Skip to content
Silicon Valley GirlSilicon Valley Girl

The 3-Year AI Deadline: How to Stay Relevant (2026-2028)

📌 Today's sponsor is MongoDB. Their Atlas Learning Hub has free, hands-on courses and skill badges — spin up a free cluster and earn your first one today 👉 https://fandf.co/4f5t1Hi Daniil and David Liberman are building a decentralized alternative to Big Tech AI — and they came on to answer the question: who gets access to AI, and does it stay open to everyone, or slowly turn into a privilege for a few? We tried to figure that out together. We also got into the one-hour-a-day habit that helps you get through turbulent times, how AI is already changing hands-on jobs like plumbing, and what our near future actually looks like. *Timestamps:* 00:00 Meet the Liberman brothers 01:20 Is AI actually overhyped? 03:22 ChatGPT solved a 3-year house problem 04:48 How much time do we really have? 06:36 The AI moment that hits everyone 07:58 Anthropic's founder owns just 1.5% 10:34 How China's AI game really works 12:48 Why leaving your AI gets hard 14:20 Who really owns 70% of our future 14:50 Two futures. Which one is yours? 17:08 Who controls your answers controls you 18:12 Should the government step in? 21:23 What one person can actually do 24:37 AI that's more empathetic than people 26:48 How Bitcoin beat the untouchable giants 31:54 The coming boom in small businesses 33:00 What to do in 2 years 33:47 Where AI can take you: real story 35:24 What non-coders build in one hour 36:43 Rapid-fire predictions for 5 years 39:53 The smartest move you can make today 🎙 GUESTS — Daniil & David Liberman - Instagram: https://www.instagram.com/libermans - Gonka community on Discord: https://discord.gg/6Tpq98jdD9 Links: 📩 Follow my Newsletter: https://siliconvalleygirl.beehiiv.com/subscribe?utm_source=youtube&utm_medium=video&utm_campaign=futureproof-sub&utm_content=The-Libermans 🔗 My Instagram: https://www.instagram.com/siliconvalleygirl/ 📌 My Companies & Products: https://partnerships.marinamogilko.co

Marina Mogilkohost
Jul 17, 202640mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:001:20

    Meet the Liberman brothers

    1. SP

      What kind of reality we will see in the future mostly depend on whether big tech companies will have control on who has access to artificial intelligence.

    2. MM

      These are David and Daniel Liberman, investors and founders who sold their company to Snap. Now they're building Gonka, a decentralized AI network designed to make AI more open and affordable so anyone can win with it.

    3. SP

      Do you know what is the share of, uh, Dario Amodei, founder of Anthropic, creator of Claude, in his own company?

    4. MM

      What is it?

    5. SP

      1.5%.

    6. MM

      Oh, really?

    7. SP

      People who build it, people who brought their values into this, they are not in control anymore.

    8. MM

      What should people be doing now to thrive in two or three years?

    9. SP

      Everything will be disrupted. Every single job-

    10. MM

      When?

    11. SP

      ... will be disrupted.

    12. MM

      How much time do we have? Hello, everyone. Welcome to Silicon Valley Girl. I have my amazing friends today, Liberman brothers. They're not only friends, who I've known since 2015, they're also my first investors, who believed in me. And the whole round that happened in 2021, which became the first ever venture round in a creator, happened because of you guys. Thank you so much.

    13. SP

      Because of you. [laughs]

    14. MM

      Thank, thank you for believing in me. And, um, yeah, it's fascinating to see how much happened after that.

  2. 1:203:22

    Is AI actually overhyped?

    1. MM

      You've been in technology circles and building stuff for a while, so I really wanna focus on what's going on right now and what's happening with AI. And AI has been the most discussed technological revolution in the past two years. But what I'm noticing, and I'm sure a lot of people who are watching this will agree, that we're talking a lot about AI, and obviously, it's changing, like, for coders, because a lot of people generate 80% of their co- code, maybe more. But when you take an ordinary person, yes, the search is better now, the productivity's maybe up, but there hasn't been a life-changing event for humanity yet. So do you think AI is over-hyped?

    2. SP

      I would start, start with, like, co- contradicting to what you said. Well, life changed 'cause people started trusting one source of information, like, in the chat. Like, "What is this? What is that?" And they don't check multiple sources as they used to with the internet. And this is something pivotal, honestly, already. Therefore, like, we will se- have some consequences about that already from this point. When was the last time you were actually doing your research in Google? Like, not through the ChatGPT or other platform with agents doing all the swarming research for you and stuff. Like, this changed. We somehow started trusting this, uh, technology, which not always is correct.

    3. MM

      Yeah.

    4. SP

      What kind of reality we will see in the future mostly depend on whether someone will be able to get access to artificial intelligence. If artificial intelligence will be freely accessible, uh, by every human on Earth, if, uh, it will become actually commodity, when it's, like, cheap, uh-

    5. MM

      Yeah

    6. SP

      ... a- as cheap as possible, then each of us for sure will use AI to increase our productivity. But definitely, another scenario is possible, if big tech firms will be able to gate the access and decide who can have the access or not.

