Running 40 Games With Zero Artists, One Launch Every Week, and AI Agents on the Org Chart

Tap to unmute

In this episode, Amit sits down with Tomer Waintroub, Co-Founder and CEO, and Avi Asher, Co-Founder and CTO, of Mad Brain Games. The two met at Playtika, where Tomer moved through a variety of positions and Avi worked as a system architect and R&D manager, and for 14 years they have started the day with the same 9:00 a.m. call on the hour-long drive to the office. They founded the studio together seven years ago, before COVID.

Today the studio has more than 40 games in production and releases a new one every week, with a team Amit puts at between 10 and 20 people and zero artists. Five years ago, it took 27 artists to make the content for a single game.

The conversation gets into how that works: the platform behind a weekly launch, where new game ideas come from, the daily report that goes to every employee, the incentivized channel they built themselves and the fraud fight that came with it, the AI agents with job titles, and the point where both founders stop trusting AI.

The Good Guys Become Superheroes
For Tomer, running small is partly a necessity, since AI has changed how many people a game company actually needs, and partly an advantage, because a small team moves faster and adopts change sooner than an enterprise. What makes it work, he says, is seven years of building the right infrastructure, creative, technical and BI, plus the processes, and tools that get the most out of every person on the team.

"So the good guys are becoming superheroes, and the less good guys which are not utilizing AI's tools are becoming not relevant."

Avi puts a lot of it down to motivation. AI took away the technicality, in drawing and in coding alike, so a motivated person who checks their own work can be very productive, and someone who is into excellence can go much further.

"You can reach five hundred percent of what you're doing."

Build the Platform Alongside the Game
From day one, the plan was multiple games on one platform, and Avi says they understood early that a games platform cannot be designed up front, because the gap between what you imagine and what you actually need is huge. So they developed a game with a platform in mind, then a second game exposed the platform's issues, and the third had a few less.

"You cannot imagine it. Our mind is way too limited to imagine what you actually need."

Each weakness you solve frees your mind for the next one. A year before the recording, putting an environment up took five to seven days, and today it takes about 45 minutes.

Tomer explains why so much of it can be shared. Most systems in a game are the same no matter the genre, ad monetization, segmentation, payments and daily missions are mostly counters and rewards, and the real work in each new title is the core gameplay mechanic, built from scratch or on top of simple code from the Unity Asset Store.

"Daily bonus is a daily bonus."

Everything Ships to Every Game
Tomer's favorite example is Wordle. Three years ago, when it was all the hype, they added it as a mini-game inside one of their word games, expecting the hook to bring marketing CPIs down. It didn't take, and the experiment failed, but Wordle now exists in every game they have, including ad-based slots games.

"It's turned off because we don't need it in the slots, but it's there because it's part of the platform."

The same logic applies to fixes. Avi reads every support ticket, and Tomer says users show you the bugs no number of QAs will catch, and the pain points too, even when it is only a player asking why they can't have more for free. Since most of what they fix or add lives in the platform, every game benefits.

"Whenever we are releasing a new feature, it automatically applies to all of them."

Amit connects it to a rule he keeps repeating to his own team at Sett while it scales, to only do what you can learn from and repeat, and says Mad Brain Games took it to the next level.

Zero Artists and Four Top Employees
The studio has no artists. Tomer credits Mor, their VP Creative, an artist who is not technophobic at a time when many artists are still wary of the tools, the same way they were when the Wacom arrived.

"The top four employees of the company is Claude, Nano Banana, GPT, and Veo 3."

Ten years ago, a single character mock-up took an artist three or four days, you needed four variations to choose from, and then the chosen one had to be polished. Today Tomer writes a prompt, asks for seven or ten themes, and picks.

Amit brings up the intro video from their first game, Word Farm Adventure, a scene of a car crash that leads the player onto the farm. It cost thousands of dollars, included a long argument about keeping the hands unisex, and came down after about a week once they saw it was hurting the first-time user experience. Tomer remembers two months and 20 iterations with the video creator.

"Today, you can tell Sett, 'I want it like this.' After two hours, you have a video."

Where the Next Game Comes From
Asked which games will release the following week, Tomer admits he doesn't know yet, but they will release some. He plays constantly, with 300 games on his phone and the 512 GB of storage he insisted on to fit them, and he says the store is no longer where games get discovered.

"Nobody goes to the store and search for something like this. You just consume it from the ad that you are playing."

So the mechanics come from the ads. An ad they keep seeing usually means a popular game, and playables, misleading as they are about the core gameplay, are often more fun than the game itself, which makes them a good place to find the next mechanic. Avi adds the business logic.

"Whoever is advertising has a viable business."

Variety matters more than volume in one genre. You can clone a game once or twice, Tomer says, but 20 games in the same genre makes no sense, and a sequel only works with a significant change to the core gameplay or the meta game. They started with a crossword game, followed it with a word game without the cross, and have since spread across genres to reach men and women, older and younger players, and whatever is trending on Sensor Tower.

Tomer plays two or three hours a day across their own games and the competition, with roughly 10% of his time on new games and 20% on existing ones, because a new game only gets a week. Avi plays far fewer games, only the ones he likes, and plays those a lot, Fluffy Doku being the current one, and spends much of his time iterating on the platform.

Every Employee Sees the Numbers
With 40 live games, weekly launches and big campaigns running, Amit asks how a team that small decides where to look. Avi says it is very hard, and that it rests on people taking real ownership of specific domains, ideally the domains that cover the founders' weak spots. The team raises what matters, the founders sit in the dailies, and when something gets ignored, the team nags them.

The other half is transparency. If a developer keeps asking for access to something, Avi would rather build him a tool that refreshes automatically and gives him the freshest picture, so he never has to ask Tomer. The same goes for player data: type in a ticket number and see everything, open any new game and see retention, levels passed, ads and revenue on one timeline, so anyone in the company can raise a flag and change the priorities.

Tomer adds that they have teams in Israel, Ukraine and Poland, and that many companies are afraid to share KPIs and revenue with developers abroad. They learned the opposite at Playtika, and the studio's daily report goes to management and to the newest employee alike, with the daily spend, the daily revenue and the impact of every new game or feature.

"They are connected to the business and they care more."

Building Their Own Incentivized Channel
Neither founder had worked in marketing before. Around late 2024 and early 2025, Tomer took ownership of marketing himself, closed everything that wasn't working and went all in on incentivized traffic, which he describes as math rather than marketing, with no creatives, videos or playables involved.

"So I don't know about marketing, but I know a little bit of math."

At the peak, they worked with more than 60 incent channels, which was its own kind of chaos, and it got the company to profitability for the first time. Avi calls him the best marketing manager in the industry, and Tomer corrects him: junior, he hasn't made it to lead yet.

On incent, Tomer watches CPIs and the funnel more than retention, specifically level churn. If players drop off after level 50 and the next milestone sits at level 100, add a step at level 70 to push more of them to the top.

Then came the ceiling. Ad-based games cannot pay the CPIs that IAP games with whales can afford, and if you can't crack Google, AppLovin or Meta, scale stops somewhere no matter how far you push it. The funding market has also changed, he says: two years ago investors could digest a marketing mix that was 70% to 80% incent, and today that makes it hard to fundraise or do M&A. So the founders took what they knew how to do and built it themselves.

"Let's do our own incentivized channel."

The plan started in August, almost a year before the recording. The structure of the company didn't change, but the need for content did, because an incent app needs a wide variety of games to keep the users it acquires.

"Content is the king."

The Daily Fraud War
Wherever there are real rewards like gift cards, Tomer says, there are scams, fraud and VPN users, and the fraud infrastructure built for games was not made for that. Avi describes it as an ongoing game of cat and mouse, run with his team and the QAs, where every decision has to protect the good players while catching the rest.

His example is the apps that pay people three cents a day to keep them running in the background as a VPN, which is how a user from Bangladesh can show up with a clean Verizon IP address from New York City.

"He looks amazing, but he's not."

Tomer leaves the audience a tip: a Telegram group called PSH Public Group, where he says any game making more than $10,000 a day will probably show up, with people asking for scripts to skip levels or get free coins, and sellers handing out those scripts in exchange for watching ads.

"You watch two hours, you get a script."

They log in, search for their own games and pick up ideas. It doesn't cover everything, but it is another source.

AI Agents With Job Titles
Avi says a strong team protects its own culture, and that when someone who doesn't fit comes in, the team won't accept it. Tomer says the same goes for AI. Mor set up a team of agents and told each one its role, the product guy, the marketing guy, the monetization guy and the R&D manager, and every new release goes to them for a discussion between the bots. Three or four weeks before the recording, Mor told Tomer he had fired one of them, after asking the agents which team member was the least efficient.

"So everybody said the same, so we just kicked him out."

