“If every website becomes just a chatbot that does everything, then what’s the point of a website?”

Kirill Zubovsky seems to have a lot going on. He’s the mastermind behind the thriving co-working space in the picturesque town of Breckenridge, providing a vibrant hub for creativity and collaboration. But that’s just one facet of his portfolio. Kirill is also at the helm of SmartyNames.com, an advanced AI-powered domain search tool designed to streamline and simplify your hunt for the perfect domain name.

Mike: I want to start out by acknowledging that you own your own name as a domain, KirillZubovsky.com. That alone tells me you know a thing or two about domain names. Can you tell us more about your background?

Kirill: It’s funny you bring up my personal domain because it’s 2023 and nowadays most people either have a personal domain, or they don’t care or don’t even understand the joy of owning a name on the internet. But back in 2008, give or take, I was applying for internships, and I thought that having a personal website to host my resume and email would give me an upper hand. In practice, the companies I ended up applying for (as an industrial engineering student at the time) didn’t really care, I could’ve used @AOL, but the domain stuck. Over the years it’s been a personal blog, a collection of random projects, a business card, and a bunch of other things. I change it at will.

Since we are going to talk about domains here, I gotta say I miss the old internet. I started on the web in late 90s, when the internet was cool. My first website, hosted on some shared server, had a wallpaper, a custom cursors, a wall of confetti, and yes, it was in Comic Sans! It was magical, and weird. Then came Geocities, and Xanga, and ICQ and eventually MySpace. All of those things were canvases that enabled virtually unbound creativity. Internet today though is stale, you mostly get a choice of black or white, which is ironic because we have so many new tools that make creativity possible.

Maybe that’s why I am so excited about AI and what it could do for us; it could turn the internet on its head and force and redo. If every website becomes just a chat bot that does everything, than what’s the point of a website? They will either go away, or they will flourish with new ideas.

But I digress. To answer your question – at some point I got a degree in Industrial Engineering, and then never worked in the field a day of my life. Building products on the internet seems way more exciting to me!

Mike: I’ve always wanted to work in a co-working collaborative building. Before we get into SmartyNames.com, tell me about your co-working space. What drove you to create this? What have some of your observations been?

Kirill: Breckenridge is a small tourist town with fantastic skiing in the winter and hiking and fishing in the summer. It’s amazing for outdoor activities and for spending time with your family, but for that reason, it’s also not exactly set up for computer work. If you are wealthy, you probably have a home office there, but otherwise, housing is a problem, not to mention finding a place big and affordable enough to have space to work. My wife and I needed that space.

One thing led to another and we ended up starting Breckworks, first as a single room with just six desks, to test out the concept. We put a Facebook ad for tourists like ourselves, tech people who needed to work in a ski town. Right away, we got a number of responses. From there, we grew the space one room at a time, adding more features and offerings. It’s a constant work in progress, as we change the space to accommodate changing user needs.

I found working on a physical space to be quite enjoyable. We know exactly who our users are. They are people in tech who need a quiet and comfortable place to work. They need to appear professional, but also to be close to town so they can run off to a hike as soon a Zoom meeting is over. When we deliver on the promise and they have a good time, we get good reviews. While we do need to make sure the space is clean, and it works as expected, we don’t really have to invent much else. I don’t have to convince anyone of anything, and that simplicity makes it very pleasant to operate. Also, because we are a premium space, we get premium business customers, who are very friendly and take care of our space. We don’t have to worry about them trashing the space, or anything like that, because every member wants to maintain a high bar of quality for themselves and the environment around them. It’s just overall a very happy experience.

One thing that I find hard, personally, is to put time into a product that won’t grow beyond itself. What you put in is what you get out, but you can’t get 100x more with the same amount of work. Even if we decided to open more Breckworks around the country, the opportunity is still capped. On a bright side though, there’s going to be a point when we won’t need to spend any time on making Breckworks work. It will just operate and make money, on its own. I think that’s the ultimate scale for this type of business, where it works while you sleep.

Mike: SmartyNames.com is an AI backed platform for finding domain names related to your business. It works by entering, in plain English, a description of your company. The platform then finds available names. What feedback have you received from users?

Kirill: You know, feedback depends on who gives it and when. Smartynames was intended at first purely as a consumer product that would help future business owners to find great available domains. People thought it would completely destroy domain hoarding by making it easier for consumers to find something that was not yet taken. But, as Smartynames got traction, more and more domain investors started using Smarty to find new and unexpected names to buy.

At the very beginning, you could only ask Smartynames one question, and you would only get 10 dot-com domains. It was that simple. Now, we offer tools to generate TLD-specific responses, to generate new domain names based off existing domains, we list daily unique and available pre-generated domains for sale, and there are more features in the pipeline. All of those came as a result of user feedback.

Instead of guessing what the users need, I try to release these features and then watch analytics to see how they are being used to decide what to keep, or how to change them. There’s an upside to domaining being a very old business; because people don’t really change fast, I can make features and observe, without rushing to add more features every morning.

That said, I love simplicity, and personally I think the product got too complicated. I want to get back to the roots of push-button-get-results. There’s a time to listen to feedback, and time not to. For me, every day is an experiment, and there are no bad outcomes, just learning experiences.

Mike: The names generated by the tool are not necessarily available as hand regs, but as premium names. Is it fair to say you make an affiliate commission on names that are purchased? Is that the business model?