  3. 3:224:48

    ChatGPT solved a 3-year house problem

    1. SP

      I'll give you an example. Someone recently asked us about, like, what about the plumbers? Will they be actually replaced with the robots or, like, with AI? What, what's gonna happen there?

    2. MM

      Yeah.

    3. SP

      How, how can AI actually make innovation in there? Uh, in our house when we mo- where we moved three years ago, for this past three years we had the problem with, uh, hot water, and we were complaining to the landlord. He called for the plumber several times. They came there. He's like, "Oh, you need to replace the entire heating system. It's gonna be very expensive." And then at some point, we basically pointed the ChatGPT through the camera into the piping system and said, like, "What's wrong here?" And, and the chat said, like, "Looks like it's missing one small part here, like a pipe which will actually close a loop here," and blah, blah, blah, blah. So we started discussing this, w- this problem with the plumbers. After three years-

    4. MM

      Oh, my God

    5. SP

      ... we now have instant hot water-

    6. MM

      Oh, my G-

    7. SP

      ... with just one tiny-

    8. MM

      But you still needed phys- physical labor, yes?

    9. SP

      Absolutely.

    10. MM

      Mm-hmm.

    11. SP

      But the, the, imagine that it's not, it's not us who should convince the plumber that they should do this job, but the AI in their own earpiece, which is like, "Looks like there's this-"

    12. MM

      Yeah

    13. SP

      ... "stuff missing there." And they fixed everything. And the first time, when we called them three years ago-

    14. MM

      Yeah

    15. SP

      ... and they were like, "Oh, best guys ever."

    16. MM

      Mm-hmm.

    17. SP

      "Five stars. You fixed the problem."

    18. MM

      Yeah.

    19. SP

      "Hooray." Instead, we were suffering for three years before actually AI came and-

    20. MM

      Do you see

  4. 4:486:36

    How much time do we really have?

    1. MM

      any jobs completely being disrupted when we don't need those people at all?

    2. SP

      Like, I mean, drivers.

    3. MM

      Oh.

    4. SP

      Like, the cars will be driving themselves, but the people inside of the car can still be there.

    5. MM

      Yeah.

    6. SP

      I mean, like, for services, safety, whatever.

    7. SP

      This is, um, actually a good example of the industry, uh, which, uh, had a chance to-

    8. SP

      Adapt

    9. SP

      ... gradually adapt. Uh, if all the drivers will see the future, the future will be, uh-

    10. SP

      Definitely self-driving

    11. SP

      ... self-driving.

    12. MM

      Mm-hmm.

    13. SP

      Uh, and if they would, uh, invest in joining forces together, in, uh, building the future together. Instead, w- when people, they're reluctant to participate in the innovation-

    14. SP

      In the change

    15. SP

      ... and, uh, when they, um, use innovations in the passive, uh, form, when they, they just, uh, consumers. They, they don't understand that, uh, they actually can also contribute. Eventually, 15 years later, their profession can be disrupted.

    16. MM

      There is no such job. Mm-hmm.

    17. SP

      So actually, for every profession, there will be enough time Uh, to adapt.

    18. MM

      Mm.

    19. SP

      If, if you actually see that the future is coming. Like, if you-

    20. MM

      Mm.

    21. SP

      Like, clearly accept.

    22. MM

      Yeah.

    23. SP

      Everything will be disrupted. Every single job-

    24. MM

      When?

    25. SP

      ... will be disrupted.

    26. MM

      How much time do we have?

    27. SP

      It depends on what the profession.

    28. SP

      For some profession, it-

    29. SP

      Like, for some profession, it's like next year. For some profession, it's, uh, maybe in half a year from now. For some professions, it, it will take five to 10 years because the robots are first gonna be in the factories. Then it's gonna be there, and then it's gonna be there. Uh, a specific professions, uh, will require more time for the physical world to adapt.

    30. MM

      Yeah.

  5. 6:367:58

    The AI moment that hits everyone

    1. SP

      because it's software.

    2. MM

      Do you think in the next two or three years we're gonna see an event that is directed by AI that's gonna change lives for everyone? Because again, for example, going to, I don't know, my mom's example, right? She uses AI to search. Her life hasn't really been affected with AI. What do you think the timeline is when everyone experiences something COVID-like, when they're like, "Oh my God"?

    3. SP

      Honestly, in the next two years, probably we'll experience several events like that. We'll see a lot of breakthroughs, like, a, a, an eye-opening breakthroughs in biotech and, like understanding genome and a lot of chemistry as well, uh, new materials. Like, we will see a lot, mind-blowing, something which we considering possible before events within the next two years. In the software engineering itself, I think that we will see the entire operating systems and coding languages to be rewritten and, like written from the scratch, something which we're actually changing everything we understood before.

