Mykhailo Marynenko (Misha) was one of our very first O’Shaughnessy Fellows. Today, he joins Infinite Loops to discuss why AI tools should expand human control rather than replace human judgment.
From growing up in his father’s phone repair shop in Ukraine to building experimental AI systems at OSV, Misha has spent his life taking technology apart, figuring out how it works, and rebuilding it in unexpected ways.
We explore creator tools, privacy, data ownership, synthetic audiences, Infinite Canvas, and what it means to build AI interfaces that help people navigate complex information without giving up control of their work.
I’ve shared some highlights of our conversation below, together with links & a full transcript. As always, if you like what you hear/read, please leave a comment or drop us a review on your provider of choice.
— Jim
Links
Highlights
“When I was nine years old, I got my first gigs as a software engineer.”
Mykhailo Marynenko: Oh, there’s so much to say about this. So a little bit of background on this part is that most of the consumer electronics that get into Ukraine aren’t officially there. And most of the things that, if you, for example, bought a phone somewhere in Europe or brought it from America, you would want to have Ukrainian language in settings.
The first iPhone would be a perfect example. When the first iPhone just came out, it was an exclusive contract with AT&T. Without an AT&T SIM card, you wouldn’t be able to use it. And people were bringing these phones back to Ukraine, ordering from the States. And essentially you would want your phone to call someone, not just to have it as a shiny object. So these phone repair shops would not just repair something in case of an accident. You would be able to bring in your phone and get Ukrainian language, get it to work with Ukrainian SIM cards, make it actually usable. And there’s a lot that comes into this.
Most of these shops, like where I came from, those are not just, let’s order a new part and make it work. It’s actually going down, understanding how the phone works, how the operating system on the phone works, how components interact with each other, where are the weak points, how you can flash it to do more than it did before. In a sense, that requires a lot of deep technical knowledge and a lot of tinkering in order to get it right.
My initial steps would be just getting consumer hardware and trying to solve a problem for a person. And if the person would be just like, okay, I don’t have Ukrainian maps, how can I just use this phone to drive my car? Most people can, why can’t I? Then you go and, okay, how is the building maps working? Why is it geo-locked or similar things? And you go down and you look up different ways on how people could have written the app to make it work. Then you would go and actually try to fit it in order to satisfy the customer’s need.
Jim O’Shaughnessy: While you’re doing all of this, you have no formal training in this at all?
Mykhailo Marynenko: Of course not, no. And to remind, it’s... again, I was really young. It’s before I was nine. When I was nine years old, I got my first gigs as a software engineer.
Jim O’Shaughnessy: Wait, stop. You were hired as a software engineer at age nine?
Mykhailo Marynenko: Not hired. Technically you cannot hire a nine-year-old person. So no, it was more like freelance and doing some basic web development or also helping out with my father’s phone repair shop as some source of income to fund my toys.
Jim O’Shaughnessy: Was it just a natural talent on your part?
Mykhailo Marynenko: I think that it has a lot to do with my father having a lot of tools and toys and me just wanting to touch and experience those. There’s an early picture of me being really young and playing around with electronics or disassembling my toys just to break them, to do anything with them. Just like, I need to know what’s inside.
Jim O’Shaughnessy: So you’re familiar with Claude Shannon who came up with information theory. That’s kind of how he spent his early days. He was a free-range kid, but he just wanted to figure out how everything worked. Did you have a process or was it literally, were you just tinkering?
Mykhailo Marynenko: Just tinkering. There is no process. There wasn’t an end goal to satisfy the needs. Sometimes... for example, most of the background on me having access to tools, it was either having something from my father’s phone repair shop or later on in life I was living with just my mother. I wouldn’t have access to a laptop to code and there were going to be only two sanctuaries in my life in order to go and actually learn and tinker more. It was going to be either my IT classroom or my mother’s work. And at my mother’s work you would have a lot of corporate protections. You would not be able to do many things. And this is where most of the cybersecurity background started to stem from also. And sometimes you meet an end goal. You want to get some random corporate object to do things it should not do in order to just code. In some scenarios, you would just disassemble it to understand how the mechanism works, just out of curiosity. At some point you would just want to see how this thing looks from inside and there is no purpose to do so.
Transcript
Jim O’Shaughnessy
Well, hello everyone. It’s Jim O’Shaughnessy. I’m very happy, Misha, to welcome you.
Mykhailo Marynenko
Thank you, Jim.
Jim O’Shaughnessy
I know you quite well, but our audience doesn’t. So let’s go with your origin story.
Mykhailo Marynenko
My origin story starts back in Ukraine, where from a very young age, I was always in my father’s phone repair shop. It was a place where I was hanging out since I was an infant. And primarily, it was a place that inspired me and shaped me into who I am today. It’s the place where I discovered numerous technologies, went down into how things work. And that is pretty much where I came from.
Jim O’Shaughnessy
Talk a little bit about how you discovered how these technologies work, because I think a lot of our audience, they go and buy their phone at the Apple Store and they think, okay, it’s all done and I don’t have to do anything. And there might be some things in that hardware that regular folks don’t know about.
Mykhailo Marynenko
Oh, there’s so much to say about this. So a little bit of background on this part is that most of the consumer electronics that get into Ukraine aren’t officially there. And most of the things that, if you, for example, bought a phone somewhere in Europe or brought it from America, you would want to have Ukrainian language in settings. The first iPhone would be a perfect example. When the first iPhone just came out, it was an exclusive contract with AT&T. Without an AT&T SIM card, you wouldn’t be able to use it. And people were bringing these phones back to Ukraine, ordering from the States. And essentially you would want your phone to call someone, not just to have it as a shiny object. So these phone repair shops would not just repair something in case of an accident. You would be able to bring in your phone and get Ukrainian language, get it to work with Ukrainian SIM cards, make it actually usable. And there’s a lot that comes into this. Most of these shops, like where I came from, those are not just, let’s order a new part and make it work. It’s actually going down, understanding how the phone works, how the operating system on the phone works, how components interact with each other, where are the weak points, how you can flash it to do more than it did before. In a sense, that requires a lot of deep technical knowledge and a lot of tinkering in order to get it right. My initial steps would be just getting consumer hardware and trying to solve a problem for a person. And if the person would be just like, okay, I don’t have Ukrainian maps, how can I just use this phone to drive my car? Most people can, why can’t I? Then you go and, okay, how is the building maps working? Why is it geo-locked or similar things? And you go down and you look up different ways on how people could have written the app to make it work. Then you would go and actually try to fit it in order to satisfy the customer’s need.
Jim O’Shaughnessy
While you’re doing all of this, you have no formal training in this at all?
Mykhailo Marynenko
Of course not, no. And to remind, it’s... again, I was really young. It’s before I was nine. When I was nine years old, I got my first gigs as a software engineer.
Jim O’Shaughnessy
Wait, stop. You were hired as a software engineer at age nine?
Mykhailo Marynenko
Not hired. Technically you cannot hire a nine-year-old person. So no, it was more like freelance and doing some basic web development or also helping out with my father’s phone repair shop as some source of income to fund my toys.
Jim O’Shaughnessy
Was it just a natural talent on your part?
Mykhailo Marynenko
I think that it has a lot to do with my father having a lot of tools and toys and me just wanting to touch and experience those. There’s an early picture of me being really young and playing around with electronics or disassembling my toys just to break them, to do anything with them. Just like, I need to know what’s inside.
Jim O’Shaughnessy
So you’re familiar with Claude Shannon who came up with information theory. That’s kind of how he spent his early days. He was a free-range kid, but he just wanted to figure out how everything worked. Did you have a process or was it literally, were you just tinkering?
Mykhailo Marynenko
Just tinkering. There is no process. There wasn’t an end goal to satisfy the needs. Sometimes... for example, most of the background on me having access to tools, it was either having something from my father’s phone repair shop or later on in life I was living with just my mother. I wouldn’t have access to a laptop to code and there were going to be only two sanctuaries in my life in order to go and actually learn and tinker more. It was going to be either my IT classroom or my mother’s work. And at my mother’s work you would have a lot of corporate protections. You would not be able to do many things. And this is where most of the cybersecurity background started to stem from also. And sometimes you meet an end goal. You want to get some random corporate object to do things it should not do in order to just code. In some scenarios, you would just disassemble it to understand how the mechanism works, just out of curiosity. At some point you would just want to see how this thing looks from inside and there is no purpose to do so.
Jim O’Shaughnessy
So let’s talk about that for a bit. What kind of vulnerabilities did you find and continue to find in commercially available software, phones, hardware, where the innocent buyer is like, oh, this is totally secure, and it’s really not?
