Ep 168: Vytautas Kubilius

 

Google's Vytautas Kubilius on the risks of too-slow AI adoption, opportunities for startups to build on top of BigTech AI & everyday productivity hacks

Vytautas Kubilius is the Country Director for Google Baltics, playing a pivotal role in driving Google's mission to organize the world’s information and make it universally accessible and useful within the Baltic region. He is also an angel investor and working with such organizations as Change Ventures, StartupWiseGuys and Vilnius Tech Park in various supporting roles.

On this episode we talk about:

  • AI Innovations from Big Tech

  • Responsible AI Deployment

  • The Challenges of Managing Confidential Data

  • Competing for Talent on a Global Scale

  • Future of AI in Startups and Enterprises

  • Evolution of Google Search with AI

We are on YouTube and Linkedin as well

 Watch select full-length episodes on our YouTube channel > https://www.youtube.com/channel/UCP6ueaLnjS-CQfrMCm2EoTA 

Connect with us on Linkedin > https://www.linkedin.com/company/pursuit-of-scrappiness/


Read the full episode transcript below

 

Uldis (00:02.151)

Hello, hello, hello, dear listeners. Welcome to another episode of the Pursuit of Scrappiness podcast. Whether you're building a business, running a team, or just starting out in your career, we are here to bring you scrappy and actionable insights to help you become more productive. My name is Uldis Teraudkalns and my co -host is Janis Zeps. Hi, sir.

Janis (00:24.686)

Hello.

Uldis (00:27.271)

Before we start, a quick reminder, click follow on the platform of your choice for listening podcasts, Spotify, Apple, we're there. In exchange to that, you will get more than 160 episodes covering all kinds of topics to help you build a business or run a business or just do a good job at a company. So follow us and you will be the first to know when we come out every Tuesday morning. About today's topic. So...

As you know, on this show, we are very proud to talk to a wide range and profile companies. We've had early stage founders, we've had founders who have just closed their companies, we've had unicorn founders and people from large corporations. And today we're adding one missing piece in this puzzle by talking to someone from Big Tech, a startup that became one of the world's largest and most profitable corporations, a company that has changed the way we search, email,

store documents, use maps, and probably do most things on the internet. And a company that is now engaged in the fierce AI race for world domination. Or did I make it sound too dramatic? We're happy to have someone from Google today. And in the true spirit of this podcast, we have someone who can discuss global events and trends from the Baltic angle. Very happy to welcome Vytautas Kubilius, country director of Lithuania, Latvia, and Estonia at Google. Hi.

Vytautas Kubilius (01:57.905)

Hello, hello, good to be here.

Uldis (02:01.689)

So Vytautas has been at Google for over 12 years and is also very engaged in the local startup scene in various capacities as a board member, mentor and also investor. So today we want to talk about AI innovations coming from the big tech, how it will change our work life and societies. And we'll also go into details about how startups can benefit from AI as well as what challenges do we face here in the Baltics when trying to build globally competitive solutions. So let's start.

So what's happening in the AI race at the moment? OpenAI, the big five, can we say that someone is ahead or any advantage someone might have currently that will matter in the long term?

Vytautas Kubilius (03:02.769)

Look, it's a good question and I'm not entirely sure if I would call it a race, to be frank. I think there is probably a little bit of a vision of a race, right? It looks like a race. If you think about it, actually...

Google has announced itself as an AI -first company back in 2016. So it's been going for quite a while already, right? So it's not something new. It seems like it's new. It seems like it suddenly happened in 2021, 2022, when LLMs kind of broke through.

Janis (03:31.183)

No.

Vytautas Kubilius (03:34.481)

But again, LLMs are just one of the AI types. So we see right now probably a phase of early deployments of products, of services, and this is why we kind of see this massive pace, massive kind of something new is always happening every couple of weeks. And that's why it looks like a race. But even if you would treat it as a race, within Google, we have a little bit of a different perspective on that. So the way Sundar Pichai, the

the CEO of Google is talking about it, is that the only race that we want to win in this situation is the race of deploying this in a really responsible manner. It's a massive technology, right? So this is a technology that gives us an absolutely unprecedented opportunity as a human race. And I know that this sounds grandiose, but really this technology is just so shifting, you know,

Janis (04:30.479)

Yeah.