    4. MM

      Mm-hmm.

    5. SP

      It will happen likely in the next couple years.

    6. SP

      For sure, in the coming year or so, uh, we will, uh, get through, uh, several tipping points where there is no way back. It will be really hard to, uh, influence, uh, where this progress, uh, will lead us. Um-

    7. MM

      Like what? Do you have any examples?

    8. SP

      So for example,

  6. 7:5810:34

    Anthropic's founder owns just 1.5%

    1. SP

      uh, whether big tech companies will have, uh, control on who has access.

    2. MM

      Mm.

    3. SP

      Um, so now it still can be changed and, uh, consumers can affect that. Uh, consumers can demand rules.

    4. SP

      The question is, like-

    5. MM

      On price

    6. SP

      ... whether, whether m- market defines the access to the technology or a handful of people.

    7. MM

      But if they all become publicly traded, then it's up for stakeholders.

    8. SP

      This is the most interesting conversation. Who are the stakeholders?

    9. SP

      Um, yeah, so one of the subjects which is not really discussed right now is that, you're right, all of these companies, they are-

    10. SP

      Publicly traded or will be

    11. SP

      ... they will become publicly traded. Who will control these companies then? Uh, for example, do you know what is the share of, uh, Dario Amodei, founder, uh, of-

    12. SP

      Anthropic

    13. SP

      ... Anthropic, creator of Claude?

    14. SP

      In his own company?

    15. SP

      In his own company, what is the share?

    16. MM

      What is it? Uh, I know Google owns a lot. [laughs]

    17. SP

      1.5%.

    18. MM

      Oh, really?

    19. SP

      [laughs]

    20. SP

      So when you understand that it's just 1.5%, you understand that, uh-

    21. SP

      People who-

    22. SP

      ... those people even-

    23. SP

      ... people who build it, people w- who, who brought their values into this, they are not in control anymore.

    24. SP

      Uh, who then in control?

    25. MM

      Who is? [laughs]

    26. SP

      [laughs]

    27. SP

      Uh, then you, you find some, uh, it called, uh, uh, long-term trust or something like that.

    28. MM

      Mm.

    29. SP

      And you see people in this trust, and you won't recognize anyone on this list.

    30. MM

      Let me pause for a second. Outside this podcast, I spend a lot of my time around early AI startups, and almost every founder I talk to is building the same kind of thing, an app with AI inside it. And almost every one of them hits the same wall. Their product data lives in one database. The second they add an AI feature, they spin up a separate vector database just for the embeddings. Now, it's two systems to sync, two systems to scale, and the AI answers from a stale copy of the data instead of what's actually happening in the app. That's exactly what MongoDB Atlas, today's sponsor, is built for. Atlas keeps your operational data and your embeddings together in one system. You run semantic search and build RAG apps without ever leaving MongoDB. Because everything lives in one place, your app answers from real-time data, not an old copy, and it scales as you grow. And you don't have to figure this out alone. MongoDB built the Atlas Learning Hub, free hands-on courses and skill badges, like Vector Search Fundamentals, about an hour each. The link's in the description. Spin up a free cluster and earn your first badge today. Big thanks to MongoDB for sponsoring this video. Now, let's get back to our conversation. Why would those people

  7. 10:3412:48

    How China's AI game really works

    1. MM

      who control those organization decide to stop or pause access?

    2. SP

      Because if you create artificial scarcity, you can increase price for, for as long as, uh, you want.

    3. MM

      But then an open model, China releases an open model, and people can-

    4. SP

      But China is not releasing open models. Chinese doing the, the, the old corporate trick, except they're doing this on, like on the government level. They lure you with a semi-open sourceness, because none of their model is actually open source. They're open weight, so no one can repeat. And then when the next version is released, they can easily not open weight it just like that.

    5. SP

      It's, it's actually-

    6. SP

      And no one will be able to even continue or do anything with this.

    7. SP

      It's not theoretical.

    8. MM

      Mm.

    9. SP

      Uh, you can see how most of the US-based companies, they used to have open models.

    10. MM

      Mm-hmm.

    11. SP

      Now, not anymore.

    12. MM

      Now, none. Yeah, it's so interesting when you see, like the-

    13. SP

      The Chinese models start to be not open from the v- very first day of the model release. They put it on the market for a month or two or three, only accessible from their own servers. And then they still play with the open weight, uh, uh, game. Because what they play, they want for as many people in the world as possible to use their models to actually create the competition to the, uh, uh, Am- American models, and then basically lure everyone to use their servers. Not open weight on your own server, but our servers in China with Chinese data centers and Chinese chips.

    14. SP

      For any Tech organization which sells digital products. The first thing that they need to acquire users. As soon as they acquire users, then, uh, they usually change the rules. Remember, like in 2003, 2004, still Google had, uh-

    15. MM

      Ads

    16. SP

      ... a pro- promise that ads will never be in the main feed.