Mykhailo Marynenko
A couple of weeks ago, this is one of the most recent examples, I got myself an AI speakerphone, a microphone. The microphone itself is a very cool piece of hardware. It can listen and pick up your voice from 16 feet away. It’s a nicely marketed product that would go into enterprise meeting rooms and be used as a speakerphone. I didn’t use it for a meeting yet. I didn’t utilize it in any way. I just got it and being a security professional, I would go and check if there’s any firmware updates and I wouldn’t do so on the microphone, but I would go onto the manufacturer’s website and see, can I download it and just flash my microphone with new firmware offline? And when I saw the file, I saw that the file was trivially unpackable so you could see what’s inside. And my natural curiosity, I unplugged the microphone, I started it up and I saw it can connect to Wi-Fi. I’m just like, okay. Most of the AI recorders are usually just a recorder, just an audio recorder that you later on connect to your phone or a laptop and you upload audio and some server processes it and returns you an AI transcript. This microphone stood out because it had the full ability to have this networking stack that your laptop or your phone has in its full capability. And that drove my curiosity a bit more. So, what does this microphone actually do? So I got this firmware file and since it was trivial to get to know what’s inside, I unpacked the firmware and I started looking around what’s inside it. Maybe there are some modifications I can do to this thing. Essentially, I got to know that the microphone does transcription inside itself. So the microphone doesn’t go to servers like any other voice recorders. You would get the file uploaded. This microphone would do the transcription part, not summary, but transcription part on device itself. And I was curious how they achieved it because it’s a relatively cheap device, it doesn’t consume a lot of power, it can be on its battery for a long time. And I’m just out of curiosity opening how their ASR pipeline, automatic speech recognition pipeline, looks like inside. And there’s one file that is not obfuscated in any way. It’s not hidden under... I would have hidden it better. That contained two politically hot words that are specific to China that essentially were just there. So the firmware, even though the device was bought on American soil from Amazon, contained code words that... I don’t know what happens next if you say them, but the device is definitely designed to recognize these specific, highly censored words in China, phrases in China, and something happens. So I didn’t dig deeper yet into what comes after if you say these words, but it’s definitely designed to do something.
Jim O’Shaughnessy
So I’m assuming that the microphone was manufactured in China.
Mykhailo Marynenko
Yeah, the company behind the microphone is, I think, don’t quote me on the number, but it’s multi-billion, definitely multi-billion. I think it’s $10 billion. It’s a publicly traded company in China. It produces primarily 360 cameras. I think they’ve gotten into the audio space a little, or at least the enterprise space a little. And yeah, it’s a Chinese company.
Jim O’Shaughnessy
So what countermeasures? I mean, obviously I’m lucky to have you to let me know whether there is a bunch of spyware. What can regular people do?
Mykhailo Marynenko
Nothing.
Jim O’Shaughnessy
Nothing.
Mykhailo Marynenko
There’s literally nothing. There was another part of it. As soon as I got these words, I was curious to see what’s inside the privacy policy. So I got the privacy policy, I got the Chinese-English version and the Chinese version contains a bit more sub-processors. In China, most of the public cloud providers are sharing data with the government. It’s a publicly known fact. And their privacy policy basically mentions that yes, all of the things or part of the things would be transferred to Chinese servers for actually processing things, where the American privacy policy mentions that essentially it would go and get processed on AWS, on Amazon Web Services, and their marketing material on U.S. soil shows that this is AWS. But most companies don’t do this. Most companies have redundancies, backups. They don’t store information just in one country. There’s global replication and a lot of need for this information to be available in multiple spots. So I would assume that the Chinese privacy policy states facts that are true about this microphone, hence having the firmware that works the same way for any region, for any customer that detects these words. Essentially when you upload this transcript to their cloud, or maybe the microphone does it in some other secretive way that we don’t know about, you might be flagged. And also the microphone has an ability to... your voice is part of your biometry. There’s a couple of things about this microphone that allow you to segment people, means that if it transfers this information to China, it also essentially gets a part of your biometry, part of what you said, your words to identify you. And this is what’s scary. It’s not that it just would note that someone said something, any recorder would in a sense, but that it would transfer who you are, where you are.
Jim O’Shaughnessy
And your voice.
Mykhailo Marynenko
Yes.
Jim O’Shaughnessy
So like, that’s fucking terrifying. And does the United States in particular or the EU, do we know about these problems?
Mykhailo Marynenko
I think that comes a lot into business decisions, right? As a manufacturer, you don’t want to complicate the process of manufacturing these microphones. You want to simplify it as much as possible. I think the proper way of doing this is to make a regulation for these things, like having how we process voice, how we process all of this data, especially if it happens on device as advertised. The device is capable of working offline. These things should be documented. There should be a legal obligation to document these things. Even if this feature is somehow disabled, if you chose United States in settings, it still should be documented because me as a consumer, in the firmware of the device I have on hand, the keywords are present.
Jim O’Shaughnessy
Well, and having two different privacy policies or terms of service and the one in Chinese saying something very different than the one in English, that seems to be not a trivial problem to me. It would seem to me that these are things that... you know, I’m not a huge fan of regulation, but I mean, this seems like a regulation that might make sense for American consumers.
Mykhailo Marynenko
Yes, it definitely does.
Jim O’Shaughnessy
And why isn’t there one?
Mykhailo Marynenko
I don’t think there’s any particular reason for it rather than it just isn’t there yet.
Jim O’Shaughnessy
Let’s move on to things like DeepSeek. DeepSeek was super popular when it was launched. Do you think that people in this country or Europe or elsewhere knew that virtually every one of their queries was going right back to the CCP in China?
Mykhailo Marynenko
I don’t think so. I don’t think people think about it. They should and at the same time, they shouldn’t. DeepSeek depends on how you use it because DeepSeek models have been open weight. You would be able to download the model weights and even self-host those. And there were numerous American providers that provided access to the model that were territorially located in the United States. But most people went to download the Chinese app from the App Store that would process all of this information on Chinese territory.
Jim O’Shaughnessy
That’s the thing that always astounds me, right? Like we have DeepSeek, but it doesn’t go back to Chinese servers. But I mean, let’s be very clear, there are very few people who have resources like you. And I was thinking about it this morning as I was thinking about what I wanted to talk to you about. Why don’t big corporations have a bunch of Mishas?
Mykhailo Marynenko
Mishas are hard to find. You’re lucky. That’s probably the reason.
Jim O’Shaughnessy
Well, yeah, we discovered you as you know. But our audience might not know you were one of our first fellows for the O’Shaughnessy Fellowship program. Actually, it was more of an art project. Let’s take a moment before people are like, what did he just say about all of these things that we can buy in the App Store? Talk a little bit about that because I was immediately hooked. You had me at hello when I saw what you wanted to do. But tell our audience about that project because I think it’s fascinating.
Mykhailo Marynenko
I will connect it a bit to my life story as well. The war in Ukraine started and before the war started, I would go to visit France and Sweden, where I would end up for a year more before I would move to America. And during Sweden, I would call that period creative boredom. I had a job that I was quite good at. It stopped requiring a lot of attention. All of the things I could have done at the time in order to drive innovation of the company or generally contribute as much as I could for the business, I’d done. And I essentially was left to do all these legacy systems just to keep up the project in a sense. And I was also left to experiment with a lot of new things because they wanted to see if I could do something else as well for them. And being in Sweden, the weather was atrocious, it was sad, it was dark, or it was bright 24/7. I essentially just got bored because back in Ukraine you get all of the friends you ever need. You’ve got your life history, you’ve got your classmates, you’ve got your family, you’ve got everyone there. I had only one friend and his girlfriend. And I started to explore Sweden much more. Essentially I came to one of the raves, musical events in Sweden where I met a lot of wonderful people. My first night at that particular rave, there was a guy getting me into the room where all of the preparation goes. So it’s not a venue space, but a new space. And I see a vibrant hacker and artist community. First thing I see is people disassembling a Tesla battery from a car just randomly on the table. You enter the room, there’s on the background you here loud, beautiful techno. You open the door and you’re like, okay. One person draws, one person makes a video, weird big sculpture. Another person just goes and probes the Tesla battery. And essentially I got along with the creative community there. It was amazing. I met a lot of wonderful people who weren’t incentivized a lot to do what they did, but yet they did it and they made a lot of art. And I never thought of code as a routine for myself, or engineering for that matter, as a routine for myself. I always try to not think outside the box, but be attentive to things I have around me in order to drive my work. Since I knew a lot of artists, I tried to paint, I tried to make sculptures. I visited numerous art events, both very high-end and quite shitty as well. And essentially there were a couple of artists that I wanted to work with and I was doing a lot of design work as well. I wouldn’t just engineer software, I would design how users would interact with it, how to make the business decisions work with design. And then I was just like, okay, let’s apply this to art. Let’s see how my engineering skills can work to create something that serves zero purpose in order to make any revenue. I met a Ukrainian artist and essentially we wanted to collaborate on some AI things. We found that generating images at the time wasn’t as appealing for her collection. We tried training models on her work, tried approaching it from so many different perspectives and ways. And one of the guys in the space was just shouting, “Use brain!” It was like, what do you mean use brain? We are using our brains. And the dude was like, well, quite literally buy an EEG helmet and try to fuck around with that. And I’m like, oh yes, this sounds fun. And that’s how the descent into one of the projects for the fellowship started. Essentially we wanted to replicate what your visual cortex processes and correlate it to a model that would be able to generate approximately the same image. I think that back in Sweden, the project wasn’t as quite successful as it should have been. I think that most of the work was more exploratory. And when I came to the States, there were two performances that went amazing where the project would actually be used, where we would get an EEG helmet and do real-time processing to get as much of your visual cortex while musicians or artists perform and present it on stage to the audience to see the imagination of an artist in real time. So yeah, that’s one of the projects under the fellowship. The other one was also quite fun. The second project was analyzing crowd behavior. Primarily, I was just amazed and fascinated by raves. Most of the musical events you would visit in the modern day, you would see people just staring at their phones and going there to post a new shiny picture. They wouldn’t go there because they enjoyed the music, or they enjoyed the music but didn’t appreciate it that much. They didn’t appreciate the artists. They would go there as a social gathering, yes, but not for art. And essentially most of that underground scene for raves, you would find people who are actually enjoying weird alien sounds that make zero sense to normies, just not staring at their phone at all, where you would have a sticker on your camera that would prevent you from taking pictures. And people would be free. People would not care about time, people would not care about social media. People would not care about anything but having a great time. I found this as, okay, so let’s compare. How do crowds behave in these two different scenarios? What happens in normal musical events and why are our attention factors quite less when we not just have the phones, but what’s different between specialty underground, hard-to-get events and general public, huge festivals? And I got my hands on a couple of sensors like seismographs, lidars, infrared cameras. And I would start tracking people. I would see how people react to different parts of the music, to different styles. In some events we would collect Spotify, Apple Music to get their general taste as well. In some events we would even get an optional face... a person would be able to submit to correlate a specific person to their preferences and to music. And essentially the first initial version of that project would be an artist that plays in real time would get a distribution and real-time charts of how different parts of the crowd engaged, the parts that are not engaged, what do they like, all of these cool and nice things. And then later on we would bring generative AI to generate music in real time for people who are less active at the overall event to actually drive not just attention factor, but actual emotion, try to push people more and more, covering more of the audience. And the generative part went insanely well. There were numerous successful events where people would be enjoying and taking their attention fully just to music and to themselves, dancing instead of doing anything else.