Where They Stop Trusting AI
Tomer sees creative as the easiest place for anyone to start with AI, because it is visual: you write a prompt, look at the image and decide if you like it, while vibe coding still requires understanding how code is structured. The limitations, he says, are Avi's department.

Avi starts with hallucinations, especially in video generation, and says AI cannot go unsupervised yet, comparing it to full self-driving, which has been almost there for over a decade. Some limits will get solved, others come from the nature of a statistical model, so the process has to stay dynamic.

His example is a game that needed levels described as specific images and JSONs. Mor built a flow that worked well, Avi put it inside a loop with three competing prompts and several models, and the first run produced around a thousand levels. A simple Python program put versions A, B and C side by side in one image, so he could see that C was usually best and still cherry-pick the best version of each of the first thousand levels.

Innovation is where he draws the line.

"AI will never have innovative ideas, at least not in its current... Not the LLM."

Avi separates craftsmanship from creativity. He can't draw at all, but with AI he can produce images because he can think, and AI handles the technical part. The creative direction has to come from people, and AI only delivers once you push it there. Amit agrees, and shares that the research team at Sett is working on getting real novelty out of LLMs, a problem with no solution today.

"LLMs are very repeatable. They might sound innovative because they will say it's an innovative idea."

He warns that the model will even announce an idea as fresh out of the oven, which can fool anyone who forgets it is a statistical model completing the next token.

The Rollercoaster
In the card game, Avi picks the question about a small decision that ended up mattering more than expected. On the career side, he was a bored computer science student on reserve duty who shouted that he wanted to start a startup, and someone answered. The startup was an early predictive analysis company in 2003, they raised money and lost it, and he rode his scooter to the office with a desktop computer on his back because he couldn't afford a laptop. On the personal side, it was a random apartment in college, where he met lifelong friends and his wife.

Tomer's card asks about a line he still carries, and it came from his cousin before he joined the army.

"Don't lose it. Never lose it, no matter what happens."

He describes founder life as an emotional rollercoaster, to the point where the two of them send each other a rollercoaster GIF on WhatsApp once a week, and sometimes the ups and downs happen in the same day. What separates good founders, he says, is the relationship between the partners and resilience to extreme changes, without letting work and home bleed into each other. Avi says Tomer's real skill is flattening the rollercoaster in real time, so he can act how he wants instead of acting out of necessity.

Should You Open a Studio Now?
Tomer's first answer is a joke: never. Avi says yes, because entrepreneurship is the biggest turn-on for him and people should pursue their dreams, even though the bar is much higher now, with more regulation and even an iOS account hard to open.

"This is the only way we'll bring in innovation. Definitely not with AI."

Tomer's real answer is about passion. If games are your passion, do it, and if you are coming for the hype, the exits or the money, it won't work in the long term, in games or anywhere else.

On the future of the studio, he says they have been scaling significantly in the months before the recording, and the focus is on growing, building more quality games and enriching the platform. He also has advice for founders going out to raise, that smart money matters, because founders are mostly alone with their problems and need investors who back them on the good days and the bad ones.

"You need a good product, you need guts, and you need good investors in order to keep on pushing and scale high."

Wrap Up
What runs through the whole conversation is that Tomer and Avi build things once and let everything benefit. A fix to the platform reaches every game, a tool built so one developer doesn't have to ask becomes a view anyone in the company can open, and a daily report sent to the whole company makes every employee part of the business.

The same thinking reaches their AI. The tools doing the work that once took 27 artists, the agents that sit in on every release and the level pipeline that produced a thousand levels in one run are all built around a process the founders control, with the creative direction and the final pick kept by people.

For Amit, they are the couple he and Yoni take most of their inspiration from, and Tomer's plan for the future is simple.

"As long as we are having fun, we are here."

---

The Full Transcript:

Amit: Oh, hello, guys. Uh, so today we have Tomer and Avi from, uh, Brain Games. They are the co-founders of that. They work together in, uh, Playtika, and, uh, they have a vast history of, uh, their career. Uh, Tomer and Avi is also our co-founders role models for Yoni and myself. And yeah, it's great to have you here. Please, uh, tell us about yourself.

Tomer: The old version of you and Yoni.

Amit: The old version.

Tomer: The old version.

Amit: The taller version, let's put it like. So please tell us about yourself.

Tomer: Uh, so thanks for having us. Um, I am Tomer. Um, I've been in the games industry for, uh, almost, uh, 15 years, I think. Started in Playtika. Did variety of, uh, positions in Playtika. Um, over there I made... met, uh, Avi. Uh, we are like the married couple. And been together for a long time now, and, uh, I'm married to Anat. I have three daughters. Um, and we founded, uh, Brain Games, uh, seven years ago.

Amit: Wow.

Tomer: Yeah.

Amit: Seven years ago, huh?

Tomer: Seven years ago.

Amit: Wow.

Tomer: Before COVID.

Amit: Before COVID.

Tomer: Think about it.

Amit: But COVID probably was insane, no? But it was a crazy time for games.

Tomer: Yes. The hype was, uh-

Amit: The hype was good.

Tomer: Yeah, it was.

Amit: Avi.

Avi: Uh, I'm, uh... Avi, I'm, uh, more on the technical side. I've been, uh, in the games industry, um, as, as long as Tomer has. With, uh-

Amit: He's older.

Avi: But, uh, before that, we actually, uh, I was in a startup that did the, uh, ads in Facebook Canvas games.

Amit: Really?

Avi: Yeah. And, uh, that was, uh, also, uh, intense. That's actually how I got to the world of, uh, web and, you know, all of the, uh, more fun stuff, I would say. Um, I, I was a system architect, planned a lot of systems. Uh, I like game simulation, game modeling, uh, doing things like this. Um, and, uh, I managed, uh, R&D teams and, uh, did a lot of, uh, a lot of, uh, other stuff. I'm very interested basically in anything I see. So I like to investigate anything that I see. Um, uh, I think my last position was very interesting in Playtika. We did the due diligence for, uh, uh, Super Treat, uh, which is Solitaire Grand Harvest. And then, uh-

Amit: Mm

Avi: ... it was a very interesting deal. We, uh, uh, Playtika a-acquired, uh, the company, but we actually had to transfer the knowledge to us, was very interesting. We... I was in Austria for, like, three months, and-

Amit: Austria, it's an, uh, Austrian company?

Avi: Yes.

Amit: Wow.

Avi: And, and, uh, we built a, a team. We had to transfer all the knowledge from the Austrians to us. It was very interesting. And then, uh, me and Tomer, we planned, uh, our, uh, uh, chapter together.

Amit: Chapter two. Chapter two.

Avi: We have... For 14 years, we have our daily together every day, 9:00 a.m. we go to the office, even in Playtika and also over here. So we have one hour drive. We have our daily call every day- ... every thing that we plan how we can conquer the world.

Amit: Yes.

Avi: What we can do, the whole-

Amit: Exactly. Okay, let's speak now about how you're conquering the world. So you're very small team. Uh, you have, I, I'm not, uh, going to say the numbers, but, but more than, uh, like six figures daily spend and revenue currently, something like that. Uh, and very, very small team. Can you a bit tell us about how you can operate in such a large scale with small team?

Tomer: Yeah, I think today, uh, one is it's a necessity. Um, today- ... you don't need, with all of the things happening in the world, uh, with AI and how it reflects on the general industry, especially in the games, um, you don't need so many people. Um, you operate... When you're a smaller team, generally you operate much faster, you are much more agile. You can react fast to changes and adopt things, uh, much faster than company, which is already, you know, a big company and enter-enterprise. Um, and, um, we've worked very hard in the past seven years to build the proper infrastructure to actually do this. And, and, um, we have a very, very, very good team, which I can say that the biggest thing, uh, about our team is that, uh, they are super committed. They care about-

Amit: When you said... Sorry for interrupting, but you said the proper infrastructure. What is the proper infrastructure for, um, like operating in large scale with small team?

Tomer: Um, it's a variety of things. One, for us specifically, since we are generating lots of games, uh, you need to build the proper infrastructure in order to deliver, deliver all of these games. And you need to have the right infrastructure to support it. And you need to have the right pipeline for content creation. And you need to have, um, the right tools for each and every one of the guys to get the 500% from the person and not the 80% from the person. Um, so it's an ongoing process that you just keep on, um, improving the team, the tools that the team have, and the infrastructure, which can be the creative infrastructure or the technical infrastructure or the BI infrastructure, uh, and... or the process is also part of the infrastructure we have, different processes that we are doing. Um, and that's something that, you know, uh, today with AI It's becoming more extreme. So the good guys are becoming superheroes, and the less good guys which are not utilizing AI's tools are becoming not relevant.

Avi: Right.

Tomer: Um, so the gap in between a super, a professional good, uh, team member versus one that, um, cannot stand and cannot, you know, keep the pace.