    17. MM

      Mm-hmm.

    18. SP

      Ads only-

    19. MM

      It's-

    20. SP

      ... will be on the side.

    21. MM

      But they're still at the top, and they're highlighted ads.

    22. SP

      But-

    23. MM

      So they're not in the search.

    24. SP

      No, but, but it's-

    25. MM

      But, like, on top of the search.

    26. SP

      But, but the promise was that it w-

    27. MM

      Oh, that it would never

    28. SP

      ... it will never be in the main feed. Then the rules changed, and, uh, that's what we see in many cases.

  8. 12:4814:20

    Why leaving your AI gets hard

    1. MM

      But because it's so... It's kind of easy, not very easy, but it's kind of easy to move from one model to another. We still have competition on the market. We have Google, Claude, ChatGPT, right? And from my own experience, I switch maybe like every two weeks. [laughs]

    2. SP

      [laughs]

    3. MM

      Uh, yes, with now learn turn, long-term memory that goes across all your chats, it's becoming harder, but if one model decides to 10X their price, I can easily switch. Do you think it's not gonna be the case?

    4. SP

      But look at then what is also happening in parallel with this, uh, luxury of changing, uh, between models. Tools which used to be independent for coding, like Cursor or, or Figma for design, those tools are being steadily integrated into the main product or similar product of the frontier models themself.

    5. MM

      Yeah, they're becoming the feeders for your chat window. Yeah.

    6. SP

      Everything you use, from the calendar to the email to documentation, to picture generation, to music listening, to, like, buying products, as soon as all of this is already in one ecosystem, that's when it's gonna be much, much comp-

    7. MM

      Oh, but you can-

    8. SP

      ... much more complicated for you-

    9. MM

      You can connect your Google Calendar to another

    10. SP

      ... to change between the models just because all of your ecosystem will be in one. So from, from this perspective, uh, still, uh, you're right that, uh, existence of all of these, uh, independent companies, it helps.

    11. MM

      Yeah, they prevent-

    12. SP

      It helps to-

    13. MM

      ... authoritarian regimes inside the AI companies. Paint me

  9. 14:2014:50

    Who really owns 70% of our future

    1. MM

      this worst case scenario. Is it one company owning-

    2. SP

      No, it, it do- it doesn't have to be one company

    3. MM

      ... or it could, or three company coordinating their efforts.

    4. SP

      Or three company coordinating their efforts. Like Google, Microsoft, and Amazon today combined is 65 to 70% of global compute, cloud compute.

    5. MM

      So they ba- basically own 70% of our future.

    6. SP

      When we're talking about AI-

    7. MM

      It's everything

    8. SP

      ... now it's everything. It will disrupt every single industry over the course of next five years. Let me give you, uh, like

  10. 14:5017:08

    Two futures. Which one is yours?

    1. SP

      two versions o- of the future. Uh, in each version, three of us increase productivity four times. In one version of the future, this productivity increased because of this, uh-

    2. MM

      Three corporations

    3. SP

      ... three companies, three corporations. What they can easily do is to get the entire-

    4. MM

      Markets

    5. SP

      ... roles of this, uh, margins by, uh, increasing the prices of their-

    6. MM

      Of their services

    7. SP

      ... uh, ecosystems. If your productivity, my productivity depends on whether we have access to those models or not, they can set any price.

    8. MM

      Yeah.

    9. SP

      Uh, and that's what we experienced recently. Models haven't become cheaper. Opposite to that, there is version of the future, uh, when access guaranteed, and we have, uh, it as cheap as possible. In this case, each of us become four times more productive. We produce four times more goods.

    10. MM

      And the only question is gonna be-

    11. SP

      And with, with the four times wealthier

    12. MM

      ... whether we need that much or we will work less. Oh, we adapt to new stuff. [laughs]

    13. SP

      We will adapt, and we will have more and more and more. Like, it's just, like, details of every single product will be more details, better quality, all of that. Maybe, like, our houses will be bigger. It's gonna be easier and faster to build them.

    14. MM

      Or we'll have more inside our houses, like with cars.

    15. SP

      We will have more stuff inside our houses. Our house finally will become smart houses. The fridge which actually buys its own grocery.

    16. MM

      The, the pipe that fixes itself, yeah. [laughs]

    17. SP

      The pipe that fixes itself. All this stuff. The promised future of the IoT, like, likely see a, a, some kind of a combination of that. Like, the worst case scenario is not going to look like abruptly everything is stolen. The worst case scenario will look like one by one, industry by industry, absorbed by this new technology. Like, big groceries will be replaced with, uh, online delivery, which will be then give more jobs at some specific period of time, but then less as soon as the people will be replaced with robots as well. So, like, there will be a spike of additional small businesses and additional jobs in, in some areas, but then decline as well. So it's not going to look like, oh, we're going to hell. Something which we're not going to notice, our freedom will be taken

  11. 17:0818:12

    Who controls your answers controls you

    1. SP

      away.