Jim O’Shaughnessy
Obviously the capitalist in me sees something marketable there.
Mykhailo Marynenko
Of course, of course.
Jim O’Shaughnessy
Because how much did the scene change? I mean, can you quantify it for us?
Mykhailo Marynenko
From 60-ish percent to almost 90.
Jim O’Shaughnessy
And did people know what was going on?
Mykhailo Marynenko
No, they did not. Not a single event was talking about what’s actually going on. When you would buy a ticket, Spotify would be an option, but it wouldn’t be required.
Jim O’Shaughnessy
I see. And how many people gave up the Spotify playlists and everything else?
Mykhailo Marynenko
Less than 50%. There wasn’t a higher number than 50%.
Jim O’Shaughnessy
So you’re operating with a minority of the Spotify data. And I of course love that it’s all voluntary and that if you don’t want to give your Spotify, you don’t have to. But what was the unlock, what was the key that allowed you to move from the 60s to the 90s? Because that, I mean, that’s the difference between the most popular rave promoter and all of the also-rans.
Mykhailo Marynenko
Right. It depends a lot. Of course, the bigger the event is, the harder to drive people’s attention to the actual performance. The general idea is that you as a person, if you like specialty music, you wouldn’t like just one genre or just one artist. You would like many different things. An artist doesn’t know that, especially they don’t know that in the moment when they perform. They don’t know what the audience might expect or like. And sometimes it’s not about music at all. Sometimes it’s about natural rhythm, sometimes it’s about the group a person came in. We would track groups, we would recognize groups of people sticking together and generate more relevant suggestions for DJs, for example, to select from their existing track selection. So you still have the creative choice in style or music, but to suggest better times to kickstart more and more attention. So essentially the breaking point was that you can hook people more and not lose this and you can still have the control over your performance.
Jim O’Shaughnessy
And did the participants... did you talk to them afterwards? Did anyone talk to them?
Mykhailo Marynenko
Of course, yes.
Jim O’Shaughnessy
Yeah. And what were the...
Mykhailo Marynenko
Just general amazement. Just like, this was so good. There wasn’t a single time where it was just like, it just came, kept going. That would be the tagline usually from people. It would just keep going, driving attention or generally just keep enjoying.
Jim O’Shaughnessy
Lots of use cases here. Let’s shift gears into what we’ve been building because we’re taking a very different approach to building out the AI suite at OSV. Freestyle on that for a little while.
Mykhailo Marynenko
Freestyle. AI Lab. Well, we call ourselves the Department for Engineering, mostly AI Lab. We are a small team that develops a lot of fundamentally new ways to look at data, to look at how you approach AI, yet not trade off on existing things and existing approaches. Being a relatively small team, we covered so many things. I’ll tell about our successes first, then I’ll tell about what’s going on, not so good, both with the market and what prevents us from delivering the product we envision in a way. So from successes, we dug deeper than just using LLM APIs. The first initial thing of how AI Lab came to be is that we wanted to get LLMs to write movie scripts. That was the initial...
Jim O’Shaughnessy
That’s right.
Mykhailo Marynenko
...for the engineering residency. So I would do this part-time. I would try to create synthetic datasets and actually drive the model to create a cohesive screenplay, which would not go well.
Jim O’Shaughnessy
Yes. Well, for listeners and viewers, our mantra is human-in-the-loop, a centaur model. We think that AI just left to its own devices, you’re going to get mostly a tsunami of slop. But with a human in the loop, if it’s built the right way, you can get magic.
Mykhailo Marynenko
So essentially we went to agentic approaches before most of the models we have today are much more suited for agentic stuff than when we just initially started. Our approaches to doing screenplay writing and generally approaching doing this kind of writing work that should be connected, that should have a higher-order understanding of what’s going on in a sense... general completions, the way how language models work, in a sense where you would get a context and try to predict which tokens come next, would be good for writing, but terrible at many different number of things. So essentially we tried some agentic approaches. We’ve built out our AI chat thing we still have today, which I think that many people would say, okay, building an LLM wrapper is easy. I’m going to say, try to make it right, try to do it properly, try to make it useful and try to make it work on all devices in poor network conditions in all the possible edge cases. This is what a product is. You polish for people. You don’t want people to get stuck in chat and then try to say to me how easy the AI chat is. Although there are very much different sets of alternatives in the market, like Claude. Claude is so much better in my experience than OpenAI’s models when it comes to work. When it comes to personal things, somehow ChatGPT is a little bit better. I don’t know why. OpenAI did an insanely good job on making a good product. The chat, the user interface, the responsiveness. When you open the app, you have first time to interaction, you have everything set right. And that’s one of their insane advantages. They were the first one to try to make this product and they did it good. It would be really hard to compete with them. I think that we are on the level where our chat platform is still quite buggy because again, it’s in beta, but we’re at a technological level where we cover most of what their team did in terms of product, yet we added our nice things and our unrestrictiveness and all these scenarios. So that’s part number one. It still exists. We need to collect training data. And there’s part number two. How do you approach AI products with all of OSV’s craziness and all of the “try to create a movie script, then try to actually generate video based on that”? And there’s a lot of agentic products out there. There’s a lot of open-source projects that would try and do this, but they’re fundamentally wrong. They’re not the way you should be doing that. And the deeper we dug, the deeper we found issues that relate to how general businesses work, how big data was structured, how it was initially optimized for hardware that was much slower and much less flexible and much less resilient than we have today. How all of this legacy bulk continued to exist and how most researchers don’t want to go into software engineering and they need to. So most of the researchers would think well about math, they would think well about general concepts, they would do good research, but they would never do good engineering. You would get a great model, but not great engineering behind that model. All the tooling that you do when you train models nowadays, it’s still this way. It’s not universal, it’s not portable. You would adjust so many different things and you don’t want it to be universal, but you want it to be portable, you want it to be extensible. Right now it feels more like people are building temporary solutions for small events to drive their business and experience better, but not build something that drastically shifts how they do their work. That’s where we come in. There’s a lot of our new sets of tools, approaches to how we store, index, process, train, do all of these sorts of things and even maintain our own hyperconverged infrastructure and bare metal infrastructure. And yep, that is also the lab in a nutshell, where we try to approach it fundamentally different in terms of how we approach data. And with that comes a set of challenges, remaking the whole ecosystem, trying to make novel approaches that I probably will mention a bit later, which is Interplanetary Link Knowledge, or IPLK, and our upcoming flagship product, Infinite Canvas, that basically change how you look at AI in general, how you look at data, how you look at data sharing, how you look at data brokering and exchange, how you look at data indexing, how you look at AI products where it’s not just chat.