Avi: I would say, I would say that, you know, uh, a lot of it is also, uh, um, related to motivation. Meaning that if you're very, uh... if, if you have high motivation, you're just as good as the smart guy because, uh, the AI took away the technicality, you know, uh, of it. Just like in drawing, it's the same for coding, I would say. So you can actually produce things. So as long as you're motivated and you check what you're doing, you can actually be very, very productive. Um, but as Tomer said, if you're, if you're, uh, into excellence then, then, you know, you can, you can reach five hundred percent of what you're doing.

Tomer: Yeah.

Avi: And a lot of it is also about how do you, uh, uh, manage your process? How do you do, you know, risk management? You know, h-how do you do all of this together in order to, to achieve a very, very efficient, uh, process? So, uh, I think, I think, you know, that, that it, it's something that you... it grows with you kind of. You know, it's not a, it's not that you sit-- We don't believe in-- And this is something we didn't-- You know, we, we understood from day one we wanted to have multiple games. We wanted to have a platform. We wanted to have all of this. And we understood that you cannot just develop a games platform. It doesn't work like this. You, you, you know, you know, it's, it's what you imagine versus the reality, and the, and the, the, the gap is huge. So what you do is you just do that. You know, you develop a game with a platform in mind, and, and then you evolve it into something that is, you know, much more, uh, uh, much better. And then, of course, you introduce a second game, so your platform has issues. And the third game, so it has a little bit less issues. Uh, uh, but then it grows with you.

Tomer: It's like an iterative, uh, learning process also.

Avi: Yes, definitely.

Tomer: Yeah. Okay.

Amit: And what positions are a must for you? Like you, you have you two, of course, that you are, you know, hands-on and the leaders and amazing and blah, blah, blah. But what position in the studio itself you, you're saying it's a must to have, and it's the minimum to operate in such a large scale? Uh, yeah. On which areas, like, I mean.

Tomer: Um, so I think that the trivial in every company, like you have, you have the R&D guys, uh, which is a, a client developer and a backend developer, or depends on what you have. You can have a full stack developer instead. Uh, you need to have the QAs, uh, although the developers are doing QA to themselves, but you still need to, you know, to cut down the pipeline of how do you build, uh, the APK, the version, and how you push the production, then you're doing QA for that. And you ne- uh, we have also the BI side, which is the infrastructure of the data and the analyst to have a look on the data and to, to give you the inputs and help you to get to the right decisions. Like looking on the data, everybody today are like data-driven, and this is a must for any company. If you own a games industry or if, uh, you own supermarket, uh, if you don't look on data, you will never be able to improve yourself. Um, we have, you know, monetization and product, uh-

Amit: The same, uh, same person?

Tomer: Yes, I am the monetization and he's the product. Uh, so, uh-

Amit: No, but you have another-- Like you have person in the company that is, uh, doing this job or not? It's only you.

Tomer: A monetization product, I would say it's, uh, Elias.

Amit: Mm-hmm.

Tomer: Um, uh, we, we have, uh, one which in, which in... can, can call her also monetization and, uh, um, creative, which is very important when you are creating content.

Amit: Mm-hmm.

Tomer: Um, I think this is the, the minimum. So if you are just starting a company, I would say five people is, uh-

Amit: Wow. And designers, I'm mostly speaking about, uh, the game itself. Art director, you know, those stuff. Is it, uh, crucial these days?

Tomer: I think-

Amit: For your type of games also, I know it's different from, uh, other games.

Avi: It's crucial, and, uh, we have an amazing, uh... Our VP creative is-

Amit: Ah.

Avi: He's amazing.

Amit: Oh, oh.

Avi: He's sitting-

Amit: Best of the best.

Tomer: He's a ninja.

Avi: And, you know, he sits on the edge of, of, of everything, basically, of AI and everything. And we, we collaborate a lot. We can talk about it later. We, we collaborate a lot and, uh, he's doing a amazing, uh, really a, an amazing job. I wanna go back to what you said about what-- the question about the scale and a little bit the, about the positions. I would say that there is a lot of, uh, uh, part of it of, or part of, uh, allowing this to happen is the fact that you have people that can take ownership on specific domains. Okay? Real ownership. And you need these people to be, uh, people that, uh, take your weak spots. So you need to understand, you know, uh, what, what works for us might not work for you because of different type of management, because different type of, you know, behavior or, or what your strengths are. And, um, and then on the other hand, what you need to give is kind of transparency to everybody. And this is, I think, a lot easier with AI. Meaning that if you're, I don't know, if you're a developer and you need access to something, make him a tool that he can do it. He doesn't need to ask Tomer for it. You know, if he asks f-a few times, "Wait, wait, wait, we have a problem," let's resolve it for him so he can see it. So, you know, have like a something that automatically refreshes, gives him the freshest copy of what's going on and, and involve him in the, in the not just, you know, the, what he's doing, but allow him to reach whatever he wants.

Amit: Mm-hmm.

Avi: So transparency in the organization, I think, is something that is very, very important.

Tomer: Yeah. Lots of the companies are, like afraid. Our team, we have a team in Israel, and we have a team in Ukraine and team in Poland. Um, and I think that lots of companies are afraid to share the KPIs, uh, or the amount of money they are making with, uh, Eastern European, uh, developers.

Amit: Yeah, like not local, uh, employees.

Tomer: Um, we believe otherwise. Like actually, I think that we learn it in, in Playtika.

Amit: Yes.

Tomer: Um, it was like very transparent. Everybody knew what's happening, what's going on. And also in our company, they have a daily report which is coming to the management and coming to the last employee also in the company. They see the daily spend, they see the daily revenue, they see the impact of a new game that they are releasing-

Amit: Mm

Tomer: ... or the impact of a new feature that we are doing. So-

Amit: They are connected to the business.

Tomer: They are connected to the business and they care more. I think this is the main-

Amit: Then they work more

Tomer: ... way-

Amit: And they want to be more creative. No, no-

Tomer: Yeah

Amit: ... I understand you. It's amazing. Okay, amazing. Uh, let's speak about, uh, Workmore and stuff that you do very uniquely. I think every time, not that I'm coming to your office, but, uh, we have also, for people who doesn't know, we have mutual investor, both Brain Games and Sett, and, um, like Bessemer and et cetera. And every time I speak with anyone that invested in, uh, Brain Games or that I meet you, you show me a new game. You did every type- ... of mechanic that I ever saw. You just told me before we came inside that you did, uh, two RPG games now. Eh, I have few question about game launches. First of all, how you build this system of launching new games that fast on, I don't know, weekly, bi-weekly basis. And, uh, after that, I wanted to ask like how do you choose the, the next game to develop? But let's start with how you build this really infrastructure but for launching new games like crazy.

Tomer: So the best way to, to work fast like this is to start as fast as you can and increase the pace- ... when you are going forward. Uh, no, literally. So as we said, we, we've worked hard in the first three years to build a, a proper infrastructure. If you think about it for a second, most of the things in the games industry are very, very similar. Daily bonus is a daily bonus. You log in every day, you get something, no matter if you're a sniper-

Amit: Yeah

Tomer: ... shooter-

Amit: Okay

Tomer: ... or you are a world game. And, uh, ad monetization system is the same, and segmentation is the same, and payment system is the same, and daily missions is the same. So it's, it's mainly about count- counters of something and g- and giving reward for things. And the main thing that we are doing is developing new core gameplay mechanics. So it's either you can, you can acquire a simple code in Unity Asset Store and then, and then to rebuild on top of that, uh, or we just develop, develop, uh, from scratch. But when you have the right infrastructure, it's very, very easy to release. Uh, not very easy. We work hard. And, uh, but it makes it a lot more simple, um, uh, to deliver new games. Um, and, and we are heavy consumers of AI. Uh, as Avi said about, uh, Mor or VP Creative, I think that m- one of the most unique things about Mor specifically is that he's not as... He's an artist, which is not, uh, technophobic. Most of the artists in the world like their art and like the creation, and a little bit afraid. Like, look what happened when, when, uh, they introduced Wacom 15 years ago. Um, so he is like, um, doing everything with AI. The top four employees of the company is Claude, Nano Banana, GPT, and Veo 3. Uh, so, uh, and think about it also, i- if you... And we don't have artists in the company today. We have zero artists in the company. We used to have 27 artists five years ago-

Amit: Wow

Tomer: ... which created content for a single game. Today, we have more than 40 games on production, and we are releasing one game per week. Uh, and we have zero artists. And we are doing everything with AI. So it's a hard work. We've built it, uh, a unique tools that utilizes, uh, AI. But if you think about it, 10 years ago, when you wanted to have a character for your, uh, a game or a theme for a game, you would spend three, four days for a single artist to create a mock-up for a single, uh, character. And you used to have like four of them to create different variations, and then you need to choose. And then after you chose, you need to go and to polish that. Today, if you want to have a theme for a game, you just write a prompt, "This is my game. I want five, se- seven, 10 different themes," and you can just literally choose. And-

Amit: Do you, do you remember the... In the, in the first game, we had Word Farm Adventure.