Vytautas Kubilius (04:34.243)

so many things, so many technologies around it, we have an opportunity to really have massive impact on society, on businesses, on sustainability, on climate change. And so making sure that this is being deployed in a smart way, in a way that we fully understand that we fully control all of the risks associated with it, all of the quality parameters, that we train the models in a way that it really represents the world.

in a correct way, that it's not like super centric to Silicon Valley, and that it incorporates all the nuances and all the things around the world, and it serves the people around the world in a way that they would anticipate. So AI is not something new as well. It's a technology from 1960s. And now if you think about it, where it's being deployed, it's actually quite more probably in...

that you would think of. So if you think about Google search, you already have AI deployed there for quite a while in order for it to understand the nuances of language, in order for it to understand what is it that you are potentially looking for. So again, these are the capabilities that we see right now in LLMs, just as well, right? It needs to understand the context of your query in order to provide you the response. I don't know how about you,

I'm really, you know, I'm a marketeer by trade. So I built myself a habit of really trying to notice my own kind of usage of things, my own patterns. And I notice right now, like for example, Gmail.

it's absolutely mind -blowing how good autosuggest has become. It takes the context of this entire email spread and it starts to suggest you the next kind of bits and pieces. And I'm noticing myself being like, dab, dab, dab, yeah, that's good, that's still good, that's good, yeah, signature, that's good to go. So again, this is AI in action. I don't have a cable TV for a couple decades now. I'm trying to limit a little bit my

Uldis (06:38.983)

you

Vytautas Kubilius (06:52.259)

local news kind of exposure. I do watch a lot of YouTube and I'm watching more YouTube lately. Again, that's where AI algorithms are being deployed in order to do better, to do better suggestions for you. So.

Janis (07:07.342)

That's true, I've noticed while YouTube has improved and in I would say last year or so, yeah, quite a lot.

Vytautas Kubilius (07:11.153)

Massive.

I'm subscribing to a bunch of new channels. We have a habit in our household, we're doing a little bit of a lazy learning Sundays where we are putting just videos, documentaries, Vox, et cetera, so just interesting stuff. It's amazing how much quality stuff there is and it's amazing how much more new stuff I'm getting exposed to. So again, algorithms are getting so much better. So we see AI being deployed in so many places already that,

Janis (07:16.399)

Yeah.

Vytautas Kubilius (07:44.003)

we don't really think about too much, right? We are so focused on this. They released a new model, they released a bigger model, and these guys released another model. I think it's wider than that. And again, I am really of a firm belief that we do need to make sure that we are smart about how we deploy it. Google has earned a reputation for quality over the years, right? So the expectation towards Google is massive.

We have, what is it, six products now with over two billion users, I think 15 products with 500 million users. You wouldn't want to have Google Maps suddenly sending you to a wrong location or telling you that it's a 10 minute walk when it's actually an hour. You wouldn't want to be directed on search to something that isn't relevant to you. So we have this responsibility very...

Uldis (08:26.823)

you

Vytautas Kubilius (08:42.129)

deep in our company DNA that we need to make sure that this is really done correctly. So yeah, in terms of race, I don't know. I think it's a matter of perspective a little bit.

Uldis (08:56.487)

And in terms of applications and usages, you're building both kind of infrastructure and applications. I mean, they used to say that people who invented the fridge made some money, but then Coca -Cola made all the money by serving a cold drink. So have you seen the Coca -Cola of the AI starting to emerge? What will actually take the most value?

is created with this technology.

Vytautas Kubilius (09:33.905)

Interesting analogy with Coke and fridge. Okay, haven't heard that one.

Janis (09:38.863)

That comes from our friend Chamath.

Vytautas Kubilius (09:41.969)

Okay, okay.

Janis (09:43.662)

I sometimes see some clips online from him and I took this one but yeah anyway please carry on and just...

Vytautas Kubilius (09:51.249)

Yeah, yeah. Look, I mean, again, I thinking about AI, we haven't we are just starting, right? We are just starting already right now. It's mind blowing the amount of applications that we can see. What's going to come up still and where we're going to see new applications is unknown or rather, it's a good question where you cannot apply it. So.

I wouldn't, I would bet that we're not gonna see a single...

a single or one single industry who is a winner, I'm rather of an opinion that AI can be really a technology that is going to be helpful for everyone and that we can apply to many, many industries. And so if, you know, even off top of my mind, again, I'm a marketeer, I'm not an engineer, I'm not a biologist, but it still blows my mind when I'm reading

like announcements like alpha fold three, right? So it's a model that helps to predict proteins and DNA and RNA interactions between each other. This used to take massive amount of time for scientists to do even one model. Now suddenly, apparently it's able to predict the folding of proteins of all of the known proteins, a single model. So we're talking about medicine, we're talking about

materials, we are talking about so many potential applications across all of the fields, right? And again, this is not something that you would have thought probably when you started thinking about, okay, so we are going to, you know, into LLMs and they are able to answer our queries. Again, that's a great example of where this can be applied. Another example was project relate, I believe it's called. So,

Vytautas Kubilius (11:57.699)

There are 250 million or so people around the world with speech impediment. So they have difficulties communicating with other people. Again, AI is able to...

improve the capabilities of understanding their language, of understanding their full sentence, and is able to help them communicate, which is absolutely amazing. Again, we are talking about a massive amount of people, of global population, whose life can be improved. We're seeing projects where AI is able to reduce the traffic by optimizing the city traffic lights.

