work with), what are my KPIs, what are my features (my skills, my experience, but also things like having friends in communities that let me bring innovation in fast, great communication skills, the ability to influence without authority). And I also have two kinds of bugs: knowledge gaps — those I just need to close (learn what PLG is, for example); and weaknesses — say, if I'm introverted, no course will turn me into someone assertive on stage. You can strengthen it a bit, but this is who I am. So I look at myself as a startup looking for product-market fit — I'm the product, the market is the company I work at. But two important things: first, the long view — having PMF today doesn't guarantee I'll have it in five years (Google had great PMF with Google Search, and it's weakened since, which is why they kept building Gmail, Ads, and now Gemini). Second, if I'm ambitious, I want to keep growing my PMF all the time — not just do the tasks I was hired for, but also learn to do the tasks my boss would love, and even ones nobody in the org imagined I could do.

Features, bugs, strengths and weaknesses — diagnosing your "product"

Product-market fit for your career — and why you must keep growing it

Eyal David: [21:44] That reminds me of the Lean Startup Canvas, which also has risks in it. When OpenAI came out, Google all gathered in a room and said — how do we analyze this, how do we deal with this risk? That's basically risk management, and it needs to come into your career too.

Risk management: why run a premortem on your career

Lior Frenkel: [22:47] I really like that idea. The significant thing I want to add — unlike a business, people are people, and what we're really talking about here is being stuck.

Eyal David: [23:10] Okay, so let's try to talk about both — let's start with risk management.

Lior Frenkel: [23:17] A good CEO doesn't wait to get hit by the market — she sits a year or two ahead and says, "I feel something starting to weaken, let's launch another business line." She also does premortem exercises, not just retros: before going into a project, ask "six months from now the feature didn't ship, customers didn't adopt it — why?" We're good at doing this for our business, our product, but we don't do it for ourselves, exactly because we're human. Few people are strategic about their own careers — most of the time we repress it, and only wake up once every few years out of FOMO (a friend who switched jobs, someone we knew who got promoted, and suddenly "wait, what's happening with my career?"). I'm 47, and back then there was no internet or podcasts to teach you how to plan a career — you had to work hard to get that information.

Worldview: fear vs. opportunity

Lior Frenkel: [28:03] Luckily, there are a lot more open ears now to talk about building a career strategy. There's a skill called "career building" that we used to wake up to once every three to five years, and now it needs constant development. My book "I Want It All" came out in 2016 and has a chapter called "Here Come the Robots" that surveys the AI revolution — the second episode of my podcast "Popcorn" was about the AI revolution with Uri Alyahav from MLDI. I was too early — like it often happens with entrepreneurs, I saw the picture but had terrible timing. Now the revolution has caught everyone, and at the same time I see the volume of emails I get from parents — "Lior, can you talk to my son, he doesn't want to enroll in university" — there's a kind of doom that you can feel.

Eyal David: [29:22] You talked about career strategy — when you did it, it came from a place of choice. Now it's a strategy from a place of fear. How can you still get out of that and plan something you actually want?

Lior Frenkel: [29:56] There are always two sides — it's a matter of worldview, before anything else. One worldview says "everything's happening to me, I'm scared, how do I get out of this" — another worldview says "how cool, there are tectonic shifts, there are opportunities." An analogy: the Strait of Hormuz closes, oil prices rise — there are the ones who panic (the market's too high, Nvidia will never earn back the investment) and the ones who say "this is an insane opportunity." Entrepreneurs always look at the world as opportunities — they see something changing and start rubbing their hands together. Same thing with building a career: you can put on the glasses called "they're paying you to learn" — you're at a company, there are changes (AI or not), your bosses also don't know what to do but they say "just try your best," and if it doesn't work, that's fine — "we didn't know if it would work, go mess up again." It's a school, and even if there's fear, there's also a fun side to it.

"They're paying me to learn" — experimenting inside the job

Eyal David: [34:25] So there's inflation and a ton of noise, and maybe that's exactly what pushes you into this place of looking at the opportunities.

Lior Frenkel: [34:34] Let's take a spoonful of positivity — there are a lot of opportunities beyond robotics and space: the whole physical world. I heard Snapchat's CEO talking about reconnecting to physical experiences. Anthropic posted an "Events Manager" job at $400,000 a year — a role that in tech is usually given to whoever gets the lowest salary in the office. That statement, to me, reinforces what you said — the physical world is also the best place, say, for marketing. Another thing: inside your job you can start experimenting (with coordination, of course) — and another thing, instead of sitting on the couch watching Netflix, go do side projects. I've supported this for years — I have a little digital booklet called "Just Do It Already." Today my take is that in a lot of jobs you don't need to do the project on the side — you can do the project inside the job. There's a simple model called the Value Gap Framework: what is everyone complaining about, in my immediate team, in the department, across the whole org? And if you can't find it, you can always follow the money — where is money going down the drain. There's a difference between tools that are production-ready and customer-facing (a thousand times more complex, data security, etc.) and internal tools — there's a lot more forgiveness there, you can build something that looks like an Excel sheet, but with a brain behind it. The secret: be a bit more entrepreneurial — identify problems and build good solutions for them, even as an employee. There's a Venn diagram here between "builders" and entrepreneurs.