Jim O’Shaughnessy
Since we’re talking about it, let’s do it now because every time that I go in there, I’m blown away by what you’re achieving because I of course use all of the commercial large language models and yeah, you’re right. ChatGPT I always think of as more in white tie and tails and, you know, Claude is much more relaxed. But we’re taking a fundamentally different approach.
Mykhailo Marynenko
Yes. So fundamentally different approach. There’s so many things this can mean in a sense, right? There’s all the technological things I just mentioned, but there’s also a user side of things. There’s a way how we build our AI products today, trying to connect to technologies that fundamentally weren’t designed for AI, that fundamentally didn’t predict AI would exist in such capacity we have today. So people connecting agents to databases and all of these things, it doesn’t look and feel right if you’re deep into engineering. And from my world, it just doesn’t feel, it doesn’t sit right even by today’s metrics. We have, what was it, around four years now, three years now, where people are trying to do agentic products. We see success stories like Cursor and I’ll go deeper into why this is a success story. And we see a lot of shitty products that are essentially just either trying to be products like Cursor who took an actual proper and innovative approach to do what they do, or just try to glue shit and stacks in order to make it work, connecting things that weren’t designed for each other to do something. Infinite Canvas is probably a new category of AI product that didn’t exist before. It’s what I like to call it, an infinite-dimension reactive computational whiteboard.
Jim O’Shaughnessy
Very Philip K. Dick of you, Valys.
Mykhailo Marynenko
Essentially it’s a whiteboard that understands what you do. It’s not a diagramming tool. It’s not a tool where you would be drag-and-dropping your tools and connecting things together. It’s a tool that orchestrates itself, that coexists with you, where you are capable of changing any part of the process and complex orchestrations, which I will go to in a second, down to their smallest details without having any domain knowledge. Infinite Canvas should be able to provide you with an ability to even train your own models, orchestrate a lot of different...
Jim O’Shaughnessy
And let me stop you. You mean like people...
Mykhailo Marynenko
Yes, like Jim O’Shaughnessy.
Jim O’Shaughnessy
If you achieve that, you’re going to win a Nobel.
Mykhailo Marynenko
No, I’m not. So when it comes to these products, we see a lot of really cool new approaches to collaborative environments. One of the best products we have today that I enjoy a lot, that I’m sad about because I feel like they are taking the wrong turn right now, is Figma. Figma is essentially a tool for designers where they could wireframe, design really complex things, build prototypes of apps without having domain knowledge. It’s not a no-code tool, it’s essentially drawing rectangles but doing it well and doing it like 100 people can do it in parallel. And what’s cool about Figma is that they didn’t write another shitty layer like Adobe does. They would take their codebase that exists from the ‘80s or ‘90s to this day and try to continue on the legacy approaches just to not break their existing user base. Figma did a lot of drastic changes to how you approach general UI/UX design. At the time it was a killer product that you don’t need to install. You open it in a web browser and it works. It magically can do much more than any other product was able to in your web browser. And we drew a lot of inspiration from a lot of products like Figma that are essentially infinite whiteboards, infinite canvases. There’s a whole curated list of websites that mention Infinite Canvas products where you would be able to diagram, to design, to prototype, to do all different sets of things. But the biggest issue with these is that, for example, Figma, they would add AI features, but they would not fundamentally shift some of the things in their product to make these AI features right and sit well. What does Infinite Canvas provide for a user? Essentially, you can dump any information you want. You can use it structured or unstructured, both. You can do spreadsheets, you can upload manuscripts, you can just drag and drop it. And it’s a spatial space that would use machine learning to automatically arrange your space as well. So as soon as you dropped thousands of files, it would essentially just organize information for you. Then essentially you, as with any other LLM, you’re capable of doing anything like any other LLMs do. So in a sense, what’s in these files? But what’s crazy about this product is that you would be able to say, I want all of my characters to be a different way in my script. And it would go and parallelize itself and use a semiotic approach. Semantic, primarily what we have today, we have symbolic. So symbolic is something like images, words. Semantic is them, but understanding of them in context. But the missing part is semiotics, where you would be able to connect and change the meaning of a specific word, not just in context, but globally. Where you are capable of reacting dynamically to new information or making adjustments and still operate on symbolic and semantic environments. And also the symbolic part of most of the AI today is not on the level required for complex automation orchestration or being able to also extend it not just to basic business automation, but to do it for actually writing a book. Some of the most wonderful use cases I’ve tried previously that work well today and some of the things that are upcoming, essentially you write a book, you have many different characters and you uploaded your first outline or asked it to do a first outline of your idea. As you expand, you don’t lose any other properties. You can easily navigate between your drafts. You can put there a lot of information, tens of thousands of pages that you can still easily navigate. And then you are able to tell it, okay, let’s write it. Let’s actually write it. And you have a nice, what you see is what you get, text editor that actually would properly work, that would allow you to see your final book in a sense or allow you to modify anything in that book. And maybe there’s some character that owned a network of hotels or some traveling company. And the traveling company would be named after your character. And around that there stems a lot of different parts in your story about how he might have fun making fun of the company name, for example, or he was bullied in childhood, or different sets of things that rely solely on understanding and deeper meaning in the general overall context, higher-order level thinking about it. If you change the character’s name, if you ask the canvas to change the character’s name, you have the flexibility to change all of the points of view in your story cohesively in order to make it right. Not just ask, yeah, how would it look like? Or let’s do it, or let’s adjust this or these parts. It just magically would go and rearrange your story beautifully and cohesively into one piece. Imagine putting these capabilities to a test. Let’s say in five years you would be able to generate a movie, or maybe a bit more than five years, who knows, maybe it will be tomorrow. Essentially when you would want to generate a movie, you would want to have an ability to have a stable trajectory for people or crowds moving because it’s essential for your story. You would want a lot of these symbolic things, all symbols to be in place. And you want your video generations or image generations or character appearances to be consistent and you want them to be physically consistent as well. One of the technologies I’m proud of that we made in AI Lab, we call it Canvas Kit. Canvas Kit is essentially a multimodal environment for information to exist and be represented and be operated on. Canvas Kit allows you to visualize almost literally anything when it comes to structured or unstructured information, both to handle real-time stock information and to having a scan of the room or a 3D space and seeing the process of how the 3D space would be converted into a scene in the movie, for example. And this is all one cohesive product that you don’t need to go into something like a 3D editor, like Blender, then go into Photoshop, then go into all of these places. And yet you will not trade off on features. So essentially you would potentially be able, with this technology, to simulate real physics for your movie, or also provide VFX for your movie, provide even aperture or lens parameters for your shot that are not purely dependent on AI generation, but on actual post-processing that make it possible. So this is what Infinite Canvas is, a completely new approach to how you interact with complex, structured, creative, boring, whatever information in a huge and scalable manner that can self-automate, self-orchestrate, interconnect with other processes and allow you to orchestrate not just boring business work, but also creative stuff.
Jim O’Shaughnessy
Let me translate it into normal human speak. If you’re a writer using this, you can be writing your novel or whatever, your screenplay or whatever, and you change that guy’s name, right? And rather than having to hunt and peck throughout the entire document, let’s say it’s at 500 pages now and this guy is a really important character, and then you have a writer’s room and everyone in the writer’s room is like, oh, this character is awful. We’ve got to really upgrade him or her. They’ll be able to do that. But then it understands context and it goes and makes those changes on your behalf.
Mykhailo Marynenko
It doesn’t only understand the context of semantic text, right? It understands the context of symbolic things. It can invert symbols like change a relationship between a character or someone and convert it into semantic text that would be consistent.
Jim O’Shaughnessy
So give another example of that, because as you know, this is one of the things that I get really excited about when you tell me, but the first couple of times I was like, Misha, talk to me like I’m a small child or golden retriever, to steal from the movie Margin Call.
Mykhailo Marynenko
Sure. Imagine your company is a complex system that is alive that operates both with or without you. The company has multiple verticals. The company might have thousands of different processes inside all of these verticals. And essentially you want to merge two verticals or you want to change one process and you are able to predict every single part of what it means and put it into effect. This is another example. If you do complex business automation, it wouldn’t be just, okay, let’s intake emails now. It would be also higher-order part about what it means for the business. What’s the actual meaning of why it exists? What’s the purpose of all of this? You want to change a higher-order process, but how do you re-puzzle all of these things together? This is what Canvas is capable of. It would go and rearrange all of the pieces to fit right into the actual proper network of things as you set it, not as it just predicted, but also to have the flexibility to set things as they go.
Jim O’Shaughnessy
So in, again, translating, it changes not just the semantic part, it changes everything.