We see pilot projects for airlines that are helping them to choose the right altitude because apparently the majority of the CO2 emissions are happening when the jets are flying through a specific year, year streams and this is when they are producing these long tails, you know, and you can actually avoid that by predicting the weather better and by choosing the right altitude. So again, we are reducing massive amount of carbon dioxide going to the air.

into the year. So, and then you have all the very basic stuff, you know, what we see right now, the widest application, which is in the productivity tools, right, in the creativity tools, where people are able to use AI just to speed up the work, just to get a little bit of a creative juice going for them.

I do a lot of slides myself. I love those beautiful single picture slides with one sentence on top of it. For me, aside from the content and figuring out the full kind of flow and the idea that I want to convey,

Vytautas Kubilius (13:55.889)

I think a very good chunk of time would be going into me going through all the databases of images trying to find exactly what fits me in that particular slide, right? I need a...

Janis (14:07.758)

And it always takes time.

Vytautas Kubilius (14:09.233)

Exactly. I need a chimpanzee in an austro astronaut costume next to a board looking upwards, right? I mean go figure try try and find that Now I can do that directly in the slides. I mean, I just Describe it and I have the I have the picture You think of you again the same, you know email writing?

transcription of video calls, which is amazing. Like this is, you know, we obviously are working on the Google productivity tools with all the, some even of the test for a new kind of features launch, but this is something that we recently announced coming to Google workspace. So for example, you can do the full recap of the context of the video call.

And if I'm joining a little bit late into the call or I'm going through a recording, I can see the summaries. It's not a word by word transcript. It's the contextual transcript. This is what's being discussed here. This is what's being discussed over here. And I can quickly jump through the video. And it blew me away when I was having a flight and I had to listen in into a management call. And I went through a one hour call in less than 15 minutes.

I'm fairly certain I didn't miss much because I either read it or I went into exact spot where I wanted to hear it in more detail. So these are amazing, amazing time -saving kind of moments that we're gonna see applied across many, many people and the broader population. So coming back to your question. Yeah.

Janis (15:49.902)

Yeah, I'm old enough to remember that actually people used to employ people who take notes in the meetings, like protocols, right? Not that long ago.

Vytautas Kubilius (15:56.241)

100%. 100%. Yeah, you're right. You're 100 % right.

So, you know, coming back to your question, like, is there gonna be a single winner? I'm sure that there are gonna be companies that are gonna do very well in this age, right? The ones who are gonna provide the infrastructure, so, you know, the chip makers, the model makers, the platforms, probably a bunch of other companies around it. And again, you know, if I think about how Google is approaching it, we do have this Vertex AI platform, and we are very clear, like, we wanna deploy

AI on our products, but at the same time, we cannot be everywhere and we shouldn't be everywhere. So we are giving all of these models to companies, to enterprises, to startups with an idea, like take them and build stuff, build stuff on them because we are fairly certain that this global creativity is boundless and we are unable to think of all the possible use cases.

Janis (16:58.158)

and then you can buy them. Which is how nice for the startups.

Uldis (17:00.391)

Speak.

Vytautas Kubilius (17:06.033)

I would love to see Google, frankly, buying some of the Baltic startups. Would absolutely love that. And that's something that is still on my list because I do believe in Google. Well, my goal is for Google. 12 years in the company, you need to have some sort of a, you know, when can you consider that you have achieved what you came here to do?

Janis (17:17.102)

shopping list.

Uldis (17:31.847)

So speaking of building one aspect of building a business and selling online that almost everybody relies on is the main product of Google and Google search and I bet that you are getting a lot of incoming questions and concerns about companies that have spent years optimizing their content.

spending money on Google search campaigns and different kinds of ads. And now there's AI coming in and people are getting worried. So are all of those investments are going down the drain? How do you see this at Google of how this whole industry that has created some companies don't even have any other marketing budget than just for Google. So how are those going to be affected?

by AI, obviously.