Anthropic and the $400K job — why the physical world is coming back

Side projects: inside the job and outside it

Value Gap Framework — where to look for unmet value

The story: the WhatsApp game Lior built for his daughter during the war

Lior Frenkel: [43:13] Side projects don't have to be business-related — during the war, I went down to the bomb shelter with my daughter, and I knew she wouldn't have school for a long time. We built a WhatsApp "who knows who" game for her class — kids sign up, a short quiz with a few questions. She lost interest pretty quickly, I stayed excited, and sent it to her class. It was fun because most of my career I was a PM, and suddenly I could build it myself, without a developer — that's the fun of vibe coding. When you measure motivation at work, one of the things that keeps coming up in research is how much you use your abilities — as a PM who knows product marketing, PLG, virality, and is stuck in one narrow corner at work, I feel "underused." In a side project I can use everything, so motivation goes up. The game never went truly viral (I'd be paying Cloudflare $5,000 a day, not $5), but it went pretty far — I discovered our ICP was 6th-grade classes, between 5th and 9th grade. I walked away from that with two things: noise-canceling — go to a place that's easy for you, that interests you, where you're a domain expert; and there's no ego risk there — it's not like promising your colleagues at work that something will work and then failing.

Eyal David: [48:37] The second thing you did is stand out above the noise — VPs gave you feedback on that, I saw it on Twitter or LinkedIn too. It's also a great distribution path: someone inside the industry with a LinkedIn full of experience, versus a 24-year-old junior coming in — it'll also be cheaper to work with them.

What to do if you're a junior right now

Eyal David: [49:47] Want to talk about juniors for a second? What would you suggest to juniors starting out, to stay relevant, when they don't have work experience yet?

Lior Frenkel: [49:58] On the current narrative — "the talented ones will stay, we won't bring in juniors anymore because we don't need to add headcount" — because in the end the experienced person runs Claude, and the junior runs Claude, so why pay them more. Juniors have one big disadvantage and one big advantage. The advantage — Unlearning: a product manager who's been in the role for ten to fifteen years has habits at every level, that they now need to question. A lot of product work is "passing the baton" — alignment between design and developers, and AI can replace a chunk of that, so anyone who's only passing along information is at higher risk. A junior comes in as a blank slate — like the difference between Wonderful, which was built AI-first and built agents from scratch, and Google, which has to build this on top of old infrastructure and culture. I remember myself: when I moved from a hardware/software company to a startup, I kept writing 300-page PRDs, and the whole team waited — it took me a few days to realize this habit was stupid, that there weren't 400 developers to align, and a whiteboard and three hours would do. Now the disadvantage for juniors — the middle has disappeared: they don't yet have "taste" — the ability to decide what's good and what's bad. There's a writer at This American Life who explains the Taste Gap: you know what something good should look like, but you don't yet know how to make it — the only way is to do it again and again, until after 20 years you get there. Juniors haven't developed that yet, and that connects to the most important trait today — judgment, precision. If I ask the same question to Claude Opus, GPT-5.5, and a few other models, they'll all give me different answers, and each one is basically "the average" — and you need experience to know that the real answer is sometimes 180 degrees from what you got.

Unlearning — juniors' secret edge over the experienced

Taste Gap — what juniors still lack, and why it's the most needed skill today

Eyal David: [59:14] I sort of connect this to pilot training, when you were talking about it — what do you do in pilot training? You go to the simulator. So kind of — start creating things, build up experience, and arrive with a more accurate "pilot," which helps with exactly this precision work that businesses need today.

Lior Frenkel: [1:00:00] Exactly — juniors need to do things that also go out into the world, because that meeting with real customers is where the learning happens.

Eyal David: [1:00:22] I want to steal a few more minutes to talk about something personal for every listener — finding what you're good at and understanding your strategy. That's hard — how would you go about it?

The trick: your frustration is the compass to your next skill

Lior Frenkel: [1:00:57] I'll give you one trick: there's something in psychology that speaks strongly about desires — things we carry inside that we're not really open to. Desires can be "I want to be famous," "I want to be rich," "I want to build a product that reaches millions of people." When I choose a skill to learn, there are two options: one — say "what does the market want" and stick with it, which takes a lot of willpower; the psychologically easier option — learn a skill connected to your desire. The simple way to find it: notice when you leave a meeting frustrated — "I had something really smart to say and I didn't say it, because everyone was more senior than me" — that means there's a skill called Executive Presence worth learning. Or if you're frustrated that you can't convince your team — maybe the skill is Influence or Communication. Frustration is a place with a lot of energy in it, which keeps driving you to learn the skill over the long run. That's the Product-Market Fit of your career — the Venn diagram between a skill worth investing in and where your motivation actually is. The cliché I love: ask yourself, in the last month, what did you curse about in traffic on the way home — "why don't I," "why don't they" — and it's worth documenting that, in a Google Doc that collects your frustrations, and measuring yourself against it every so often.

Closing, how to reach Lior

Eyal David: [1:07:13] Listen, this was fun — we didn't get to talk about a ton of things I wanted to. Write to us and we'll try to bring Lior back. Lior, it was a pleasure — if people want to get in touch with you?

Lior Frenkel: [1:07:35] LinkedIn, and you'll find me easily on Google — I make a lot of noise. I'll do one small plug: a lot of what we talked about here is in my course, "My Next Step," which is for employees who want to understand how to move forward at work. And if you have no idea what you want to be in the world, there's my book "I Want It All." But other than that, you don't need to buy anything — I write a lot on LinkedIn, feel free to learn from there for free.

Eyal David: [1:08:34] And that's it, how fun. Alright, thank you so much, bye bye.