Tomer: Yes.

Amit: We had a, we had a video scene. We, uh, we took it off really fast, but we had a video scene when you start and you see, you, you know, you're driving your car and then you cra- or crash or something, and, and, and then you go into this farm, and this is how it starts. We... It took us... We paid, I think, thousands of dollars for this. We argued so much about we wanted the hands to be unisex. We didn't want, you know. And it was, it was crazy. And then once we put it on, we realized that it's completely killing the, you know- ... the, the first time user experience. So we took it off immediately. It was up for like a week, and I think this is like, uh-

Tomer: Yeah, it took us like, uh, two months I think- ... with 20 iterations with the-

Amit: I love it

Tomer: ... video creator. Today, you can tell Sett, "I want it like this."

Amit: Yeah.

Tomer: After two hours, you have a video. Uh, yeah. So we created a proper platform that helped us to easily create more games.

Amit: How easy is it to create this platform? Like let's say I'm a new company today, I'm taking the genre of, you know, of ads. So-

Tomer: If you don't have Avi, it's impossible.

Amit: No.

Tomer: If you have-

Amit: Avi and GPU

Tomer: ... Avi and GPU.

Amit: Sam says. Sam says.

Avi: I would say, I would say, um, I would say that, uh, you need to build it, as I said before, alongside a game and to understand-

Amit: Like not starting with building infrastructure.

Avi: No, because you cannot imagine it. Our, our, our mind is, is way too limited to imagine what you actually need.

Amit: Mm.

Avi: So you just need to do it side by side, and then to develop it, and then to basically understand in, in each part of the way what are the weaknesses and how do you handle them. And you know, you know, like a year ago, I would say that it would take us like, uh, a five to seven days to put an environment up. Today, it takes I don't know, 45 minutes. So, you know, so you have to, you have to understand what are your weaknesses, where do you wanna... wh- h-how do you release also your mind into other things, you know? Once, once you resolve something like this, so suddenly your mind can work on something else.

Amit: Mm.

Avi: And, and, and so this is how you do it, I think.

Amit: Amazing. And how do you today choose, uh, the game genre mechanic that you're going to launch? If you're doing it every week, you need consistent ideation, I assume-

Tomer: Yeah

Amit: ... consistent theses.

Tomer: You know what are the next games that we are releasing next week?

Amit: What?

Tomer: I don't know. But we will do some. Oh, there you are. We, we play a lot. Every founder in the games industry have to play a lot, right? My kids like my phone so much because I have 300 games on it. Uh, when, when I bought the phone in the phone company, I said, "I want, uh, 512, uh, gig, uh, memory."

Amit: Ah.

Tomer: And they said, "Why do you need it? You have the cloud." I said, "No, I need it for games." Uh, so you play a lot. And, and think about it, today the organic factor in games is super low.

Amit: Yeah.

Tomer: Because nobody goes to the store and search for something like this.

Amit: Okay.

Tomer: You just consume it from the ad that you are playing.

Amit: Mm.

Tomer: So we play a lot of game. We play a lot our games. Um, and we have a variety of games to play. And we just consume the ads. A-And we look on the ads and, you know, if there is a specific ad that you like, you'll do that game. If you-- if there is a specific ad that you see a lot, it means that this game is probably popular, and you will-

Amit: Ah, so the main, uh, your... One of the main sources of ideation is actually the ads that are in the game?

Tomer: Yeah. So the mechanics are coming from the ads. We have playable ga-

Amit: Wow, okay

Tomer: ... playable ads, for example. Playable are super misleading. They are not related to the core gameplay itself.

Amit: I know.

Tomer: But they are fun. Lots of people like to play the playable more than they like to play the game itself. So you ha- you find a, a nice playable game, say, "Let's do it."

Avi: And, and also whoever, whoever... Think about it, whoever is advertising has a viable business, so-

Amit: Ah.

Avi: You know, so-

Amit: So you're saying it's working, he's advertising that.

Avi: Yeah.

Amit: I see it a lot.

Avi: Absolutely.

Amit: Looks fun, let's do it.

Avi: I, I, I, I actually, uh, built a bot, uh, uh, like an agent that will give a CI and, uh, and it's making this website and then, uh... But it only got me to play games. My, my kids are making fun of me. I'm making all kinds of, uh, money-making ga- I don't know, all kinds of crazy games. So, so far it's a failure. Maybe if it's, uh-

Tomer: Maybe it will evolve.

Avi: Maybe it'll work in the future, yes.

Tomer: And, and, and for our-

Avi: I will keep trying to improve that

Tomer: ... uh, constellation, we do need a variety of games. So to clone the same word game over, and over, and over again with different variations is not something that's relevant for what, what we are currently doing, so-

Amit: It's not relevant, you're saying? To do the same-

Tomer: For us, no

Amit: ... iteration of the same, uh, game?

Tomer: Um, if you find-

Amit: The same mechanic

Tomer: ... if you find a business value for it for your own company, so it might work. Um, but it's not like, um, you know, Marvel releasing, uh, the Avengers 1, Avengers 2, Avengers 3-

Amit: Yeah. Okay

Tomer: ... Avengers 4. You need to have a significant, uh, change in either the core gameplay or the meta game in order to make the second, uh, variation of your first game better. Um, and there is a limitation of how much you can innovate in a specific game.

Amit: Yeah.

Tomer: So you can clone a game one time. You can clone it twice. But 20 of the same, uh, genre is, is not relevant. And also, we want to target a broader audience. So if we'll have only word games, for example, t- we started with a word game, and then we cloned it to a different word game.

Amit: Yeah.

Tomer: First one was crossword, the second one was word list without the cross. So it's different, a little bit different mechanic. But let's say you have, uh... You don't like word games. You don't like word games, it's okay. You can play something else. Uh, so we are really trying to get to a much broader audience. Um, uh, males, females, older, younger, uh, and follow trends. You know, you can look on Sensor Tower.

Amit: But how do you focus? This is, uh, also a fo- a question for me as a founder. You have 40 games. I assume you have top ones, no? Like, ones that are working, uh, better, I assume.

Tomer: Yes.

Amit: How do you put your focus on between... With such a small team, Ken, just to give context, it's like between 10 and 20 people. It's insane. You have 40 games, you're spending a lot of money, you're gaining a lot of money. How do you really have... I don't know. How can you choose where you put your eyes on when you have a week, every week new games that you launch, and 40 games that are live, and huge campaigns that are running?

Avi: It's very hard, first of all. No, it's, it's hard.

Amit: No, it's insane. It's ins-

Avi: It's super hard. The second thing is, I think that, uh, again, once, uh, you have a good team and you have a good relationship with your team, so important things, uh, come up and, and we're very involved. We, we're on dailies. We're, we're, we're super involved. And so, so our team, you know, uh, brings to us a lot of times, uh, things that are, they think are important, and we trust them, and if not, they nag us. So, so they just do it, you know? And, and, uh, this is one thing. And the second thing, you know, it's the million dollar question, to get, you know, what, what is important out of life versus what is less important, and to take that important part and do it. You know, this is, uh, I think the, the question of general success and-

Amit: And if you-

Tomer: I would, I would add one more thing to that, um, which is Avi doing much more than me. We listen to the users. So Avi is reading every support ticket that we are getting.

Amit: Mm.

Tomer: He doesn't sleep, he doesn't eat. He doesn't go to the toilet.

Amit: I... You can see.

Tomer: But I think that, you know, uh, users know games, and user know, like, if you have bugs, in most of the companies, no matter how many QAs you will have, you will find lots of bugs coming from your support ticket saying, "I got to this level, it's stuck. I didn't get my reward for this." Uh, and you do need to filter a little bit of the noise that is happening from the support. But, uh, but when you, you, you understand the user and you understand what the user, what's the pain points of the user and, and, uh- Even if it's not a bug, even if it's why can't I have more free So when you listen to the users, uh, you know where to add it. And, and since most of the thing that we are adding or fixing is part of the platform, whenever we are releasing a new feature, it automatically applies to all of them.

Amit: Mm-hmm.

Tomer: I-I'll tell you like a simple, uh, simple example. Three years ago, there was a big hype on Wordle, which was acquired by-

Amit: Mm-hmm

Tomer: ... the New York Times. And so we said, "Okay, it looks like a good marketing trend. Let's, uh, we have a Word game. Let's add Wordle as a mini-game inside the Word game. And then we'll have a pitch for the marketing. The marketing CPIs will go down because it's very hyped." So we added Wordle inside the game. We put it in the marketing. It didn't really take So the experiment failed. But today we have technically Wordle in all of the games. We have, uh, uh, i- ad-based slots games with Wordle inside. It's turned off because we don't need it in the slots, but it's there because it's part of the platform.