Vytautas Kubilius (18:31.793)

Yeah.

Look, I think one thing that is important to recognize here is...

Internet is always changing, right? So this is not the first kind of big change, if you will. So if you think about it overall, how Internet came to be and how companies started moving online, then we had the mobile shift, if you remember, right? When there was also a little bit of a resistance and all the doomsayers like, so what does that mean? Is desktop dead? Is everyone going to be on mobile? Is like the commerce dead? Is everything going to be on mobile commerce?

Janis (19:06.959)

Yeah, I'm good at times.

Vytautas Kubilius (19:08.785)

So there are always these kind of a little bit of anxieties that are heightened a little bit, but I would say, you know.

Google's search algorithms are always changing. Are always changing and shifting because people are changing and shifting. It's still one of the one of the coolest stats that I have from Google, which is mind boggling to me. Every single day, billions and billions of searches are happening. 20 % of those every single day are completely and totally unique. They have never been seen on Google. Never.

So people are very creative in terms of how they are looking for stuff, how long, how precise, how imprecise, for example, those queries are. And so it's an ongoing iteration for Google to always figure out the search experience. Again, we probably, if you would try to look at the Google results from today, and you would look at them in 2018, 2015, 2010, we probably do

don't feel that there is a much change but if you think about all the inclusions of like knowledge panels and if you are looking for like a molecule model you're gonna get like 3d models and all of these little things that are happening so Google search is changing constantly regularly and right now even when we see Google search we are actually thinking still about going on your desktop or mobile into the browser and typing something in but that's

That's not entirely the reality. So now you have the voice search, you have the auto experience, you have the circle to search. I don't know if you've tried, that's pretty new. When you have on your phone, you are able to just circle around on your screen. Google Lens, the visual search on your phone, just like, what's that? I'm pointing my phone at something, show me that. I think we have like 10 or 12 billion searches.

Vytautas Kubilius (21:16.627)

like that a month now. Again, that's pretty new. So there is a lot of new stuff that is happening on search regularly and I don't think that AI is just gonna be something that is gonna change overnight and is gonna, you know...

have companies completely surprised. We are starting these experiments with AI responses. So far it's only available in US. And fairly certain it's gonna be coming to European Union fairly soon. What we do see in those initial kind of searches, and again, quite a lot of searches are happening on these tests kind of queries as well. We see that people are...

going a little bit deeper than before, if you will. So they are going into like multi -step searches. They are going into kind of more exploration mode, if you will. So there, you know, let's say if you are interested in, I don't know, I'm creating right now. Instead of going just, you know, for best gym around me, you would go into best gym and clothing for the gym and then perhaps the best new interesting

programs for getting fit fast, stuff like that. So people are generally using Google search to find deeper answers. And Generative AI is able to complement that a little bit more and give you kind of more perspectives on the topic.

Uldis (22:55.367)

and for the companies? What should they be most mindful of when adapting to this new reality?

Janis (23:07.503)

How do they gain edge? Is there a tip you can give? To be on top of Google now.

Vytautas Kubilius (23:14.161)

Look, the tip hasn't changed since the very beginning. Be relevant. Be relevant and provide high quality...

fresh information to users that is actually useful. That's the only and the best tip. On a day -to -day basis, I generally work mostly with ads. So we, as the commercial arm of Google, we don't do a lot of SEO consulting. We don't know, first of all, how the algorithm works fully. We can give pretty...

generic kind of recommendations, right? Mobile is important, speed is important, good navigation is important, security is important, but it's always the number one thing is good content. So I really don't believe that this is gonna change whether AI is gonna become more important or less important if we're gonna have one algorithm or another, that's the core principle here.

Uldis (24:24.295)

We stumbled on a report recently that in the Baltics employees seem to be more engaged in using AI tools than their managers. If this trend persists, how do you think we can turn it around and how do you think employees can actually get their companies to be...

more faster in adapting AI tools and more progressive and using the benefits of AI.

Vytautas Kubilius (25:02.673)

Yeah, look, I think you are absolutely correct. And there are multiple reports coming in this direction. So I've seen I really like European Commission's DESI. So Digital Economy and Society Index. I'm always very keen to see whenever they are updating lots of great data on enterprises, on small and medium business, on society and how digital we are and what's happening. I've seen a fantastic report now on the Baltics and the CEE.

countries by Women Go Tech, a Louisiana -based NGO, which also showed...

that people are using it quite heavily and at the same time they are not getting the appropriate support from their companies. And that is a little bit of a hush -hush kind of moment of using it in many of the enterprises. And now, very recently, we brought in a report as well from Implement Consulting for CE and Nordics and some of the other European Union countries, which is showing some very interesting data. And I know...

it's a little bit of a public policy kind of country strategy type of data, but nevertheless, I think it's important for us in the Baltics to fully comprehend the idea behind it. So it's very numbers driven, right? So it's like economists, Goldman Sachs methodology on impact on the economy of adoption of new technologies.