    2. MM

      Can you talk more about the loss of freedom? 'Cause the analog situation that I see here, social media, right? We all have access to these amazing videos, but then some people can't stop consumption, and they just spend their day scrolling, and fake information is affecting how they make decisions. Do you think it's gonna be worse with AI?

    3. SP

      I started with this. Decision-making and your research, if you actually ask the question, get the singular answer to this, the source of this information, if it's controlled, defines what you think about the reality around you.

    4. MM

      That, that's what we're talking with some of my friends, 'cause all of our AIs gave us advice to work less. We're like, "Are they coordinating?" [laughs]

    5. SP

      [laughs]

    6. MM

      Like, what's going on? [laughs]

    7. SP

      We know how, uh, powerful narratives are. It's not, uh, because of AI in general. It's a part of, uh, uh, human nature how storytelling affects us. The question is, like, who's behind it, and what's the

    8. SP

      Purpose-

    9. MM

      Yeah

    10. SP

      ... of people who provide users.

    11. MM

      Does it

  12. 18:1221:23

    Should the government step in?

    1. MM

      have to be regulated then? Do you think the government should step in with, like, all the declarations that we have that we agreed on together as, you know, Western world? Should that be the guidance for AI? Um-

    2. SP

      The government's already doing this.

    3. MM

      Well, in the US, I think it's hands-off approach. [laughs]

    4. SP

      If you listen, uh, all the current leaders of AI industry, they all want the industry to be regulated. Um, we are not so fond of governments in general. Uh, we believe that, uh, uh, some sort of regulation, uh, is nice, but it will be, uh, really hard to prevent governments to abuse this power, uh, as well. We really like being in United States because of, uh, how much freedom people have, how much, uh, our government is actually decentralized. Uh, not everything is decided on federal level. Uh, many things decided on state level or even the municipal level. At the same time, when we talk about, uh, AI regulations, likely to be federal levels, [laughs] so not m- not much, uh, freedom out there. Um, so we are not as, um-

    5. SP

      Optimistic

    6. SP

      ... optimistic about, uh-

    7. SP

      On the regulation side

    8. SP

      ... um, regulations, where they will, uh, take us. Uh, but at the same time, we are quite believers in consumer power, uh, how consumers can change how all these companies behave, uh, how they can demand what should be-

    9. MM

      How?

    10. SP

      Right. And actually even just, uh, being vocal about what, um, they want to have, uh, in their life.

    11. MM

      But I'm thinking about OpenAI, Anthropic, and the defense situation. Yeah, people cancel their subscriptions. Actually, we see... I, I don't know how much that affected, uh, what's going on internally, but-

    12. SP

      Somewhat affected. Uh, clearly somewhat affected. It also, it, it also the, um, it affects employees of these companies as well.

    13. MM

      Hmm.

    14. SP

      Like, because this time it's all about the researchers. Like-

    15. MM

      Oh, yeah

    16. SP

      ... the only successful AI labs are those which manage to gather the brilliant minds of the, uh, AI researchers. Not even engineers, but AI researchers, because the engineering part is now done by AI itself. So, like, you need to actually think about this, uh, rather than code.

    17. SP

      We saw multiple times how those researchers left some organizations because of the-

    18. SP

      Disagreeing

    19. SP

      ... ethic, uh, disagreements

    20. MM

      Disagreement with the founders, yeah.

    21. SP

      Um, so I, I do believe that, uh, right now it's still, uh, consumers have quite a big of a power. It won't be like that in a couple years even, so we don't have, uh-

    22. MM

      Oh, you think so?

    23. SP

      Yes, I, I'm pretty sure that-

    24. SP

      Wow

    25. SP

      ... the, the power of these organizations will only-

    26. SP

      Increase

    27. SP

      ... uh, increase. And our dependence on particular models and our dependence on having access to the latest, uh, model will only increase over time. And, uh, I think that's, uh, probably the most urgency in the process right

  13. 21:2324:37

    What one person can actually do

    1. SP

      now.

    2. MM

      What can we do about it?

    3. SP

      We can be vocal. We can support, uh, every, uh, independent projects. Uh, we do believe that decentralization, uh, will play significant role, uh, when people, uh, will build, uh, AI not based on data centers, but on the community of people who joining forces, um, together.