Amit: Like you build a huge, uh, feature release, uh, platform that you can do... Basically, what-whatever you do, it's repeatable.

Tomer: Exactly.

Amit: It's the-

Tomer: Yes

Amit: ... actually, it's something that we work on at the company. We're in a stage at the company that it's, we're scaling. It's, uh... I know it's not complete chaos, but it's a lot of chaos, and I'm all the time saying to the team and to myself mostly, when I'm saying it to them, I'm saying it to myself, that we sh- we should only do stuff that we can, first of all, learn from, but mo- like mostly, uh, repeat. Like if we're doing something that is not repeatable, meaning that it's not scalable, meaning that it's not relevant for our stakes. Uh, and you took it to the ma- to the, to the other level, that everything that you do, you can really replicate it to all of the game. Uh, in high level, it's very high level. I know it's not, uh, like you don't have the real answer, but in high level, how much, we will ask each one of you separately, but how much of your time you're spending on, uh, in percentages on the new games and the existing ones?

Tomer: Um, look, I'm not playing games. This is my excuse for my wife. I'm checking competitors- ... or doing QA, so I'm never playing games. Uh, but I would say I play at least two, three hours a day in our games, the new ones, the existing ones, a-and external games also. Um, so I would say I invest like 10% of the time on the new games, 10% or 20% of the time on existing games. Uh, because, you know, a new game, since the pace is so hi- so fast, a new game is a week, so we cannot spend more than-

Amit: Yeah

Tomer: ... a week. So-

Avi: And he's on, he's on the wall in, uh, Royal Match.

Amit: Yeah?

Tomer: Royal Match is the best game.

Amit: Yeah.

Tomer: It's the best game. They are not releasing enough content. Uh, so I would say 10% on the new games, 20% on the existing games, and the rest is-

Amit: The marketing-

Tomer: Yeah

Amit: ... and all of the fun stuff. Avi?

Avi: My, my approach is, uh, I think, I think, uh, is a little bit different. First of all, in terms of games, uh, I don't play all games.

Tomer: He didn't play the last three games we released.

Avi: At least.

Amit: The best thing. Really?

Avi: I... But games that I like- No, but you need to understand. So games that I like, I will play.

Tomer: Games of production, but he's-

Avi: I'm not, I'm more like a, a... I don't know how to call it. Like, uh, I, you know, I, I let my, uh, myself, you know, go with the game all the way. You know what I mean?

Amit: Say what?

Avi: It's intimate, intimacy. No, but games that I like, I'll play and, and, and, but I... and I'll play them a lot. And-

Amit: What is your favorite game?

Avi: Right now? My favorite game is, uh, um-

Tomer: Find the Dogs by Playrix.

Avi: No. Fluffy Doku.

Tomer: Fluffy Doku.

Avi: And I play it a lot. I think it's, uh... And, and it makes me feel good with myself. I really like-

Amit: But you say they can cut it and do it in marketing. Put that like Fluffy-

Tomer: Fluffy is a nice example because it's like a replicate of Queens from LinkedIn, uh, which is like a Sudoku with more level, uh, levels inside.

Avi: Yeah. So, so I really like it. Um-

Tomer: Because of the character.

Avi: And also, and also regarding your question- ... I wanna, I wanna, uh, also answer your question. Uh, in terms of the time spent, um, so we... I spend a lot on the new games because you have to, but we also spend a lot of times in, in, in iterating on the platform, okay?

Amit: Mm.

Avi: And maybe this is also going... You know, imagine, imagine again that your QA has, uh, uh, sees an issue or, or somebody sees a support ticket. How long does it take you to examine this player? You know? It can take you a long time. It can be like, "Oh, let's open a ticket to SLS," or it can be just you type the ticket number, and you see everything. So we try to reach that, and we, we improve this all the time also for new games, you know? So whenever you have a new game, you will immediately see the retention, this, that, you know. You check it on a timeline. Which timeline? The, you know, the, the retention, the levels, you know, the, the levels passed, the ads, the revenue, you know. You, you... If, if it's very easy and it's a very common thing and you unify it and you allow anyone to look at it, anyone can also raise a flag within the company there is an issue and actually, like, change your priority and, and, you know-

Amit: Mm-hmm

Avi: ... if it's important enough.

Amit: Amazing. Okay. Let's speak about your unique strategy, and you can share whatever you want to share because your competitive edge. But, uh, I don't, I don't remember when, maybe a year, two years ago, I don't remember when, uh, you started a pretty unique strategy that not a lot of companies are doing, some are doing, but not a lot, about having your own, uh, app for incent, uh, to incent users and, you know, to get them into your game. And then you have this, uh, you're improving basically the margin of the ads, like ad-based games. So can you maybe tell us, like, completely from scratch about this strategy? Why did you choose it? How it looks like? Whatever you can share, of course. And, uh, what are your main learnings so far?

Tomer: Yeah. So, uh, I can tell you that both me and Avi did variety of, of positions in our career history. None of it was marketing. So we don't know much about marketing.

Amit: This is the best, I think- ... because it's enable you to be creative. No, really.

Tomer: And, uh, and, um, about a y- two years ago, late '24, early '25, um, we got to a point that, you know, like every other company, you have, uh, highs and you have downs and, and we had to do, like, a restructuring of the team. So, uh, I took the ownership of the marketing myself and then, um, just literally closed everything that didn't work and started doing only incentivized marketing, um, which is not good, I can say. But, um, incentivized marketing is not marketing from a UA manager perspective. This is pure math, uh, 'cause you don't have... not creatives and not videos and not playables. So you just work with a external, uh, company. There are plenty of them out there. Some are good, some are less good, some are super big, some are small. Uh, on the climax, we worked with 60 plus different incent channels-

Amit: Wow

Tomer: ... um, which was a disaster on one hand, but it brought us to a place that the company got to, uh, um, to become a b- a profitable company, uh, for the first time. Uh, and again-

Amit: Woo!

Tomer: ... this is math. This is math. So-

Avi: For the record, he's the best marketing manager-

Tomer: Junior

Amit: Yeah.

Avi: ... in the, in the-

Tomer: Junior. I haven't made it to lead yet, but-

Avi: In the industry, okay?

Tomer: I'm, uh- So I don't know about marketing, but I know a little bit of math. And when you're doing incentivized channels, it's only about the CPIs and the funnel that you're giving to the users, and you theoretically don't need to look on retention by doing so. You need to look on, uh, levels churn, I would say.

Amit: Mm-hmm.

Tomer: And then to understand how much revenue you are generating per X levels and when they churn you, and then let's say you have level 50 and level 100, and you have a big churn after level 50, put a step in the middle. Put level 70. So that will-

Amit: Mm

Tomer: ... push a little bit more users to the top. Now, when we understood that we are limited on, on, uh, scale mainly because we cannot compete with CPIs of IAB-based game... IAP-based games. Uh, IAP-based games can pay higher CPIs because they have whales. The return-

Amit: Mm-hmm

Tomer: ... over there looks different than ad-based games. So we understood that we are super limited by scale, and it's something that is not interesting, uh, for us as founders. We want to grow. We want to build something very big. And, uh, then we understood that if we cannot crack Google, AppLovin, or Meta, um, it's something which is not scalable no matter what you get to, $1 million monthly, $2 million months. It's still... There is a limit to that. And more than that, I can tell you that in the general industry, companies today and VCs are looking differently than what they used to look two years ago. So two years ago, if you had 70, 80% incent, uh, traffic on your marketing was something that people could digest. Today-

Amit: Mm

Tomer: ... it will be hard for companies that are spending 70, 80% of the marketing spend on incent no matter what the scale. It will be hard to fundraise. It will be hard to do M&A for that company. Um, so once we understood that we are limited on the scale and limited on, on the interest of us as founders, and we said, "Okay, we don't know how to do regular UA marketing on Google, but we know how to do the, um, incentivized, uh, channels. Let's do our own incentivized channel." Uh, so this is how we came up with the plan to build it. It was last August, by the way.

Amit: Wow.

Tomer: Uh, yeah. So it's a year now.

Amit: Wow. Almost a year.

Tomer: Almost a year. Almost a year.

Amit: And the scale is insane. We won't share it, but it's insane.

Tomer: Scale is-

Amit: No, it's insane.

Tomer: Uh, if you ask our investors, which are the best in the world, the scale is not high enough. To get higher-

Amit: They want more, more, more.

Tomer: Yeah. Um, and they are right, by the way. They are always-

Amit: It's a good indicator that it's-

Tomer: It's a good indicator

Amit: ... going to the right direction when they want more.