The premise there is that if we are to adopt generative AI, which is massively boosting our productivity, helping us to free up multiple hours per week for other tasks, more creative tasks, that we, within the 10 years, so adoption is going up in this curve, kind of S -curve, right? So the impact is a little bit lower in the beginning until it reaches the peak.

Vytautas Kubilius (27:03.667)

peak. At the peak moment, we would be getting anywhere between 5 and 7 percent extra in GDP. So for Lusinea Latvia, we are talking about 3 -4 billion euros per year at the peak of the adoption and of the impact. Now, so that's fantastic.

Janis (27:24.174)

so we can finally afford the rail Baltic.

Vytautas Kubilius (27:27.857)

Here you go. We know where to spend it. So that's the good news, right? The not so good news, which really worries me a little bit here, are...

Uldis (27:28.071)

you

Janis (27:30.478)

Yeah.

Vytautas Kubilius (27:41.393)

a few numbers and the numbers are on one hand the European Commission's data on the adoption of AI technologies right now. So, Lithuania, Latvia, Estonia, we are roughly at 5 % of companies, like company owners, boards that are saying that yes, we actually have adopted, we tried AI, we have a strategy for it, we think it can actually benefit our company.

When we look at the number of employees, we actually see 75, 80 % of employees were saying, yeah, we are using AI at work. Not necessarily because our company has any policy around it, but we are just using it because it boosts my productivity because I actually like what it can do for me. So that's already a gap and that's not very healthy when people are bringing in private tools and private accounts and different tools.

where they are potentially dropping in confidential data of the companies, which is unmanaged properly, which there is no policy around in the company. Can you upload your financial data into an LLM or not? Should we be doing that or not? So that's why. Right.

Janis (28:53.774)

What's our password even for that tool? Maybe, you know.

Vytautas Kubilius (28:56.881)

So, God knows, so that's generally not a good hygiene in IT to have such situation. Second thing is people are signaling very clearly that we believe that this technology is good for my work. It's good for my role that I have in this company.

while the company is not recognizing that. So we have an issue with the boards and the owners. We have an issue on the company level. And if we look down the road, so there was, together with this study, they did a survey among companies within the CE region. And this is what worries me the most is that we are very much lagging behind other CE countries when it comes down to the question, will you?

be investing into AI, into strategy, into deploying AI within your company. And we come down to something like 29%, even lower in Estonia, while we see Poland 47%, 48. We see Slovenia higher, we see Denmark very heavily ahead. And so this econometric model also has two scenarios, or rather three scenarios. You have the base scenario, we adopt at a normal pace, and we

get this benefit. There is a scenario of not adopting it and what happens then? And then your impact is very very small on your economy. And then there is a leapfrog scenario. What happens if you actually push it? What happens if you actually have government support, society support, you know, company support, effort like let's do it, right? Let's adopt it. And so that's when you get the most kind of benefit out of that.

The...

Vytautas Kubilius (30:53.041)

The good moment is we know how to do leapfrogging. We in Baltics, not many people remember, we used to be called the Baltic Tigers at one point. So we knew how to jump the technological kind of errors. We didn't go for the normal kind of deployment of the phone lines or maintaining that. We just went for fiber optics. We went for mobile super fast. We were ahead of Europe when it came down to mobile payments, mobile.

you know, mobile parking, you know, and then we went for e -government solutions. We know how to do it. And what worries really me is it looks like from the current data we need to do the leapfrogging scenario because other countries are already doing that.

And even if you add even more, it's amazing to see how far we have come as countries, how fast we are developing, how fast we are growing. You look at the OECD data in terms of compensation, Estonia, Latvia, Lithuania, number one, two and three countries among the well -developed countries globally in terms of the compensation growth in the last five years. And at the same time, productivity has

grown and in some countries has shrunk.

And this is the technology that is coming our way that talks exactly about productivity. So it looks like we are hitting a problem as countries, as societies, which challenges our ability to be competitive on a global market, to be able to continue building up this well -developed countries with nice societies and well -being increasing across the countries. And here comes the technology.