    4. SP

      It's happening step by step. So, like, several companies are right now, uh, providing the service of actually building s- some micro data center in your house. Or, like, uh, it's a grid of micro da- data centers. When, when it's less about, like, a centralized control of, uh, a specific corporation. Uh, I mean, not necessarily these particular, uh, companies are the ones which will, uh, lead to this sort of decentralization, but it happened with the power, with the, like, disconnecting from the grid, uh, having the, having the solar panels and the batteries actually providing all of the electricity at home. This is happening with increasing speed now as, as the battery technologies and the solar panel technologies are getting cheaper, cheaper, and more efficient. Similar is actually boiling in a lot of people's mind on how do we decentralize a lot of this. And there is also a chance that because of, uh, significant centralization, there are some additional, uh, physical barriers, like enough of electricity for such large data centers. Well, if you actually decentralize this b- home by home with, uh, with micro data centers, then e- every single house solar panel with a battery actually can power the massive infrastructure of, uh, of the AI, uh, cloud data center. So participating in projects like this, like even just browsing and searching and researching on what, what's possible and what's going on in there, and then jumping onto being, uh, early adapters of technologies like this, this will actually help a lot in creating a, a pressing alternative. Like, the alternative which not necessary will become bigger than the b- large corporations themselves, but will become visible and annoying enough for the large corporations to adapt and change and be less monopolistic. The trick is AI itself, as soon as it's achieving, will achieve soon, like in a year from now likely, a certain level of, uh, precision and, uh, false tolerance and mistake tolerance. They will be making decisions More uncorruptible, more immutable, uh, than people themself. More, le- less biased. Like, all of this. Like, we can see now how the call centers industry has been disrupted by AI. Like, it's sort of invisible for majority of people because the call centers are somewhere there, like in, in Asia, somewhere far away. It's been significantly disrupted. Uh, we, we know several multi-billion dollar companies and smaller businesses which are right now completely switching to the AI call centers. Moreover, it's more efficient. So, like, friend

  14. 24:3726:48

    AI that's more empathetic than people

    1. SP

      of ours, their, their banking system, uh, they're giving loans, and they used to have the collectors, uh, call center. Like, people would call you and say, like, "Hey, we can help you to restructure your debt." And usually the people in the bad situation, when they already under the pressure of their, uh, actual life situation, plus the bank is calling and they're demanding the money-

    2. MM

      Yeah

    3. SP

      ... they are not quite willing to actually even, even reply to the phone call. So what they managed to do for themselves, they completely replaced the actual real human beings on the other side of the phone. Why? Because the AI happened to be more empathetic. They designed it to be more empathetic. The AI getting into the human being situation, uh, much deeper and help the human being to actually get out of the shitty situation of owing money to the bank.

    4. MM

      And they laid off the agency that they used to work with, the call center?

    5. SP

      Yes. India, somewhere like Philippines. Like, a lot of this somewhere outside, so that's why we probably don't see this.

    6. MM

      So we don't really... Yeah.

    7. SP

      We don't, don't feel this change, but there is a lot of moving into the call centers, uh, because of this, as I said, in empathy. It's a human being, uh, quality to be empathetic, but AI can be much more empathetic. A- and, and AI can be much more truthful. And as soon as we're getting to this l- level, the governmental services, logically, should be AI driven. A lot of it.

    8. MM

      Mm-hmm.

    9. SP

      And there will be a huge resistance from, from the very government itself-

    10. MM

      Yeah

    11. SP

      ... to be replaced, to be augmented-

    12. MM

      Yeah

    13. SP

      ... to be, to become uncorruptible. And that's where we can see that there will be tension.

    14. SP

      If you want to, to create this, uh, uncorruptible bureaucracy, it's quite important that, uh, uh, this type of, uh, artificial intelligence shouldn't be controlled by anyone.

    15. SP

      Should not be centralized or controlled.

    16. SP

      Be central- centralized. And, uh, none of us wants, uh, the, the future bureaucracy to, uh, to be, uh, controlled by a corporation. Um-

    17. MM

      Which is, which won't happen, right, already? Like, because these corporations exist, do you think there's a chance that there's a world where it's decentralized?

    18. SP

      But remember

  15. 26:4831:54

    How Bitcoin beat the untouchable giants

    1. SP

      what was with the Bitcoin. The financial world was the most powerful. The Wall Street and the, and the-

    2. SP

      Central banks. No one-

    3. SP

      Central banks of the world

    4. SP

      ... can challenge them.

    5. SP

      No one could have challenged them. And then Bitcoin just happened. Yes, it started, like, a little bit early, but it was like in- invisible, uh, a, a, a toy.

    6. MM

      Yeah.

    7. SP

      But right now it's the rails of financial world. No one can affect the re- records in the ledger.

    8. SP

      For example, privacy en- encryption in messengers. The, the, the big players didn't want encryption. They didn't care.

    9. MM

      Then Telegram came [laughs] .

    10. SP

      But then smaller players-

    11. MM

      Came

    12. SP

      ... came.

    13. SP

      Signal, Telegram

    14. SP

      And they, they all had to adopt it-

    15. MM

      Yeah

    16. SP

      ... and they all had to, uh, have it. And this is a example of, uh, consumers demanding.

    17. SP

      At some point, there will no be difference for, for majority of the consumers whether it's Opus or GPT-5 or something which they don't even understand the name of it, but they prefer-

    18. MM

      For a lot of people, it's already the case.

    19. SP

      Exactly.

    20. MM

      Wow.