Mykhailo Marynenko
Yes. Essentially a symbolic thing is, I would tell you, this is a glass, right? There’s a glass that is a glass on the window. It’s the same glass. So glass means something as one thing, right? But the glass in a cup has a bit different meaning. When I speak to you and I look at the glass, I say “a glass,” which means cup. If I look at the window and there is no other glass around me, or I haven’t drawn attention to any glass around me, I would say, okay, so this glass is dirty, for example. It would give you a meaning. How does this work? Glass is symbolic. It’s a symbol. The semantic part is context. So when I look at the glass, for you to understand that I’m talking about the window, you see the semantic part where I look at the window. The symbolic part would be that this is essentially a material. But there’s many ways to look at symbolic things in our world. There are mathematical equations, for example, where you would have symbols and how those symbols would interact. And you have operations, which is also some kind of symbol that performs computation. And essentially what AI Lab would allow you to do is to change the equation for your whole book and to do it consistently and without the hassle. So if you have an equation where you have a couple of characters doing things in a perfect balance as you see it, you’re still able to follow your equation or change your equation in general in order to rearrange everything that comes into your book while still keeping it...
Jim O’Shaughnessy
Could this potentially become something where we ship a book or a video or whatever, and the end user, who we don’t know, can actually change the story?
Mykhailo Marynenko
Yes, it’s part of it, of course. Or if we ship a movie, for example, if we provide a movie, this whole technical infrastructure and frameworks and libraries we developed internally allow and predict for the future of making films, for example. And what if someone would want a different end, not just for the book, but maybe for a movie of the book, while still keeping the original intent of the author and his equation in his brain? That is the possibility, yes.
Jim O’Shaughnessy
Now, I know a lot of authors and movie makers, et cetera, can be very protective of their art. So in those circumstances, you can also ship it so they can’t change it, right?
Mykhailo Marynenko
Of course, yes.
Jim O’Shaughnessy
One of the things that we’re trying to do with all the various verticals at OSV is ultimately, we want to empower the creator. And if creator A wants it to be locked and no changes, then creator A gets that. But if creator B is like, no, let’s experiment, let’s see what comes out of that. But it will always be at the creator’s discretion, right?
Mykhailo Marynenko
Discretion, right. You and I align on that 100%.
Jim O’Shaughnessy
I, being the crazy person that I am, love the idea of being able to write a story and then see all sorts of different ends. Will we ever see a time when, let’s say I write a story and I say, yeah, you can do whatever you want to this story. Will we be able to let the reader of that story make changes that he or she wants to see as the end? Like, oh, I hated the end of, you know, fill in the blank. I really think they should have gotten together and they didn’t get together. Will we provide suggestions and support or is that also completely customizable?
Mykhailo Marynenko
Completely customizable.
Jim O’Shaughnessy
Let’s say they have a very specific... yeah, these two got to get together, they’ve got to get married or, you know, live happily ever after.
Mykhailo Marynenko
Yes.
Jim O’Shaughnessy
But they’re like, I don’t like this one. They’re going to have the opportunity to ask the AI, well, what would you suggest?
Mykhailo Marynenko
Exactly. But it’s also not “what would you suggest?” right? The important part is that most of the things... there’s the original creator’s intent, right? So you would want to see how a creator would change the ending and what were the possibilities for this ending to be. Not only training models... and we have sims. We’ve done multiple sims by now and we will do much more as well in terms of dynamically training models. Sims are essentially being able to simulate a person, right? Just getting as much data and trying to speak in the voice... it’s not essentially cloning someone, unfortunately.
Jim O’Shaughnessy
Unfortunately.
Mykhailo Marynenko
But I did a couple of experiments for my sim. I enjoy writing emails using my sim. It’s amazing. I don’t ever need to change anything about it. It just perfectly captures my humor, my points of view, things I would most probably discard just exist in that, right? And it comes not only for sims, but when you train an AI model, when you show it so many different patterns, same way as the human brain, you don’t always remember specific things. You recreate the appearance of these memories. You are synthesizing from your training data. You looked at things and you would not be able to tell exactly how that tree was looking. Same thing with AI. When you train models, it would reconstruct from the data you’ve seen previously. During training stages, I think that there’s a big disconnect on these ends. And something that we are looking at in terms of some of our internal R&D is that even if you train a sim, would it be able to look at the original things? Because you as a human, you would probably be able to go into one of your notebooks and take a look at it. And you would remember some things differently...
Jim O’Shaughnessy
Which I’ve told lots of stories about. I’ve kept journals for people who don’t know since I was 18. I’m 65, so I have lots of journals. And it was one of the things that really unlocked my understanding of human memory. I would swear on a stack of Bibles that I believed a certain thing. The story I often tell is about the first Gulf War. And I was at a party here in Manhattan and the group seemed to be like, yeah, you know, the first Gulf War, I support it because Saddam went into Kuwait and that was crazy. And I was saying the same thing. And then the group would say, yeah, but this one, like this is a bad idea, right? So I completely believed that was my memory, that I had supported the first Gulf War. And then I was looking for something else in a journal that I wrote right around the time of the first Gulf War. And when I read it, I realized I did not support the first Gulf War when you go back to the real-time writing. And so it’s quite a shock really, because it led to my theory that memory is often overwritten with our current beliefs, right?
Mykhailo Marynenko
That is true.
Jim O’Shaughnessy
And so when you think that you remember something completely crystal clear, you’re probably wrong.
Mykhailo Marynenko
Yeah.
Jim O’Shaughnessy
So let’s talk a little bit about the sims, because that was another thing that I’m very excited about. One of the things that I’ve asked you to do is create synthetic audiences so that we can stress test, do A/B testing on these synthetic audiences. And we kind of kicked around ideas like maybe we should use Big Five OCEAN profiles. Maybe we should interview varying types of people.
Mykhailo Marynenko
I already have Enneagram. So essentially you have different practices, right? You want to do what you said, the last two, Big Five OCEAN. Essentially what Canvas provides you today is the capability and you can go and basically tell it, I want an audience from these parameters. Can you run simulations on all of these things? Can you go to the internet? And we have, by our internal benchmarks that we of course haven’t yet published... we are quite early to most of the things, but we have one of the best web crawlers that is out there. We crawl and index much faster than big search engines like Google, for example.
Jim O’Shaughnessy
Brag a little. Tell me how much faster we do that.
Mykhailo Marynenko
We do it on average... there is no definitive way. But essentially the technology itself is fundamentally different in terms of how you interact with web pages and how you follow web pages. That proposes a new algorithm that provides the ability to crawl a little bit faster. Of course, we don’t know Google’s numbers. I don’t think many people do know. So we can’t say for sure that we are 100% or 20% faster than Google. But essentially we are capable of re-indexing more than 15 terabytes an hour of pure web pages. It’s an insane number. It’s a purely insane number. Billions of pages, pure text that are getting indexed not just as a typical search index, but as a semiotic index. And by that I mean if, for example, inside your canvas you’re told that, oh, there’s these effects, and you want to use internet knowledge to synthesize or run tests or create an audience based out of internet knowledge updates, you have a history and time travel across internet knowledge almost in real time to synthesize your audiences and see how that opinion changes now versus a week ago, for example. And this is where semiotics are cool because you essentially are not reacting to new data, but you react to new contents of data. And you have a network effect and cascading effect of things that change that births this new information.
Jim O’Shaughnessy
So, yeah, so for our viewers and listeners, one of the first things that I said to Misha was basically, I want the Eye of Sauron to be able to see everything that is going on. Because obviously, you know, prediction markets are really hot right now. And this maybe takes that a step further.
Mykhailo Marynenko
Yes.
Jim O’Shaughnessy
Let me restate again for maybe our listeners who aren’t as technically brilliant as you are. I am certainly one of those people. But rather than have a set audience, right? I, for example, I’m right now in the throes of writing a fictional thriller, the first fiction I’ve ever written. When the system is finished, I could literally go in there and say, find me the audience for this particular piece of work, right? And then I could change things in my story.
Mykhailo Marynenko
Restructure your audience also. And you can use it without a chat or with chat, however you want.
Jim O’Shaughnessy
And essentially it might serve me up an audience that surprises me.
Mykhailo Marynenko
Right? Yeah.
Jim O’Shaughnessy
So rather than go to, oh, yeah, here are the stats on people who like historical fiction, or here are the stats or general personality profiles of people who buy, you know, books about World War II, it will synthesize and give us a new look, right?
Mykhailo Marynenko
In a sense, yes. Yes, of course. And besides giving you a new look, that would be just text. This is where Canvas Kit, our representation layer, kicks in. You would be able to see it as a higher-dimensional space, see populations of information, see different populations of opinions, be able to dissect those and run on subsets of those, for example, or on all of these, or expand those, or subtract or summarize those, right? So you have this full multimodal flexibility, not just around what AI produced or what you want AI to produce, but around your data as well and the data that AI produced as well.
Jim O’Shaughnessy
So again, stop me when I’m wrong.
Mykhailo Marynenko
Sure.
Jim O’Shaughnessy
When it’s working the way I want it to work, we want it to work, we could in real time experiment with different ideas we have for the story.
Mykhailo Marynenko
Of course. Yes.