Vytautas Kubilius (32:44.483)

that offers exactly the solution. And we see that companies are being a little bit slow to adopt it. So this is something that I talk a lot lately. We just brought this research and it triggered a lot of conversations already in associations and with some of the ministries. So we're looking hard at it and I hope that this is gonna spark a strong conversation.

Janis (33:13.198)

That's not even a choice, I mean, it's like when electricity came, I mean, you can keep using the non -electrical plow on your field or whatever, but like, the world will just leave you behind, I mean, it's crazy, yeah.

Uldis (33:17.407)

you

Vytautas Kubilius (33:22.961)

I agree.

I agree, I agree. Look, I mean, I made even on the stage, it's not a very smart comment, but like for me, this whole 5 % of companies want you to use AI while 80 % of employees would like to use AI seems like, you know, you have a typewriter and the laptop and your manager tells you, I don't want to see laptops in the office, like keep doing it on a typewriter. This is not the right approach. I think, you know,

With all the wisdom and years of experience that our board members and company owners have, I think they need to recognize this moment and they need to open up to the possibility that this technology, that they need to allow this technology into the companies and that they need to follow the lead of their employees, that they need to allow this as a grassroots kind of moment, get them the, you know, get them the...

productivity tools, get them the subscription. It's not expensive. And this is also interesting. Why are you not using AI? We think it's expensive. We are worried about the security. And we don't have enough knowledge in the company. So it's not expensive.

It's actually bad security if you don't use it and people use it on a private accounts. And skills, how are you going to get skills unless you actually try and use it?

Janis (34:55.438)

Well, I think in Baltics overall, we have this like either or. Either the managers are young and tech savvy and they grew really fast and they are now in their early thirties, they're running companies and they know. Or I've seen a lot of managers that are very far from tech that don't know how to use Google slides, frankly, you know, and in many companies, right? So that might also be creating this situation as well.

Vytautas Kubilius (35:25.169)

I agree, but again, you know.

I don't see this as a problem as long as people can recognize that times are changing and you cannot look at the future and expect good results by employing the tools and knowledge of yesterday. So as long as there is this recognition, and again, I'm generally, even if I sound pessimistic, I'm optimistic about it because we do have a lot of great initiatives. We have startup associations. We have lots of, we have, again, the,

entrepreneurship agencies of our countries, ministers of economy who understand this topic and who are really investing into the trainings. We are bringing a bunch of free trainings to the Baltics. We just launched a fantastic program in Latvia, Grow with Google, focused on AI skills. And even in the very kind of not very advertised sessions that we wanted to test the response to see whether the content is good, whether the local partners,

We had 1 .2 thousand companies coming over to those trainings. And that's exactly what you need. I think this is the real solution. Is that we just need to increase the amount of trainings and exposure of real use cases. I feel like people are missing, it's all good and great, okay, you can solve the airplanes flying in the sky with AI. How does that apply to my company?

What can I do? We need to bring these very simple pragmatic cases. You can do slides easier. You can do calculations easier. You don't need perhaps a very sophisticated analyst. We can actually do, you know, you can run without knowing how to code and not being a statistician. You can still run some analysis with the help of AI. We need to show these kind of simple cases.

Uldis (37:23.975)

Assuming that our governments and entrepreneurs join hands and do the leapfrog.

and knowing that that AI is needed to jump but is not in itself going to do it, we still need talent. How do you see us competing for talent on the European or global scale given that talent is scarce everywhere, especially in the Baltics, I would say.

So how are you thinking about these matters?

Vytautas Kubilius (38:08.913)

Well, we need to improve the weather first. No, I don't know. The reality is we do have small countries and we do have gaps in certain areas. I didn't realize.

until quite recently, like for example, that we do have a gap with consulting companies, which is a very natural path for young professionals to join them, you know, marketing and business analyst roles and sales roles to start from a consulting company. But we don't have McKinsey, we don't have Boston Consulting Group, we don't have Bain in the Baltics. So we need to accept some of the realities that we have and we need to work with them.

So what I see actually working quite well is an approach...

across a few areas. So first of all, growing internally the talent, right? Working with universities, making sure that they are up to date and that they really have high quality. And our universities are doing quite well. I would say, you know, they are climbing the rankings and they are keen on learning and keen on participating in knowledge sharing and absorbing that knowledge. There are a lot of great private initiatives and various training initiatives. I know that

This startup school has been launched in Latvia. There is a fantastic startup in Lithuania called Turing College that we are also supporting and we are helping to prepare quite a large number of data.