    21. SP

      But they prefer to use it because of the aspects which we're saying. Some people will, step by step, like it, like it was with Bitcoin. It started with Satoshi himself, then 10 people, then 20, then 100, then 1,000. Now it's hundreds of millions of people actually holding the Bitcoin worldwide.

    22. MM

      Yeah. But Satoshi, uh, he remains unknown to public. I know you're building something in decentralized AI, and you're going to podcasts. Do you have any fears [laughs] of big corporations?

    23. SP

      Of us disappearing? [laughs]

    24. SP

      Uh...

    25. MM

      Uh, no, like, 'cause you're, uh, entering a huge market, and you're trying to disrupt and build a competitor.

    26. SP

      We are not afraid.

    27. SP

      Y- y- you know, with, uh, Bitcoin, the, the, the privacy of the owner was part of the pitch.

    28. MM

      Mm-hmm.

    29. SP

      [laughs] So i- it was a, an important aspect, uh-

    30. SP

      Of the Bitcoin itself

  16. 31:5433:00

    The coming boom in small businesses

    1. SP

      access to this technology, then everyone can achieve their, uh, their, their highest potential.

    2. MM

      Yeah.

    3. SP

      In the world where people will be losing jobs, if you can come back home saying, "I'm un- unemployed now," and then AI will tell you what next you can do, and you can just, like, choose in between-

    4. MM

      Vibe code a new startup, yeah

    5. SP

      ... Vibe code. Like, not, not, not necessarily everyone gonna be a startup, but at the same time, I personally believe that we, we will see the explosion of smaller businesses. Like, we will see the de-corporativization of, uh, i- industries if we have equal access to AI. The same way, uh, social media affected small businesses in terms of, like, now anyone can start a business online and sell whatever, like clothes, uh, jewelry, cosmetics, like, whatever it is, products for, for dogs, for kids. Like, through this new channel, more small businesses happen to be successful. We will see that tenfold with AI if access is not, uh, limited.

  17. 33:0033:47

    What to do in 2 years

    1. MM

      If access is not limited, what should people be doing now to thrive in two or three years?

    2. SP

      First of all, uh, you, you definitely, uh, need to at least try to adapt these tools. You know, in our-

    3. SP

      Use it

    4. SP

      ... bubble, in our bubble, everyone tried to use agents. [laughs]

    5. MM

      Yeah.

    6. SP

      But in reality, it's just still a fraction-

    7. SP

      It's just a couple million people for-

    8. SP

      ... of us, uh-

    9. SP

      ... in the entire world who actually-

    10. SP

      Right

    11. SP

      ... are trying to do this.

    12. MM

      Yeah

    13. SP

      Um, so people should be much more open to educate themself and to implement this in their current, uh, workflows in the way how they do their job.

    14. SP

      Plan their days, do their job.

    15. SP

      Um, definitely is this-

    16. SP

      Do their ho- household planning, like all of this

    17. SP

      ... this is, uh, quite, uh, an important, uh,

  18. 33:4735:24

    Where AI can take you: real story

    1. SP

      thing. I, I remember that, uh, when we used to have, uh, a game development studio, we hired an artist. He didn't have so much of the, uh-

    2. SP

      Portfolio

    3. SP

      ... portfolio at that moment

    4. SP

      ... like, uh, experience

    5. SP

      ... I remember just, was just couple, uh, pictures, but they impressed us, uh, quite a bit. So, so we hired him.

    6. SP

      The talent was clearly there.

    7. SP

      And, uh, uh, all o- our other artists had much greater experience at that moment. But this guy, every time when new tool appeared, he wasn't resistant but just-

    8. SP

      Would jump on it

    9. SP

      ... jump in and use the tool.

    10. SP

      Try use it, try to understand what's, what's, what's the best trick he can get out of this.

    11. SP

      The rest resistant. Uh, and in just couple years, we got a message from him that he got a proposal from Blizzard, one of the, the-

    12. SP

      One of the top game development studio in the world at the time

    13. SP

      ... uh, and he will not refuse it. And, and for us, it was for sure he will want to work in this company, especially in the cinematic division when you do these beautiful movie shots. And so he got there, and, like, and then couple years-

    14. SP

      It never stopped

    15. SP

      ... uh, in co-

    16. SP

      He advanced himself, advanced himself into robotics, into this, into that. Like, as soon as you could consume the change, uh, the new-

    17. MM

      It fuels you

    18. SP

      ... accept-

    19. MM

      Yeah. Yeah

    20. SP

      ... it fuels you in, in to continue, in never stopping.

    21. SP

      I know that, uh, for, for some it is, uh, uh, a little bit scary to use these tools, or when they are new, they're a little bit sloppy.

    22. SP

      But just

  19. 35:2436:43

    What non-coders build in one hour

    1. SP

      make it a habit to spend at least an hour a day, instead of watching Netflix or browsing Instagram, to actually see what, what you can do.