Jim O’Shaughnessy
So we have, as you know, because you sat in the writer’s room, we have what Jimmy Soni called a villain worthy of Iago from Shakespeare. That led me to believe, you know what? I don’t think the heroes are beefed up enough. And when this is working in real time, I’ll be able to go in and put my ideas for how to beef up the heroes so that they’re worthy opponents for this incredible villain. And I’ll be able to see... right, then maybe one of the heroes is skilled in aikido or whatever. But as the author, I’ll be able to see what that idea changes, right?
Mykhailo Marynenko
Yes. And more importantly, you would be able to see the cascading network effect and trace it to see how it changed. This is the important part. And you can affect that network in order to get different things if you didn’t like the end result. So you again have the full supervision. Of course you are not able to change how the model thinks, but you’re able to change higher-order structures of how the model gets all the things together and that’s essentially part of it. Yes.
Jim O’Shaughnessy
And so this is applicable far beyond writing, right?
Mykhailo Marynenko
Oh, I have something to say here. You’re coming from a finance background and of course, in a perfect world I see the core technology we have today in R&D stages be used to study network effects of traffic jams in New York on financial districts, to correlate those with stock price drops, or to be able to project what this movie would be doing to research and innovation in the biotech sector, for example. A lot of highly disparate things we have around us are shaping us in many different ways we don’t notice. But having big data systems that scale so well and have so much depth and yet allow you to orchestrate all of these steps blazingly fast and without the hassle in a perfectly normal user interface that anyone is capable of using, allows you to do some crazy stuff.
Jim O’Shaughnessy
And one of the ways that I think about this is we’re giving our user... I love music, as you know. And the more I’ve been thinking about it, the more I’ve been thinking that the user, the end user, is the conductor of an orchestra. And they can be doing like, no, you’re playing that note wrong. And no, you got to do this. And yet, as you bring up, like markets, obviously I’ve been known to have a little interest in markets. When you mentioned the idea about the traffic around the financial district, I remember reading in the Journal of Portfolio Management, a very geeky publication, but like 20 years ago, and this guy got this idea that stock prices went down when the weather was inclement.
Mykhailo Marynenko
Yeah.
Jim O’Shaughnessy
And I was intrigued because I love new data sources, I love new indicators. But what it took that guy to literally produce the... so he had a hypothesis, right? Stock prices go down when the weather is inclement or cloudy, and they go up generally when it’s sunny and nice. But it took him years to even test his own hypothesis. Spoiler alert, he was right, which kind of surprised me. But literally, when I read the afterword, he had a huge team of undergrads. He was a professor, and so he had a huge team of undergrads and grad students literally down on Wall Street taking pictures and doing all that. And the data capture... we can do that essentially in real time.
Mykhailo Marynenko
We already do most of the data capturing. This is the part. And I think that we could touch on Interplanetary Link Knowledge now. It’s a crazy name for a crazy technology that is quite new at its core concept. It encapsulates a lot of really complex systems and brings them down into one beautiful, cohesive and simple thing. I can’t wait to write about it also when we have it running and we have a bit of benefit in terms of getting our products on the market first. Essentially, IPLK, Interplanetary Link Knowledge, is one unified layer that allows you to have ownership over your own data, yet not sacrifice an ability to use this data in all of the crazy contexts we were just talking about. It allows you to both utilize very traditional approaches to data and at the same time allows you to boost those and scale those to very enormous sizes where some of this information might be even stored on your phone and we don’t even need to store it on our servers. I would make a disclaimer. It’s not a decentralized system, or as many might think, there’s a lot of projects like this that tried using crypto. I would say that this is a semi-decentralized approach where you would have parts of your information on your phone and you would agree to have this information have a fixed time of life on some other end. And either end, when it doesn’t need it, it’s ephemeral and it disposes. And when it needs it again, it would be able to go back to your phone and get it if it needs it. And you have the full transparency over these things. And this works not only for these small examples like user data, but this essentially works as a potential new framework for data brokers, for being able to stream high-frequency trading data, even in a manner that is traceable, transparent and fully appropriate for modern big data systems.
Jim O’Shaughnessy
And importantly, the key that I want built in, and you were wonderful about making that a reality, is this also allows the person whose data it is to control that data.
Mykhailo Marynenko
Of course.
Jim O’Shaughnessy
So unlike a lot of the last 20 years where essentially most of the big tech companies... we were the product, right? Because they were essentially training on all of our interactions, et cetera. This gives the opportunity for me, if I wanted to use a particular dataset that was sort of precious to me, to control that, right?
Mykhailo Marynenko
Of course. Not just control that. There’s multiple ends for end users who are actually producing data and for companies processing, handling or operating on these things, right? When in the real world you would share a secret with a person, you cannot take it back, right? But the problem with the modern world is that you’re sharing a secret and you don’t know if you even gave it. And yes, there’s multiple disclaimers and a lot of cryptic language that people have to come across, yet they don’t know if their picture was used. You don’t know if the face on your photos in iPhone was used to train a model that would recognize faces for all of the other people, which was not a huge problem. I think that ethics and the general approach to privacy shifted a lot from us having villages and everyone knew in villages who was in those villages, to the industrial revolution in big cities where people would share information with the speed of walking distance, to the modern day, which is insanely new compared to what I mentioned just before, about sharing information online. And I think that we are just not balanced, right? I think the time will come. I think that there were so many case studies right now and there should be at least four times more in order for us to catch up and take the lessons to do it right.
Jim O’Shaughnessy
Right, yeah. And you and I talked about that a lot. There is huge cultural lag built into people’s behavior. And as you rightly point out, for the vast majority of human history we lived in small villages or farming communities, et cetera. You knew who you were telling that secret to and you knew whether they were going to be reliable or you thought you knew whether they were going to be reliable or not. And so that’s so deep in our human OS code that’s how you got the last 20 years, right? You got people just naturally thinking the old way. How do you see transforming people to think the new way?
Mykhailo Marynenko
I don’t think people need to transform at all. I think that our human nature is to share. We are social creatures, a lot of us are. Not all of us, but there are some historical exceptions. When you work with private information, you share your picture online. I have a lot of points of view on privacy. Ever since I was 10, I would work on computer vision technologies. I would love to play around with facial recognition algorithms, just being able to compare faces algorithmically, try to build databases of faces. And then at that young age, I was quite scared of implications of what it meant of having an identity that could be used and primarily it’s going to be used against you. And a year after, I tried to appear less with my face in pictures. There’s been a period in my teenage years where you would not find a picture of my face. It would be either covered with my hand or I would not be in the picture at all. I would have issues in my school because there’s a class photo, you should be there, why are you an exception? And a lot of these different things. But essentially what I found for myself is that having extreme privacy isn’t the answer, right? You as a human again need to be a highly social creature, especially in the modern world to survive. When you share your face, you share your photo on Instagram, there isn’t going to be a world where a single person would be able to trace where this and the implications of posting this photo goes. But there could be technology that is lightweight enough, portable enough, cheap enough and sufficiently made, right, that would allow both companies to adopt it and users to see where did their picture travel to, how was it used. So we see Web3, which is another end of this. I honestly hate all of the Web3 space and I know too much about the technology in Web3, about all the things like IPFS, it’s Interplanetary File System, not Knowledge. And we draw a lot of inspiration from things like multi-formats. It’s formats that can describe themselves. Imagine reading something that reveals what it is not just by words, but by actually reading it, you understand it in a sense. Same thing for machines that is not based on AI, but based on basic cryptography and algorithms that allow machines to recognize what information they are reading without having any kind of super intelligence to do so. Multi-formats are very important concepts that drove a lot of our symbolic research. Afterwards we changed a lot about multi-formats. But I still love the Interplanetary prefix, so be….
Jim O’Shaughnessy
Aim for the stars, hit the moon.
Mykhailo Marynenko
Right. It’s more...
Jim O’Shaughnessy
Yeah. You want to leave the galaxy?
Mykhailo Marynenko
I don’t know about that.
Jim O’Shaughnessy
But I think it’s really important for people to understand the point you just made because I personally think it’s vital and that is the technology can be developed so that literally people don’t have to change their behavior.
Mykhailo Marynenko
Exactly. And businesses can even save money by doing that. The problem is that Bell Labs was such an amazing part of the history of innovation and general things. They never stopped exploring, they never abandoned a lot of concepts. They were aware, just the structure of the company allowed them to do crazy things. And I think that most of the research we do today and how we build things today doesn’t allow as much experimentation as we used to have. And the more we build, the more we... this works, this generates revenue. Why would we ever change that? It’s better to lobby something that would prevent this from ever changing than changing things to their core, which is deeply wrong. And I think it’s a very bad habit for us as a civilization to prevent us from experimenting. And I think if we focused, if many companies that do big data today and companies that do all the sectors of analysis and recommendations and ad algorithms, if they actually went deeper into how they structured most of the technology behind this, they would have made it much cheaper for themselves, maybe even more profitable, and yet fully ethical.