Vytautas Kubilius (39:56.817)

and AI analysts in Lucena. So there are a lot of these initiatives that are, I think, very important. It's important that we support them and that they really thrive and that they are able to...

boost the local talent. And then the second part is we need to acknowledge that there is no way that we're going to grow as many talent as we need. Our startup ecosystem is growing at a fast pace. The amount of employees is increasing quite rapidly. They are opening offices in other locations. And in many cases, if you listen to what the founders are saying, they are being pretty open and vocal about it. They are saying, we are opening an office in Bilk.

in because there is a larger talent pool. It's not because of taxation, it's not because we need operations there, we just need people. Interestingly, by the way, as well, we always thought we need more engineers. Actually, with engineers, we are quite okay. And the report that we also co -funded a couple years ago already started showing the signs that the real problem is going to be marketing, sales, and data analytics.

This is where we are starting to lack the talent pool, which could be very unfortunate. We might become the back offices here, the engineering hubs, while the management is gonna move out of the countries, which would be really a very unfortunate situation. And I see multiple companies. I mean, I've just seen Bolt, one of the executives, was saying exactly that. Marketing is becoming problematic for us in this story. Engineering, all good, great, fantastic,

highly productive people, marketing is problematic. So we need to open up to immigration and we need to be better at attracting people. Despite our weather and despite the six hours of sunshine during January, we do have a lot to offer. I think the cities are amazing. I love coming back to Riga, I love going to Tallinn, I love living in Vilnius because of how vibrant, how cool the city is.

Vytautas Kubilius (42:06.227)

how international they are becoming. And at the same time, I still think we can do much better in terms of the support system that we have for those people when the...

Bill New's office here, we are very international. We have a lot of nationalities, a lot of people coming from different Google locations. And I see that colleagues, myself including, need to help them to integrate into the society. How do you get the smart ID? How do you sign up for a local kindergarten? I mean, there are all these strange rules. What about the medical insurance? What about this? What about that? So I think we can still improve the processes. We can still do more.

much better in terms of, here is your expat package. This is how you become a valuable participant in this society. And that's also a very important moment that we see them as people who are coming to add to our culture, to our societies, that these are not employees of.

skills that we are just missing in our country, that we realize that they are living here with their families, with their friends, and that they can both contribute and that we also need to make sure that they have a great quality of life here. So there are quite a few things, but I'm positive in terms of where we are going and the progress that we have made.

Uldis (43:36.295)

In terms of the proliferation of AI,

Do you think it's a democracy, like bringing more democracy to the startups and equaling the playing field? Or do you think it's actually very resource intensive and will just accelerate the dominance of Silicon Valley and the bigger global tech hubs?

Vytautas Kubilius (44:08.849)

Look, I don't think that you can fully eliminate the head start that Silicon Valley has, right?

They do have absolutely amazing universities over there. They do have absolutely amazing community and the network that you can quickly access over there, capital access and beautiful weather as well. So it's impossible to eliminate those things, but.

If you think back to, I don't know, 2000s or 2010, it's also very different right now. The capital has become much more mobile and the funds are much more global and they're much more interested in the talent. They are less likely to demand that you move to London or to US because we just don't understand your legal system and your geopolitical kind of situation.

So the capital is mobile. Now if you look at the resources, they are much more accessible just as well. If you think about all the Googles, the Amazons, et cetera, they had to build all of the infrastructure from scratch. If you go to Stanford University, you can still see the first Google server made out of Lego bricks.

They were that scrappy. So they were just taping together a couple hard drives and putting them into a Lego box.

Vytautas Kubilius (45:46.705)

You don't need that, right? So you have all the cloud infrastructure right now, which is much, much cheaper than building your own capabilities. And again, I work on a daily basis with our advertising business. We have tens of thousands of customers across the Baltics that are using the platform to find customers anywhere in the world. And just like our economy has become one of the most

open countries in the world. We are right there together with Singapore in terms of the GDP proportion that comes from export. And I see the same happening with our platforms. I see the same how they use Google Ads, how they use YouTube. They are reaching customers globally. It doesn't matter where you are from. It doesn't matter if you are a French, British business. Our startups, not only startups, our small and medium businesses are able to

actually find customers global. And I love that. I love that there is like a real, very clear, tangible kind of benefit to both those companies and economies at large from this. So from this perspective, I would say...

Doing a startup right now is so much easier than it used to be. You are still constrained somewhat, right, by the talent, but again, we have become much more comfortable with the hybrid or even remote work, so you can actually have a team in other...

Locations and if you can access capital and even if you can access customers and the infrastructure Which is global what's there to hold you? What's the big difference then right?