Tomer: Um, a- and there are lots of issues doing that. Uh, Avi is fighting- ... on a daily perspective w-with frauds. And-

Amit: Mm

Tomer: ... whenever you are having a place that you are giving real rewards, like a gift card or whatever it can be, you will have the scams, you will have the frauds, you will have the VPN users-

Amit: Mm

Tomer: ... you will have all of these things and, uh, so it's a headache. Uh, you will need to build a good infrastructure, which we didn't have because in the games we didn't care about all of this, uh, thing. We had things for fraud, but if you are not giving away, uh, gift cards, it's, it's different. So you have to rebuild the infrastructure or adjust it to something new.

Amit: Yeah.

Tomer: Avi can say how Avi is working on that.

Avi: It's also an ongoing process. It's a- ... cat and mouse chase. And, uh, you have to be very careful because you don't wanna hurt your good players. You know, you have people that are, you know... you, you're basically, uh, giving, uh, or having a very symbiotic relationship with, and there is others that you don't. So you have to kind of, uh, pick the, the right ones and the right tools. And, uh, and again, and it's, it's a lot of work. And, uh, as Tomer said, I, I'm working on it on the development side, but it's also I'm very involved. So I do it with my, uh, again, amazing team, you know. And, uh, uh, in our company, for example, it's the k- the QAs are doing it. Uh, uh, so we do it, uh, together. I try to get on, you know, these calls and to discuss all kinds of situations. And your decision, you know, is you in front of numbers and indicators, and you have to make a very, very, uh- ... delicate decision. And, uh-

Amit: You have me.

Tomer: I can-

Amit: Yeah, you have-

Tomer: I can give a tip to all of the viewers out there. There is a group in, in Telegram- ... which is called PSH Public Group.

Amit: PSH Public Group.

Tomer: PSH. Write it down. Link, link in the bio. So, uh, this group is the biggest fraudists- ... group chat in the world I guess that all of any game that generates more than $10,000 daily will probably be there, and you will see people are requesting, um, how can I, uh, jump levels? How can I get free coins?

Amit: Yes.

Tomer: How can I... Uh, and there are lots of, you know, uh, hackers out there that just give you literally any company, log into this PSH Public Group, search for the name of your game, and you will see people requesting scripts for hacking your games-

Amit: Wow.

Avi: Yeah

Amit: Wow.

Tomer: ... these scripts, how to. A-and so we-

Avi: They're very creative people. They're good.

Tomer: Very creative. And by the way, they sell these scripts. You watch two hours, you get a script.

Amit: Oh my God.

Tomer: So they sell it by ads.

Avi: Yes.

Tomer: Uh, and, and we log into this-

Amit: That is genius.

Tomer: We log into this, and then we just, uh, get some ideas. It doesn't cover everything for sure, but it's another source-

Amit: Do you have any softwares that are helping with that? You had, uh-

Tomer: Avi is the agents

Amit: ... in the past a cyber... I think so

Tomer: Avi is the agents.

Amit: You have a, a cyber company, I don't remember its name, for, for the marketing, like for, for bots and et cetera. Do you have any software that you use or it's only you, like, trying, you know, to understand who is consuming-

Tomer: Like the, the, the party company?

Avi: No, so we, we, we use some, we use some things, uh, again, but the, the, the real problem is that, uh, um, people fake anything. So I think that-

Amit: Yeah

Avi: ... uh, you know, your solution with some general indicators just not to make it very easy on one hand. On the other hand, a lot of it is actually custom, custom rules. You have a very big problem today, uh, you know, with the identifying like VPNs and things like this. You have, uh, uh, companies that, uh, offered, even offered us, you know, they offer you three cents a day, and you keep your app in the background, and it becomes a VPN. So, so you can actually... You will see somebody who is from Bangladesh but has a Verizon, uh, IP address from New York City. You know, a very clean one. He looks amazing, but he's not. And, and there are a lot of ways that, of course, I will not disclose here, but- ... a lot of ways to-

Tomer: To cheat.

Avi: To, to, to cheat, but also, but also, uh, to, to catch, uh, uh, a lot of this. Uh, but as I said, it's an ongoing, uh, thing.

Amit: Amazing. And, uh, since you did this strategy, uh, how your company changed?

Tomer: Uh, I think it didn't change much in the structure. Like we didn't hire more people. We just made the processes a little bit more efficient. Uh, we did got to a point that we released more games because if you are having your own internal offer wall app or incentivized channel app, you cannot have two games.

Amit: Yeah, you need a lot.

Tomer: You need a variety of content. So think about in the games industry and everywhere, content is the king. You cannot be without content, and if you are a single game and you have a huge, huge amount of users, you just need to give them more and more content. As Avi said, "Royal Match, give us more content." Uh, and so we just need to release more levels and more mechanics and albums and, uh, leaderboards and things like that. Um, when you are in incent app, content is the games because you need to have a variety of games, and you need to have, you know, if you are acquiring a user for your offer wall and he doesn't like world games, and you have only world games, it's not relevant, so you're losing the user. Um, so it changed a bit of the needs, but not of, of how we work because we, we used to run fast, we still run fast, and we keep on running fast. It's part of the, I think, the DNA of the company. Um-

Avi: I think, I think that once, you know, uh, once you're, uh, you build a good team, then, then your team becomes something that is strong. So if, even if you brought somebody who is not, uh, not as good or sometimes e-even it's not, it's not related to good or not good, doesn't fit in, then the team will not accept that. So, so what I... This is, you know, this is when you know that you're in a, in a, in a good position because you have like a team that is doing what they're, what they need, and then you can run fast and-

Tomer: And by the way, it works also-

Avi: ... as Tomer said here, it's not really... Y-you know, you can have multiple directions, uh, and, and, and they will do it. Yeah.

Tomer: It works also with AI. Mor, our VP creative created a few agents. He told one of them, "You are the product guy, you are the marketing guy-

Amit: Mm.

Tomer: ... you are the monetization guy, you are the R&D manager." And then every new game that we are releasing or every something that we release, you just give it to the team, and they are having the, an AI bot discussion between them. And then like three, four weeks ago, he told me, "I fired somebody." I said, "Who did you fire?"

Amit: Avi.

Tomer: And he said, "I asked the team, the AI agents- ... which one of the team member is the less efficient one. So everybody said the same, so we just kicked him out." We have to apply the fire in bots.

Amit: If we're, uh, speaking about AI, uh, where... Okay, so everyone is using AI. I, I see a lot of Israeli studios, may- like you may be the most, and PeerPlay also is like very heavy on AI, and I know several other studios as well. But you know AI very, very well. You also come from a product and technical background, which like you're more technical oriented, let's say it anyway, um, and you're more open-minded and et cetera, and adopting these tools faster. Where do you see the real limitation in AI currently, and specifically LLMs, that you're saying also to the viewers and listeners here, this is not what you should expect from your opinion, of course, from AI, and don't put your focus on that, and put instead double down on these areas that AI can help you?

Avi: Um...

Tomer: So I think the AI has no limitations because every time I say limitations, they say, "Avi, you have a limitation." So let's solve it. So solve the limitation. Uh, I, I would say one thing, uh, this is a classic question for Avi, but one thing- ... that is simple for the users I don't like to talk. One thing which is classic for the, uh, average Joe to, to do is creative, uh, because- It's easier, it's visualized. Uh, if you write a code in vibe coding, you need to understand, uh, codes, and you need to, to know a little bit how the structure works and how to make it work better and the limitations. Uh, but when you are cre- when you want to create static image, you just write a prompt, put a static image, and then you see if you like it or you don't like it. So think about it as general advertising companies that advertise, uh, Coca-Cola, for example, used to have a Cola on the Everest mountain. You can just write a prompt, "Give me a Cola on the Everest."

Amit: Mm-hmm.

Tomer: So, uh, it's much simpler and easier to utilize for something that you visualize.

Amit: Mm-hmm.

Tomer: Uh, but it has lots of limitations that Avi would like to say a few words.

Amit: I will bring the notebook.

Avi: No, you know, I think you have, uh, you have the obvious, uh, limitations that, that, you know, e-everybody sees, which is, let's say, uh, you know, hallucinations, things like this that come into, uh... You, you definitely see them on, on video generation and, and things like this. Uh, unfortunately, I think that, uh, AI cannot go unsupervised. It's kind of like maybe full ser- self-driving. That is something that has been, you know... It's there for ten years, 12, I don't know how many, but not, but, but not really. And, and, and what happens is that, uh, um, the, the limits are, you know... They're, they're trying to solve it on one hand. On the other hand, there are some things like we discussed before, they're unsolvable because of this is a statistical model and because of, of, of what's happening. And so we kind of try to, you know, work around it. I think that this is, uh, your process of work has to be very dynamic. What you do is on some things we have, uh, we build, uh, pipelines, for example, that are, uh, um... I'll give you, I'll give you an example of a process that, that, that is, I think, maybe, uh, uh, good for this. Um, so we had a game, we needed levels. We wanted very specific images and JSONs that describe these levels. And, uh, Mor, uh, our VP creative, he built, you know, some kind of a flow that, that does something pretty well. And then, uh, we sat together, me, him, and Tomer, on a pretty long call, and basically, uh, he gave me the flow. I took it, I made it inside the loop, and I also, uh, did it with some kind of A/B testing on the prompt. Meaning that we needed... I don't remember. We produced, I think, on the first iteration, like a, a thousand levels.