Jim O’Shaughnessy
Yeah. And obviously that’s something near and dear to my heart. I think that absolutely, if you want to be in business, you should be in the business of having profits because how are you going to pay your staff? How are you going to grow or do any of those things, how are you going to support R&D? All of those things come into play. So one of the things that I like about the way you talk about this is you’re very practical about, hey, if you do it this way, not only is it going to be much better, much safer, it’s going to be cheaper and you’re going to make more money. And so is there kind of a monoculture right now in the big commercial designers and people who are producing the large language models? That’s pretty... I mean, help me out here.
Mykhailo Marynenko
People who produce, who generally... I can say only for some of the open-source world I’m tracking, maybe not all of it. I don’t think that there’s any community about how you handle data in a sense. The better the dataset you get, the better you clean it, the better you generate some synthetic data around it, the better your model is. The more data you have, the greater the model is and the harder it is to train and the longer it takes. There’s a lot of these variables and balances between these variables, dynamics between these variables that make it happen. And I don’t think that this is the case. The researchers are busy with model architectures, not data architectures, because this is not an AI engineer’s pain, it’s a data engineer’s pain and infrastructure team’s pain in order to make this data be stored and accessible. So yeah, that’s the general issue with having models. And there’s another end to this. Models would not be... we are not yet at the point with the technology we have that allows you to trace the original training information to AI outputs. There’s a sort of compression going on that prevents you from doing this. And I don’t think we will ever be able to. You cannot recall most of the things that drove your decisions every day. And it’s fine. It doesn’t shape who you are in a sense. We want AI to be able to do so, to extend its capabilities, but not care about data, right? But I think that the only way to do it right is to let people be in decision for usage of their data and how their data spends. Even if their data was used against their will, people should be able to see how the data was flowing. And that’s the thing.
Jim O’Shaughnessy
Obviously there are infinite workflows that can come from this. Let’s explore a little bit more how we are trying to, across the verticals at OSV, let them learn from one another, essentially.
Mykhailo Marynenko
Imagine that. Also one of the things we want to have inside Canvas is that basically Canvas should be able to provide you with an ability to connect your workflows together. If some person created a canvas, you would be able to embed other canvases or reference other canvases or to be able to send to something else as well. Also, there’s knowledge bases and things. How you can connect in a semiotic way of thinking about it is that if there’s something that interests our VC, we might notify our media. Maybe they want to write a book about them. If there’s something in terms of a book that would make a great company and we have a network of people who might be able to create it or a network of fellows who have finished their projects and are working at their boring job, for example, or we know some people or those people are generally available to our network, we might suggest, come together and create a company and we’ll be happy to fund you. I suppose that’s your part.
Jim O’Shaughnessy
The idea really springs from, you know, most of these things like the market are complex adaptive systems and emergence comes from below, not from the top. I’ve always been skeptical about... I mean if you just look historically, top-down systems really don’t work at all because they’re just absolutely the wrong design. The information bottlenecks that get up to those people up there are insane. And I often say to people, could a central committee with a five-year plan ever design a Rubik’s Cube or an iPhone or a pet rock? Of course not, right? Markets can do that because they’re these living organisms where all of the emergence is coming... right. And where you can try. And this is why I’m such a big fan of your approach. To learn all of this stuff, you’ve got to be tinkering, right? When you look at the biggest breakthroughs that people like Claude Shannon came up with, it was because that’s what they were, yeah, that’s what they did. And it seems to me that the reason we haven’t had the kinds of breakthroughs in more, let’s call it just basic science, right, is because the whole thing got inverted. We turned it into a top-down structure and people applying for a grant so that they can conduct their research, it’s a poisoned well because the top-down structure that has emerged is so narrow that unless you’re doing this, you’re not going to get funded, which to me is insane because you should be funding... it’s like one of the things we’re trying to achieve on a tiny scale with the O’Shaughnessy Fellowship and grantees. We want those people to get funded because that’s where all of the breakthroughs come from anyway. Do you see that model toppling and a more organic one replacing it?
Mykhailo Marynenko
I think that there are very different points of view on this. There’s a lot of personal things that drive you, right? When you create something, when you build something, when you’re tinkering with something, there’s environment that drives you, there’s opportunities that drive you. There’s so many variables that come into this. I think that bottom-up makes more sense just because that’s how I see the world as a computer scientist and how I see structures emerge. And I don’t see any particularly useful, as I imagine, less computationally intensive and more appropriate ways to scan trees or graphs or a lot of different these kinds of things. So I agree with you. Yeah.
Jim O’Shaughnessy
And really one of the things that I... it’s a thesis, right? I could absolutely be wrong, but I think that’s where markets come in, right? Because we’re in a relatively free market here in the United States. And so we’re doing this, right? And so the way the information gets into the network is we do leads to really cool things. And then others are like, oh, maybe we should be doing it that way as well. Markets are amazing at that. One of the things I used to joke about is in relatively free markets, markets co-opt everything. You could take, look at the 1960s, right? And there were these huge movements that were anti-war, make love not war, flower power, all of that. And literally it didn’t take markets but a minute to commercialize all of that. And you can be incredibly obscure. You can be Ginsberg and write Howl and suddenly you’ve got a four-book deal. But in this environment, I also think that the old... as you know, one of the things, I have six grandchildren, I do not want them to grow up where a panopticon controlled by a few controls everything. That is a nightmare to me, right? That’s 1984, married to Brave New World, married to, you know, whatever dystopian way you can think about it. But as I was thinking about it, just on your argument alone, that in its very definition is top-down and is going to fail.
Mykhailo Marynenko
Yes. And essentially we will just see scaling bottlenecks. We already see scaling bottlenecks in a lot of these systems that heavily rely on private information that start to collapse or are about to. Again, things will take their own natural order somewhere, somehow. I always believe that there is always a path to good that nature follows that essentially will lead us to better technologies, to companies making different decisions. And even having a new trend in privacy that you can just buy a product a bit more expensive and you have your privacy, like the Apple way, or you can get a cheap phone that has everything that Apple has, like Google’s way, but you can store an infinite amount of photos in your Google Photos but you’re agreeing to all of the terms that are cryptic and you never know what’s happening with your photos in the background. So yeah. And essentially I really love how Apple designs their things. And I think that many people, even many engineers, see Apple privacy tech as a thing that just deeply dissatisfies them because they need to comply with all the App Store regulations and then they need to design user flows where users would be able to opt out of their telemetry or similar things. But even Apple’s architecture down to how processors work and how they are... and what I was mostly excited, and I hope they will not fuck this thing up, is Private Compute. They essentially architected a way for you to use bigger cloud compute in a sense that even engineers having directly physical access to that server would not be able to know what exactly did you compute there. It’s a beautiful piece of architecture and technology that goes down to how processors install, execute instructions and how your phone is in control of your own cryptography. Same way as iCloud has Advanced Data Protection that essentially encrypts everything on their side. There’s a risk if you lose the key, you lose all the data. Even Apple wouldn’t be able to access it. And yes, you pay a premium for it and this should be your choice. And I think that Apple is seeing the right way here. I see that many companies are starting to see like Apple does and I think that a lot of people will pioneer much better ways. Like we are thinking and questioning all of these things while we develop our things because it actually allows us to do more and allows us to make most of our research much cheaper than it was previously with how we handle, again, novel approaches to data and how we store it, how we process it, how we stream it into GPUs than probably anyone else in the market.
Jim O’Shaughnessy
Let’s talk about Apple just briefly because to me, if Jim of 10 years ago was going to put a bet, I would have bet that Apple was the one who figured out the AI or would buy the companies that would make that happen. What’s going on there?
Mykhailo Marynenko
Apple is too ahead of their time. I think that the design and the beautiful concept behind Apple Intelligence and its promises are amazing. I think that they overestimated their own capacity to deliver it. I think that they will still deliver it just in the next 10 years. But I think that Apple would be the first company that would properly actually design it in the right way. I think that Apple got into the space thinking that it’s already old enough, but it’s too young. And it’s definitely far away from being a product that lives up to their level of depth in design and user experience.
Jim O’Shaughnessy
Yeah, because, you know, the investor in me thinks that the first company that is able to pretty much guarantee, nope, your data is truly private, we can’t get it, if you forget the key... that is going to add several zeros to their market capitalization.
Mykhailo Marynenko
It did.
Jim O’Shaughnessy
I know. But then actually proving it to be the case, you’re going to add several more.
Jim O’Shaughnessy
What we have today in terms of what struggles they have is one of the proofs they are genuinely having the freedom to work on things as they see it and they think it’s cool.
Jim O’Shaughnessy
So before we ask that final question, let’s say everything is working the way we want it to work, right? So take me through a day where we get, let’s say, let’s not use an OSV person. We get an outside screenplay from somebody who sends it to Infinite Films. And what will Nick and other people who are working at Infinite Films, once this is working, what are they going to be able to do with just that screenplay that right now seems like, you know, what did Arthur C. Clarke say, that a significantly advanced technology is no different than magic. What is he going to be able to do once that system is working?