Kirill: Yes. I have connected to various affiliate programs. That revenue stream is doing okay, but at the same time I also have some ad revenue coming from the stellar team at Lumis, and I am working on a few new partnerships to grow that side of the business. Being an affiliate is not really a great long-term strategy because while I get some revenue, I don’t get to keep most of it. Essentially, whenever someone buys a domain, I get a fraction of their first-year purchase, but the domain registrar gets to keep the rest, and they get paid every time that domains renews. There’s also a fairly large number of sales that I lose to registrations which avoid affiliate links; where people find a domain and then go to register it elsewhere, on a registrar that I don’t connect with.

While I could probably invest time in becoming a reseller, for example, I frankly find it boring, and I haven’t gotten that far yet. Instead, I am more interested in exploring domain-related features that have not been done yet, or have been done but poorly. Those are the things I want to focus on first. If I can do something a thousand times better than others, money will follow.

Mike: How did you first get involved in domain names?

Kirill: It was a pure accident. I needed to find a domain for a project, and I didn’t want to brainstorm it myself. So I built a tool to do it on top of GPT-3, at the time, and once it did shockingly well, I ended up putting it together into what you see today. In fact, Smartynames came up with its own name and logo. It was very amusing. If you are curious to know the full story, I did an interview with Geekwire, Putting AI in Domains. There’s a podcast version there too, if you prefer to listen.

Mike: Some people are oblivious to the recent surge in AI. Others are all over it. What is your prediction for where AI will take us in the next 10 years? How do you think it might impact domain name investing, if at all?

Kirill: I think AI is incredibly overhyped in the short term, as everything new and cool tends to be when it first emerges, and at the same time will have a profound effect on how we live and work.

It’s one of those technologies that you cannot put back in the bottle, once it’s available. Currently we see a lot of AI in text processing and imaging, but that’s because those are easy to comprehend and to experiment with. What happens when we apply the same techniques to any kind of data processing?

Financial trading, healthcare records, government paper work, teaching materials, news generation, tax filings, insurance filings …etc. Pretty much everything that you do, or consume, or rely on, could be changed with AI. The impact is going to be astronomical.

To really take advantage of the AI we need to build models that we can 100% rely on, and put our trust into. If you are going to file taxes in one click, you and the IRS need to know that once you’ve uploaded a stack of papers that got processed by the AI, the outcome of those papers will be 100% correct. As a way to get there, we are first going to see augmented uses of AI, something that helps you do the filing 100x faster, but leaves the last step of verification to you. That however is not going to last long; once we learn to do things 100x faster with AI, we will stop worrying about it being 100% correct. In most cases, it’s better to be done than perfect.

So as far as domain investing goes, same applies. AI can’t help you invest better or sell better just yet, but I can already take your portfolio, and use AI to augment coding in order to make a prediction of what domains are likely to sell faster, for example. I think if we put a few smart domaines together with a technologist like myself, we could quickly generate tools that would leave many investors in the dust.

You may have seen, I wrote a blog post explaining the basic math of domain investing, as I came to understand it. With a large enough portfolio, and a landing page to sell it through (smartynames.com), it’s possible to make a lot of money. However, you got to have a bag of cash to start with, and be pretty good at math to do it well.

What worries me more than AI is what big companies will do to harness that power. For example, GoDaddy had already filed a patent on a system that uses AI to generate domain names. It’s a very vague patent and probably not defensible in court, but you have to have the money in order to fight them; most people won’t. We already have a system where a handful of companies owns rights to most of domain extensions on the internet. What happens when they also dictate what you sell, and how you sell it, and what tools you are allowed to use to do it? What happens if they build the tools to buy and sell, and use it internally, leaving you with the scraps? We might be in a situation where domain investors are left holding a very expensive bag of junk. I hope this doesn’t happen, but there are no safeguards against it right now.

Mike: You are no stranger to trying new things. I see SmashNotes, Silly Cone News, and a past podcast. Do you find that trying new things is a good way to find your niche or do you pursue them for other reasons?

Kirill: Yup. The best way to learn for me is to do stuff. I didn’t know anything about domain names before I started Smartynames, but now I can talk to people on the subject and ask questions, and experiment. What you see on my website is just an icing on the cake; I’ve done dozens other projects which I either shut down, or keep unreleased to the public. I think to me, doing stuff on the internet is like working on a car to many people. I get to take stuff apart, and put it back together, and polish parts, and it gives me joy. The biggest difference is that I get to learn a bunch in the process, and I am not constrained to a single car, or a car at all.

I have been slacking a lot this summer as my family and I have been living out of a Sprinter Van for the last 2 months, traveling around the country. We have Starlink, so in theory I could work every day, but in practice between driving and hiking and playing, I put work on the back burner, for better or worse.

But again, the desire to travel like this came from experimentation. If you listen to my GeekAtSea.com podcast, I’d interviewed a number of people who live an alternative lifestyle, or those who tried something unexpected for a few months or a few years. They were all full of great stories, and learning experienced, and over time we came up with our own version of what we’d like to try.

It’s really hard to learn new things if you are stuck in the same place, doing the same thing, and so I love to move. Well, correction, I actually kind of hate moving, but I find it a necessary evil too. If you keep making yourself uncomfortable, eventually that becomes the only stability that makes you feel good, and everything stable becomes uncomfortable. Funny how it works.

Bottom line is I like to learn and to stay curious. I am currently writing a sporadic newsletter on AI called Novice.Media (yup, that’s a domain). I suggest you read the newsletters I’ve already sent out and if you find something interesting there, please subscribe and we’ll explore together. Today it’s AI, but tomorrow I might love something new, and we’ll explore together.