Amit: Oh.

Avi: And we had three different prompts, you know, with color limitations, with, you know, different, uh, uh, different ways of explaining, uh, uh, to the model how to do it. We basically ran them. Then we had some simple program, you know, in Python that unifies everything into... So you can see like A, B, C. So you can see actually, you know, level one, what each version of the model or... And we also have like different models. So you can actually like, you know, select a different model, and then you have A, B, C, and now you can see it in one image. And, and, and, and now you can say, "Wait. Okay." So I... Right now I can definitely see that, let's say, I don't know, C is the best. So in general, I wanna take C, but on the other hand, I wanna, at least for the first 1,000 levels or something like this, to cherry-pick the best one. But I know that most of them are C, so I'll, I'll, I'll start with that. So this is, this is kind of how you go around the limitation-

Amit: Mm-hmm

Avi: ... of AI, you know? And-

Amit: Mm-hmm. And do you trust AI today for innovative ideas generation?

Avi: A-AI will never have innovative ideas, at least not in its current... Not the LLM. Uh, for me, it seems like, uh, exactly this is where, you know, e-every time when you look, uh, into an AI or into something you wanna do into AI, you need to divide it into the technicality or, or the, the, the craftsmanship that you need to do versus the creativity. You know, I cannot draw... I, I, I'm, I have... I cannot draw a-an image myself. I mean, I'm the worst, okay? And then on the other hand, with AI, I can do it because I can think, right? So I have a lot of thoughts. I, I, I, I can do that. It's the same for coding. In coding, it's just, I think, not good enough yet. And, and I think that it will get better, or it requires you to have much more, uh, I don't know, technical language in order to achieve what you need. Uh, um, but I think that eventually that will get there. And, uh, so, so again, so you need to divide that. But creativity is, is not good. You have to push it towards your creative directions, and then it will give you a lot, but not-

Amit: Mm-hmm

Avi: ... not a lot.

Amit: Okay. So we are aligned. Okay. So, uh, yeah, this is also the areas that we see also in our platform where AI is limited, and there is some ways we are working on it, on the research side now to try to get, uh, like, innovative ideas from the LLM or to understand how to get, uh, novelty from them, which is, uh, like novelty, sorry. Which is pretty complex, uh, because there is no solution for it today. So our research team is working on it. But, yeah, it's a big, uh, big issue.

Tomer: Tell us when you find the answer.

Amit: Yeah, yeah. We will tell you.

Tomer: Okay.

Amit: Don't worry. No, no, it is a complex issue. But yeah, I think that also it is a message for everyone that is listening, that for ideation, no matter maybe it's ads, maybe it's a game, maybe it's a new thesis that you want to test, maybe it's a feature and et cetera, LLMs are very repeatable. Uh, they might sound innovative because they will say it's an innovative idea. Like, they will literally write it Here is my innovative idea that comes out fresh out of the oven. Uh, but it is, uh, like, uh, fools you in the eye. So it's, uh, sometimes m- it's misleading, and I think people that are not understanding that this is a statistical model that in the end just need to complete the next token. Not that I'm a, a big, uh, tech guy that understand it, but the notion of it is pretty easy to understand. Then you can, you know, not expect to really have, uh, great ideas, like really great and creative ideas from that, no matter what you're doing. Okay.

Tomer: It's like I told you before, uh, we started that, uh, if you ask GPT that you want to wash your car, and you have a car wash 50 meters from your house, what is better, to walk there or to drive there? He will tell you, "No, walk there." Until the time that you will, uh, start the engine and look for parking lot.

Amit: Yeah, now I understand that.

Tomer: So it's statistics.

Amit: Let's do, uh, a game.

Tomer: A game.

Amit: A game before the last mes-- uh, question. So you have cards here.

Tomer: Ah, was it a real game that we developed and pushed to the phone?

Amit: Since we started the podcast, two games were launched.

Tomer: Two games were launched.

Amit: Avi is in the phone and closed code like that. Uh, so you have cards there. Sometimes there are some, uh, there are qu- there is a question there in each card. Uh, the question might be, uh, weird, might be great, might be cringe, might be, uh, creative. I don't know what are the question asks. Just choose one. Avi first. Read it.

Tomer: Did LLM wrote, uh-

Amit: LLM, I, I assume that LLM

Tomer: ... who wrote the questions?

Amit: It's the best question. The most innovative question to ask Avi and Tomer. So Avi, you will do one, and then Tomer will do one. All right. Choose the one.

Avi: No, no, I want this one. Was there a small, almost silly decision that ended up mattering way more than you expected in your personal w- life or at work? That's the cringe one? Just tell it.

Amit: Let's do at work.

Avi: A small or silly decision? I don't know. It's a lot of decisions, right? A lot of decisions.

Amit: Yeah.

Avi: Most of the decisions are small.

Amit: But something small that you didn't know the impact of it. Like you said, let's do that, and then you understood it's a goldmine, something like that. Like, this is what I assume the question wants to get.

Avi: Um, I don't know.

Amit: You want another one? Choose another one.

Avi: No, there's just too many. You know, I can give you 20. I don't know.

Amit: Choose one. Choose one.

Avi: On my personal life or on my career life?

Amit: Let's do career life.

Avi: Or which one?

Amit: Career.

Avi: Career life.

Amit: One career life and one personal life.

Avi: Um, so career life, I think, you know, I was, uh, I was a young student in my, uh... And, and I was in, uh, reserves, and then I did what, you know, I thought was right at the time. So I just yelled, you know, to everybody like, "I'm a bored computer science student. I wanna do some startup. Is anyone up for-- Is anybody up for it?" And some guy came up to me.

Amit: Really?

Avi: And then we did the first startup. Yes. In his-

Amit: This is how it went

Tomer: ... in his first startup, Avi used to ride-

Amit: What was the startup?

Tomer: He was-- He used to ride a motorcycle and had a desktop on his back.

Amit: Yes.

Tomer: So a backpack with a desktop to go to office.

Amit: Really?

Avi: I didn't have any money for a laptop, so- ... so I had, uh, I had to ride my scooter with a desktop. Based on a true story.

Amit: But what was the startup? The education thing?

Avi: It, it was actually an AI startup.

Amit: Okay.

Avi: This is in 2000... We were early. This was 2003. Not an AI, like predictive analysis-

Amit: Yeah

Avi: ... uh, using, uh, logistic regression or something like this, and it was very, very, uh, innovative. Uh, I was, uh, I was very young, and, uh, it was fun, you know? Uh, we did some, uh, really, really cool things.

Amit: Did you raise money?

Avi: We raised money.

Amit: Really? Wow.

Avi: And we lost it. But I met some really amazing people there and, you know- ... and there's, but...

Amit: Amazing. You want to hear? I don't know if it's personal. I don't know if it's personal.

Tomer: That was the personal one.

Amit: The personal... Tomer, you want to do... So this was the personal or the career one?

Avi: No, this was career one. This is career. Where, where's the time line?

Amit: So let's do personal one. Then we will move to career.

Avi: The personal one, I think it was, uh, deciding on my apartment in college.

Amit: Really?

Avi: Yes, because that led into, you know, I met great friends. We live together today. I met my wife through that. You know, everything took like a, like a different path, and it was just a random apartment. It was-

Amit: In Yerushalayim.

Avi: Some-

Amit: Oh, yeah.

Avi: Yeah, something that completely changed everything, you know?

Amit: Oh, that's good. Tomer, it's your turn now.

Tomer: Bring me the middle one.

Amit: But you have a high expectation after Avi now. No, it's... You should-

Tomer: I bring the other. Uh, is there a line some-someone said to you once, a parent, a teacher, even a stranger, that you will cry, that you still, uh, carry? That I still carry? Uh, I think that my, actually my, uh, my cousin, before I joined the army a long time ago- ... told me one sentence that I still use today, and I think that's something that very, uh, reflects me. Uh, so the sentence was, "Don't lose it. Never lose it, no matter what happens." So I almost never lose it. It happens sometimes. Going on. But I think that I, as a person, I'm pretty calm. A-and, um, this is something definitely as founder The rollercoaster, the emotional rollercoaster that you are into. You see it yourself. Some days are super good, some days are super bad.

Amit: It's always super.