Mykhailo Marynenko
For example, we have an Infinite Media email or Infinite Films email, right? Intake from that email would go directly into your canvas, something that we have planned for Canvas to exist. The canvas would take in the script or the concept for a movie and you would be able to run your workflow once. Just take some script beforehand when you design your canvas that would be intake for movies, for example. You would be able to do all of the things like create audience reactions, try to extrapolate to a second part if it’s possible at all, then go do all of these things. Maybe create a dynamic set of criteria that would satisfy us enough to be interested in this and notify me if it goes this way. And then it fucks up a couple times. You correct it a couple times either in chat or manually if you value your time and essentially you get a working workflow that cost you five minutes to build and scales. Then later on if we hit an example that the system is not being able to handle, and this is the cool part about our part and our technology which is runnable graphs, a novel way to approach execution and things that can run and self-expand, we would be able to self-optimize the whole automation or Canvas’s network would be able to self-heal if things go wrong or out of the ordinary or things that haven’t been covered previously. And Canvas would either raise a warning for you that you might want to look at this or it would self-adjust and raise another warning that I looked at this and I needed to adjust because this is this and this was different compared to something that previously ran over the network of things I did.
Jim O’Shaughnessy
So essentially the leverage being provided by these tools and technology... I often talk to people who have really deep, at least business domain knowledge here, and one of the things that I struggle with is they’re not seeing the inherent leverage here the way I’m seeing it. People say, dude, you’re comparing this to Gutenberg and saying this is more important than Gutenberg. I think it is. Do you agree?
Mykhailo Marynenko
It is, maybe, yeah. The last part that I want to mention, like what would happen next, you would be able to just send it in our organization inside Canvas. You can instruct it if you see any other canvases that might use this information...
Jim O’Shaughnessy
Very, very cool. What are you proudest of?
Mykhailo Marynenko
I think that satisfaction would come when I actually deliver on most of the promises, right? A lot of things are highly experimental. A lot of things we discussed is the perfect world we want to see, right?
Jim O’Shaughnessy
Sure.
Mykhailo Marynenko
And I think that we became much closer the past year and I think that we are ahead of the market in many ways. I think that contrary to the market, we have a lot of new unseen problems to solve as well. And things that I’m proud of is that first we’ve got not a lot of hardware but some of it, right? We’ve got our own space in the data center. We built our servers, we installed those, we implemented... it’s a thing I’m so far, one of the three things I’m most proud of, is our hyperconverged infrastructure, our software that manages our data center that allows us to scale like any other cloud would be able to scale, to be able to compete in terms of compute with cloud providers yet allow our engineers to think like our own hardware is a cloud provider so they don’t need to ever change their mental model about how they interact with servers or similar things. Besides this, we’ve made a couple of innovations in the space that are internal that are both allowing us to do big data and compete with bad boys like BigQuery at Google and similar things to scaling and having dynamic GPU allocation and resource orchestration that I think is quite novel compared to the market as well.
Jim O’Shaughnessy
Okay, so what haven’t I asked you that you think is super cool about the work you’re doing and how it relates to everything that we’re doing at OSV and beyond? And beyond. And beyond.
Mykhailo Marynenko
I think that unlike any other company I’ve had experience working with, OSV is my peers. I see insanely creative people that are capable of much more than opportunities previously allowed them to do and I think OSV is great. Innovate. Bell Labs or Apple is great because we have Jim O’Shaughnessy who can help us do our projects and at the same time we actually can drive business better. One can’t live without the other. And I think that the amount of creators, creatives and the way we as a team see the world is amazing and insane.
Jim O’Shaughnessy
Well, I have a lot of friends my age who say the insane part.
Mykhailo Marynenko
Well, okay, I think it’s amazing.
Jim O’Shaughnessy
Well, obviously so do I.
Mykhailo Marynenko
It’s amazing. I wouldn’t be able to meet a lot of tech guys who would be able to operate on these concepts with me, but yet most of the creative people are. And I wasn’t able to find one person in OSV who wouldn’t be as excited as I am about what I develop. So this is definitely my ego thing where I’m like, okay, I’m recognized. This is amazing. And at the same time, the technology we are building is so disparate in terms of how many different things we touch from movie production to podcasts to general media, to markets to finance, to writing technical, non-technical fiction and how these things are connected and trying to make a cohesive system of all of this is probably a billion-dollar question.
Jim O’Shaughnessy
And you know, well, what I think about that, I mean, I think that’s what we’re after here.
Mykhailo Marynenko
Yes, that’s the way also for the future of many things.
Jim O’Shaughnessy
Yeah, I agree. What could go wrong?
Mykhailo Marynenko
What could go wrong? So many things can go wrong. We could completely overestimate a couple of current data capabilities. There’s a couple of R&D things that have shown promise, but it’s not a product, right? And we’re essentially at the stage where we polish a lot of things and we discover how things should have not been made, both existing that are on the market and products we would have loved to use and technologies we would have loved to use. And I think that the complexity of things we’re building from a technological standpoint comes with a lot of risks, but they are well justified in my opinion. I think that we are touching on so many different things and so far we are looking good on them, that even if we do fuck up, it would be a positive-sum fuck up in a way where we wouldn’t be dragged down, we would be able to move forward. I think we’ve made so many mistakes, but first of all, your policy, we never repeat the same mistake yet. And at the same time we were able to drive a lot of things we have today into a much different level. So I think we will make a lot of terrible architectural, design, product mistakes, but those are fixable. They would delay the time to market, they would maybe hinder some of the initial user capabilities.
Jim O’Shaughnessy
I will tell a story about one of our portfolio company CEOs. He met you and I think very highly of him. And he came back and he’s like, yeah, no, you need to hire him. Because I think you were showing him on your phone a knowledge graph, right?
Mykhailo Marynenko
Oh yeah. I was showing him a part of the first R&D that was later able to now enable Canvas, Infinite Canvas implementation.
Jim O’Shaughnessy
Right. And you blew him away because he came over to me and he said, I know startups that have burned through millions of dollars trying for this and they’ve come up with nothing. And he goes, that’s why you should hire him, like right now.
Mykhailo Marynenko
I was hired.
Jim O’Shaughnessy
I know.
Mykhailo Marynenko
That’s the first thing. Second is that I think that a lot matters in terms of why do you approach your work? I don’t approach my work because I needed to make the technology. I had real fucking use cases that I needed to cover. And I needed to not create a temporary solution that would grow into a shitty product. I needed to make something universal, something simple enough at its core that doesn’t require continuing to build legacy things on top. And this forces you to think a little bit out of bounds. I wouldn’t say out of the box. I am in the box. I’m in the OSV box.
Jim O’Shaughnessy
We got a pretty big box.
Mykhailo Marynenko
We got a pretty big box. It’s enough. I don’t need to go outside the box yet. And essentially when you do go outside the bounds and you have the creative freedom to take your time and try to catch what you think would be the right thing and just follow... my God, I have an ability to follow my gut. It’s amazing. And this is what drove us to making something like you just described.
Jim O’Shaughnessy
Well, we could not have done it without you, Misha. That is a certainty. And as you know, my belief is take super talented people who are very agentic and let them do their thing. I despise top-down command and control ways. If you want to duplicate and do what everyone else is doing, okay, fine, but what the fuck are you bothering for then, right? I am... I still wake up almost every day and cannot believe that I am lucky enough to be here at this point in history. As you know, we were talking about it the other day, I was going back through my journals and I’ve been writing about this since I was 21 years old. And I’m like, yeah, your age. And I’m like, finally, it’s finally here. And the unleashing of the creativity that you have, I think is the key. And I just wish more people would think like this because that’s where you get the really great stuff, right? You don’t get the, oh, you know, whenever you turn on one of the streamers, every upgrade is really a downgrade and it’s just the enshittification of everything.
Mykhailo Marynenko
Yes.
Jim O’Shaughnessy
And you have gone exactly the opposite direction, which is why I love the work you do and the way you think. And we are so lucky to be working with you. All right, we got the final question. The final question is, we’re going to wave a wand and we’re going to make you the emperor of the world. Couple of rules, you can’t kill anyone, you can’t put anyone in a re-education camp. But what you can do, we’re going to give you a magical microphone, we’re going to enchant your mic and you can say two things into it. But it’s not going to be just listeners of Infinite Loops who hear that. Everyone in the world is going to hear those two things in their dreams or however you want it to be. And unlike all the other times, they’re going to wake up whenever their next morning is and say, you know what? I just had two of the greatest ideas. And unlike all the other times when I didn’t do anything about them, this time I’m actually going to act on those two things. What are you going to incept into the world?
Mykhailo Marynenko
Don’t be afraid of breaking things and move at the speed you’re comfortable with, but don’t let everyone else slow you down.
Jim O’Shaughnessy
I love both of those. Those are actually quite unique. You might win. But the problem is, what you’re going to win is several of our books, which you already have. Misha, thank you so much for joining us.
Mykhailo Marynenko
It’s been a pleasure.