Janis (47:34.288)

Yeah, totally.

Uldis (47:34.759)

And finally, you mentioned that you love using AI to build slides. Are there any other really important productivity tools that help you become more productive?

Vytautas Kubilius (47:52.945)

So, well, I do have a couple of things. Look, I mean, despite my very good looks, obviously, I'm not such a young person anymore. I've been in business for a couple of decades and I have built my own kind of thing. So I live and die by the calendar. If it's not in the calendar, it's not happening. So that's one. My schedule is planned weeks and sometimes months ahead.

I am a huge stickler for zero inbox in Gmail. You never reach that, but you're always aiming for it. Well, if you reach it, that's once or twice a year, but I treat it as my task list. So if I had reminded myself to do a task, if it's in my inbox, then I need to actually complete it, then I can archive that message. And I use AI a lot, and at the same time, I see how much

Thank you very much.

effort it takes me, right? So right now at Google, as a manager, I'm going through probably four or five hours of mandatory training on a weekly basis. Because again, things are changing fast. I'm a marketeer. For me to understand, you know, in depth how AI works and all the applications, it takes probably a little bit more time than to an engineer of some skill. And I'm also obviously, you know, I'm engaging with our...

associations, enterprises, government authorities. So I am also looking for interesting ways how to show them the benefits, how to show them this is what AI can do for you. And this is how it connects to Google Maps. And this is how you can now do a comparison between massive docs so that you don't need to go line by line to see what are the changes that legal team has made, for example. And through giving them training on those examples and giving them those examples,

Vytautas Kubilius (49:52.083)

I see that it kind of I catch myself in moments during my day You know in moments when I'm at work or even when I'm in social environments Where I'm like crap I could actually use AI for that. So I need to remind myself I need to force myself and I'm you know, I'm considering this as building a muscle a little bit right now You know, I used to Google all the time if I'm driving somewhere with my spouse and we are having some sort of an argument just Google it check

I mean, why argue, you know, let's just check some interesting facts. What about this place? What about that church? What about this, you know? History kind of moment. So now I'm starting to think about my task as well. Can I automate it? Can I give it to AI? And this blows my mind.

This truly blows my mind how much stuff I can actually kickstart with AI. It's still an early technology. You still have the hallucinations here and there. So it's not, you shouldn't be expecting it to do the job for you. But you can expect it to be a fantastic creativity kind of, you know.

guide and support for you. So I'm asking you again, it's a matter of training yourself a little bit on how to be good on prompting. So you can just, if you just ask it a simple question, it's gonna give you a simple response. If you add a little bit more and you are like, okay, here is my presentation.

and I want you to be a venture capital investor right now. And I want you to think of all the potential issues that you have with my pitch presentation and ask me all the tough questions and prepare me for this pitch that I'm gonna have. And it's gonna give it, and it's gonna give you absolutely amazing advice. And we see, again, very, very

Vytautas Kubilius (51:56.051)

fresh stuff.

that the LLMs are becoming multimodal. So what does that mean? That it can work with multiple inputs. And so one of the examples that I've been seeing was uploading massive amounts of documents and videos into the same query and saying, this is my doctoral presentation. This is me rehearsing the doctoral presentation. Help me become better at that, right? And so it can analyze both the videos

you and the text and the information that you have provided in the PDF into one cohesive.

Response to you how you are doing and what you could improve So again, I'm looking forward to more of these user just to myself I'm looking forward to agents becoming available to us where you can train sort of the the models to to stay with a little bit of memory and a little bit of your own pre -trained data and Tasks that you want them to complete. I want to see how much can I automate myself? How much I can actually help myself to?

to be more productive and free up a little bit of my time.

Uldis (53:13.191)

I think it's a great, great way how to wrap it up. A very good point that you have to train that muscle. And in some cases, when you don't have an obvious way how you can use AI, you can ask, right? How could I automate this task with your help? And it's going to give you first do this, second do that, third do that. And it works pretty well.

So yeah, let's not hesitate to jump on the AI ship. And yeah, at least you don't need AI to listen to this podcast. So to the listeners, make sure you subscribe and catch our episodes every Tuesday. Thank you very much, Vytautas. It has been a pleasure.

Vytautas Kubilius (54:12.017)

Thank you.

Janis (54:13.104)

Thank you, bye.

 

Please note that the transcript text is AI-generated. We apologize for any potential errors or inaccuracies. Thank you for your understanding.

 
Previous
Previous

Ep 169: Martins Vaivars

Next
Next

Ep 167: Laimonas Noreika