Tomer: It's always super. It's always extreme. And, uh, you know, in, in WhatsApp you have the GIFs. So once a week we are sending each other a GIF of a rollercoaster, something going up and down- ... and up and down. And sometimes it happens on the same day. You start like super high and then, oh my God, and then everything breaks, and two hours after everything looks good. So, uh, and, uh, I think what, what uniques, um, good founders is, one, the relationship in between the partners, um, which we have the best. And- And the second thing is the, is the resilience for, for the extreme changes. Uh, because this is something that, that can really mess you up. And, um, there is a small, I think, limitation in between the very calm guy and, and the very, you know, somebody that, that doesn't care. So it, it's not that you don't care.

Amit: Mm.

Tomer: You care a lot, but you need to understand and, and you cannot have things from home affecting how you operate in work. And you cannot have things that happens from work affect how you operate at home, 'cause you have the kids, and you have the wife, and they don't know and they don't care that you had a bad day or something is broken. And it's kids, they need their, uh, you know, their attention and so-

Avi: Tomer is right. This is, like the most important thing. And the th- the crazy thing I think that Tomer is really good at is to do it in real time. You know, if I tell you right now, you know, we talk about the time when your, uh, uh, third, uh, grade, uh, I don't know, teacher yelled at you, or your principal yelled at you. You know, back then you were like, "Oh my God, my parents are gonna..." You know? Today you look at it as something like, okay. So reality, when things like this happen-

Tomer: Ah

Avi: ... at that moment you're, you know, th- this is the rollercoaster he's talking about, and this is what happens to you. But if you know how to flatten it in real time and to, to understand it, then you can actually act and not act out of, you know, necessity, but to act however you want. You kind of control-

Tomer: Mm-hmm

Avi: ... reality. And Tomer knows how to do it very, very, very well, you know?

Amit: Amazing. Would you recommend, we will go to the last questions. Would you recommend to someone to, uh, open a gaming studio now?

Tomer: Never. Don't do it.

Amit: Really?

Tomer: What are you, crazy?

Avi: Of course.

Amit: Okay. But let's say someone comes to you tomorrow, consult with you, said, "Hey, I know the gaming market is in, I don't know, uh, have some issues. Fundraising not easy as it used to be," and et cetera, and et cetera, and et cetera. Do you recommend people to open, and why yes or-

Avi: Yes, I recommend people to open a game studio.

Tomer: Me too.

Avi: I'm not like Tomer.

Amit: Why?

Avi: No, I think-

Amit: Why? Why?

Avi: Because I, I, I really think, first of all, I really like, uh, en- entrepreneurship because I, it really, it's the hottest thing. For me, it's like the biggest turn-on. It's like you see it, you know, you, I wanna, we really like to help people. We always help people. Because, again, you know, you pursue your dreams. You know? You got to a point where you can pursue your dream. Why, why, why don't you do that, you know?

Amit: But specifically mobile gaming studio. Let's say someone comes to you.

Avi: Specifically mobile game-- You know, uh, I would, I would say, and I've said this before, uh, to, to someone, you know, "Who am I to tell you not to do it?" You know what I mean? Like, at the end of the day, you know, from my position or my point of view today, the bar is much higher. It's a lot more complicated. You have regulation. It's very hard to even open an, you know, an account on, on, on iOS or, or things like this. It's very, very hard. Like, w- why not, you know? This is, this is the way I see it. This is the only way we'll bring in innovation.

Amit: Yeah.

Avi: Definitely not with AI.

Amit: Tomer?

Tomer: So I don't say don't open or to open. I, I will answer it with a little bit sophisticated answer. If gaming is your passion, do it. If you come here just because there is a big hype, and lots of companies are selling, and there is lots of money over here, this is not the reason to open a game company. If you have the passion for the games and the passion to become an innovate-- you know, entrepreneur and innovative-

Avi: Yeah

Tomer: ... uh, do it. If your passion is somewhere else, do something else. So it's, it's mainly about the personality, and don't pursue it because of a hype or because of money or... 'Cause it, for the long term it will not work.

Amit: But why specifically you're saying that to games?

Tomer: For every territory the same.

Amit: For every... Okay.

Tomer: Yeah.

Amit: Okay.

Tomer: But you ask about games specifically. Should I open a games company? If you have the passion for games, do it.

Amit: Yeah, you need a big passion because it's somehow harder-

Tomer: You have the passion for cyber, go do cyber

Amit: ... who has passion for cyber? It's not possible to have passion for that. Uh, okay, the last question. What is the future of, uh, Brain Games, in your opinion?

Tomer: Now or in 20 minutes from now?

Amit: Now. Now. Not before the startup, not after the, uh, not before the podcast, not after it, but now, now. At this point in time.

Tomer: Specifically now, I will tell you exactly.

Amit: The billions dollar now.

Tomer: No, listen, we, we are having fun. We work together, work close together. Uh, we, um, we learn a lot from our team, from our investors, from each other. And as long as we are having fun, we are, we are trying to build something big. If we will sell the company or IPO the company or close the company and, uh, bank- bankrupt, it's hard to say. Uh, we are on a good pace, and we are in a good position at the moment. Uh, and we are scaling, um, significantly in the past few months. Um, the next plans for us is actually to focus on keep on growing and keep on building more good games and more quality games and, and enrich, uh, our platform and, uh, um, I think that, uh, since we have an amazing team that we really like, and we have the best investors out there, like V Games, Eitan from V Games, and Ariel Bessemer, and, and the whole team are amazing. And when you have, uh, a good team and a good product and, uh, the best backup that you can have, that helps you push it forward You can do amazing things. And as I told you, like, we are spending, uh, lots of money today, and the main question is please spend more of, of our money as investors. Uh, and there are really ama-amazing VCs a-and for, uh, founders that are going to fundraise, I can tell you that the investors are super, super important. It's like there is a, you know, a phrase called smart money versus, uh-

Amit: Yeah

Tomer: ... any money.

Amit: Yeah, yeah.

Tomer: Um, you can get money from, you know, one, one of the people that want to invest is like a constructor building, uh, buildings, uh, like long time ago when we fundraise. Uh, but, but, um, there is no replace for, for a smart company. And when you're getting company from a VC, you need to have a VC that will back you up on the good days, and they will back you up on the bad days 'cause you'll have this and you'll have that.

Amit: You have mostly bad days.

Tomer: Yeah.

Amit: Hello, not you, but in general, no, we all-

Tomer: Yeah, that's okay. We are on our bad days, man

Amit: ... it's 90% good. It's not, uh,

Tomer: And think about it, uh-

Amit: 90% of the time

Tomer: ... as founders, you are pretty much alone. You cannot talk with, uh, your family about the problems anymore 'cause they don't understand it. You cannot, uh, talk with the employees 'cause sometimes they are the problem, sometimes, you know. Um, and, uh, uh, so if you want to build a big company and to have a strategic partner for that, uh, the VCs are super, super important part of it. And, you know, as long as they back you up, you need a good product, you need guts, and you need good investors in order to keep on pushing and, and scale high. Um, and you know, eventually, as long as we are having fun, we are here. And, uh, maybe we'll sell the company, maybe we will not. It's hard to say at the moment, but, um, we are focusing at the moment and mainly on growing.

Amit: Not the journey, not the destination, the companionship, the people that you're with.

Tomer: Yeah. Have fun.

Amit: It's the best way. Yeah, to have fun because we will die one day.

Tomer: Yeah, we are in the games industry.

Amit: Yeah, in the ga- In a fun industry, in a fun way.

Tomer: And we are never playing games. We are taking companies.

Amit: Uh, we are taking companies.

Tomer: We are never-

Amit: You're working. Uh, okay, amazing. Uh, so I wanted to thank you both for, uh, joining this podcast, and, uh-

Tomer: Thank you

Amit: ... again, for, uh, Yoni and myself, you are a... I think Yoni and myself, we are, you know us very, very well, and we're, I don't know, four years together, maybe three years, and we know each other from the, you know, the army since the age of 18. And you are, you know, a couple that we're taking most of the inspiration, uh, from.

Tomer: Our couple.

Amit: Yeah, and for, for us it's the same thing. Like, we don't really care what will happen just for, that we keep being, uh, curious and having fun. That's it. Just keep learning, curious, having fun. And-

Tomer: Think about it. If tomorrow you are selling your company, what are you doing?

Amit: No, me and Yoni-

Tomer: You will not sit and do anything.

Amit: No, Yoni and I need to think about something.

Tomer: You probably doing...

Amit: So, and yeah, the best, and I think not the only way because you have solo founders and you have more than... But, uh, if you're, uh, two very, very good friends that work well together, it's the, the best. So thank you for being a role model for us, and thank you for the awesome insights. It was truly amazing.

Tomer: Thank you.

Amit: Love you a lot.

Tomer: Likewise.

Amit: Yay.