    2. MM

      To deploy. Yeah.

    3. SP

      To deploy.

    4. MM

      Deploy, yeah.

    5. SP

      And so we-

    6. SP

      Research

    7. SP

      ... we recently found ourself with, uh, several groups of friends doing hackathons at house, at the dinner. You know, it's a culture when you just sit down, and you have one hour to build some stuff.

    8. SP

      It's important to say it's not technical people.

    9. SP

      Yeah, yeah.

    10. MM

      Mm-hmm.

    11. SP

      So we did multiple times, and it's so fun. So fun. Like, uh, we, what you can build just-

    12. SP

      In one hour

    13. SP

      ... in one hour, and you can have it on your phone immediately, and you all can try, and you c- all can laugh. Um, so-

    14. SP

      Not necessary it is going to be useful, but the fact that you have this physical experience of achieving this in just one hour With a, a group of your friends who actually never had experience of building software before, it changes the perspective. So, like, our advice would be just use the AI.

    15. MM

      I also write a newsletter where I go deeper on AI tools that I use, career strategies, and things I can't fit into a 60-minute podcast. It's free. Link is in the description.

  20. 36:4339:53

    Rapid-fire predictions for 5 years

    1. MM

      Okay, I could talk about this for another hour. I have a blitz to wrap up with.

    2. SP

      Try. [laughs]

    3. MM

      So it's like a short answer. Will AI be able to do most intellectual work for humans in three years?

    4. SP

      Yes.

    5. SP

      Yes.

    6. SP

      [laughs]

    7. MM

      If it does, what is left for humans?

    8. SP

      Uh-

    9. SP

      The same

    10. SP

      ... uh, creativity.

    11. SP

      Creativity, collaboration with-

    12. SP

      Uh, to make decisions in your life, what you want to have in your life. This is, uh, would be quite important. In world of abundance, is actually will be a quite complex task-

    13. SP

      More things to do

    14. MM

      Yeah

    15. SP

      ... to decide what exactly how you want to live your life.

    16. MM

      'Cause you have everything, yeah. Will AI create more jobs than it destroys?

    17. SP

      The nature of jobs might change, but it might change in a much longer, um, horizon. I believe AI, first, will create a lot of new jobs, and then the idea of job might, might dis- change and dissipate in 10, 15 years from now.

    18. MM

      Will it?

    19. SP

      Maybe 30 years. Like, this is lo- longer horizon.

    20. MM

      Will today's market leaders still be leading five years from now, or will we have some emergent player?

    21. SP

      Likely ... So look, uh, Anthropic haven't existed three years ago.

    22. SP

      Three years or four years ago. Like, not in AI, right?

    23. SP

      Uh, actually, if you look, if you look all the incumbents, uh, all the big companies we had before, all of them lost the AI game. Maybe Google still have some place, uh, but, uh, all the rest, they actually weren't successful.

    24. SP

      Missed.

    25. SP

      Every time when someone, uh, point out that all, uh, this, uh, big company, uh, will be competitor to you, uh, I, I, I, I, I don't take it seriously because big companies, they are really slow. Uh, and, uh, smaller players, they have, uh, much more chances to actually build something, uh, greater.

    26. SP

      So yes, we will see a lot of new-

    27. SP

      We still will steal them

    28. SP

      ... someone new. Likely those who we see right now might change a little bit in terms of, like, mergers and acquisitions, like, becoming one, some of them, or just disappearing or losing the competition to the rest. But there will still be the same 10 companies of the Silicon Valley, which are trillion, each trillion dollar company, multi-trillion dollar company, uh, in three, five, 10 years from now. Likely they're not going to go away. Will something new happen?

    29. MM

      Totally.

    30. SP

      For sure.

  21. 39:5340:44

    The smartest move you can make today

    1. SP

      Yeah.

    2. MM

      Okay. And the last question, I think I know the answer, but what's the most underrated move someone could make tomorrow to have more control over their life in 10 years?

    3. SP

      Love yourself. [laughs]

    4. SP

      Invest time in your development.

    5. MM

      Mm-hmm.

    6. SP

      Uh, for sure. It always was a good advice. Uh, currently, it's actually much more important.

    7. SP

      No, no, it's like we cannot.

    8. SP

      The world is changing.

    9. MM

      Yeah.

    10. SP

      Uh, the world is changing. And, uh, without, uh, actually educating yourself, you're losing a lot of opportunities.

    11. MM

      I thought you were gonna say run a little data center in your house. [laughs]

    12. SP

      [laughs] This is, this is-

    13. MM

      That's another one, right? [laughs]

    14. SP

      What I'm saying, it just, uh, in different words.

    15. MM

      Thank you so much, guys. Thank you for being modern age philosophers, and also for building something that's really exciting for this new age. Thank you.

    16. SP

      Thank you.

    17. SP

      Thank you for this great conversation.

Episode duration: 40:44

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode HMvTSaaO9sU

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.