"Now is the best time to start a startup": What Yuri Melnichek told the Warsaw community

How to build startups when venture capital is in crisis

Spoiler: without money.

«Venture funds, though reduced by half, still exist. Projects showing rapid growth in key metrics will still get funding.

But now is the time for startups that develop not on lavish venture capital, but on founders' money or much smaller investments, at least initially.

Thanks to LLMs (Large Language Models—devby) and AI, creating startups has become cheaper. Small teams can achieve amazing results that were impossible just two years ago. Plus, Microsoft, Google, and Amazon still offer substantial credits for using their hosting services. You can definitely build a proof of concept on these. They might not be enough for scaling, but you can create a prototype yourself with almost no budget.

Money aside, we now have a magical box with a 'genie' that can grant wishes. True, as in fairy tales, the genie doesn’t always do exactly what you ask. And there’s the problem of protecting against copying. But the possibilities remain enormous. It’s the best time to start a startup.»

What’s more important for a startup—revenue or growth?

«It depends on the investors. But generally, in tough times, everyone leans toward revenue because you can survive on it. On the other hand, there are still investors interested in growth.»

On deglobalization and its impact on startups

«It’s clear that the world is fragmenting—China is a separate region, Russia is separate, and the West is separate. Europe may even be separating from America. But I still like the idea of creating projects for the entire world: it’s easier to localize a solution for different languages and countries than to solve a problem from scratch.»

On the US elections

Spoiler: Melnichek prefers Trump (at least while Biden was the alternative)

«I really hope the US changes its leadership and economic policy, and in a year the world returns to previous economic growth. That’s my wishful thinking.

I believe the economic situation in America was better under Trump. And if Trump wins, it will improve. Biden proposed introducing a super-tax for the wealthy, which would even affect Google programmers. This tax would cause even more capital outflow from the US, leading to a greater economic downturn.

I think ultimately something reasonable will prevail. And when the new economic cycle begins, the economy will finally return to growth.»

On closed systems

«Once there was an idea of an open web: everything we upload belongs to humanity. This spawned many new businesses, including Google as a search engine for the open web. And it largely created the current AI, which was trained on the open web.

Without the open web, there would be no AI. In this sense, I’m concerned that current systems are being built as closed ones. You can’t crawl Facebook or train anything on Facebook posts. Telegram is also not an open system, which prevents the growth of companies that could become future 'Googles'.

However, there’s the magical AI that makes building products easy. Open AI trained a model, and now everyone is suing them. But other teams are distilling Open AI’s model. They ask ChatGPT questions, get answers, and train their smaller model on them. This way, Open AI protects them from lawsuits.»

On companies breaking ties with Belarus

Why a startup from Belarus suddenly becomes a «Lithuanian unicorn» and why Melnichek himself doesn’t hide his connection with Belarus

«Because if you call yourself a Belarusian unicorn, your horn gets sawed off (laughter in the audience). The valuation multiplier immediately becomes lower.

It’s one thing for me, and another for a company. I simply don’t understand how this [Belarusian origin] can be hidden. It’s easier for companies: you can relocate the office, employees, intellectual property, and say we’re local. People’s nationality doesn’t matter.

But when you’re preparing to attract large foreign investments or preparing for an IPO with many qualified investors, the fewer connections you have to what reduces valuation, the better. EPAM also doesn’t position itself as a Belarusian company. People just see risks in it now. And this immediately affects the value, especially if the company is public.

My attitude toward [hiding Belarusian roots] is this: I don’t hide it myself, but I don’t condemn those who do. I understand why they do it. Does it affect Belarus, its economy or politics? It doesn’t affect it at all.»

On Poland’s cooling attitude toward Belarusians

(cancellation of PBH and all that)

«It’s not so easy to relocate a company from Belarus anymore. But, on the other hand, maybe there’s no need for it now. We have a company in Poland with a founder and employees from Belarus, and we don’t feel any discrimination.

If conditions become bad, we’ll build the company elsewhere. There’s still Germany, Cyprus, England, Asian countries. My company (apparently referring to Gero startup—devby) moved to Singapore.

Of course, Poland has a super-close and understandable culture for us, making relocation here most comfortable. The choice of countries has narrowed, but options remain. However, I haven’t experienced negative attitudes here, either personally or business-wise. If it becomes uncomfortable, people will move further.

There are people and there are states. I think it’s right for people to put states in a competitive position. Offer me conditions, and if you suit me in culture, politics, economy, and vibe, I’ll come to you. If you deteriorate, I’ll go elsewhere. This is the force that will make states more livable.»

Futurology

What if AI takes off, AGI (Artificial General Intelligence) emerges and takes people’s jobs? What will happen to the world? And to programmers?

«We’ll see. Robotics lags far behind AI. While in AI, we’re asking whether humans can distinguish an image made by generative networks from a photograph or artist’s painting, in robotics the challenge of a robot inserting a key into a door lock, turning it and opening the door is still relevant. It’s possible now, but robotics hasn’t gone much further. So manual labor jobs are still safe for now.

For me, the main thing now is the new opportunities that allow you to create your own business. There have never been better opportunities.

What if a model emerges that handles many tasks better than humans?

«We’ll go live in virtual reality, play computer games, engage in creative activities.»

Where will the money come from?

«Money is backed by economic activity. If the model replaces economic activity, the latter will be easier to scale.»

You said that now it’s not so important to know how to create products as to understand what needs to be done.

«Yes. If previously it was impossible to create an IT product without programmers, now it’s possible, albeit crudely, and this barrier will only continue to fall. Even among my acquaintances, there are people who program using Anthropic: they tell it what to do, copy-paste code, then look at the errors.

If previously a startup needed a technical founder who hired programmers, now programmer productivity has grown, and business and product skills have become more important. I’ll probably get hated for this.»

So the programming profession might disappear?

«Not in the next few years. But the need for programmers will decrease.»

Which profession should we switch to?

«This year it’s prompt engineer. You can learn to formulate requests in a week and then create magic. You can make a bunch of products if you learn to write prompts correctly and tape together several components.

Looking at the long term: once there were family businesses (passed down through generations), then they disappeared, and people chose a profession for life. Now people frequently change professions, and this trend will only grow.

In one of my startups, we had a concept: we need people who are good at doing what they’ve never done before. That’s a good basic skill. You won’t go wrong with it.»

What comes after AI

Biotech

«I’m a big biotech fan. There are many problems to solve there. Whatever solves the aging problem will become a super-expensive company because it will create enormous value for the world.

DNA is also code. When we learn to program it, paths will open to stopping aging, eliminating diseases, and improving ourselves. Perhaps some fusion of humans with computers will occur.

There are 8 billion of us, and only 50 million will avoid dying from age-related causes—the rest will not. More people currently die from age-related causes than have been killed by dictators throughout human history. If we count lives, this [aging] is the most important problem, against which practically nothing else matters.»

Current scientific research shows that rather than slowing down old age, we can stretch adulthood, the period of productive life before frailty sets in. And further—if we reach neighboring planets, we can live much longer beyond our solar system.»

Who will pay the pensions?

«Robots will work.»

Will there be enough resources?

«Only matter and energy matter. There’s super-plenty of both on the planet. A 'thermonuclear reactor' nearby generates energy and sends it to us; the question is only our ability to process everything.»

Even if a person remains productive, mentally they still become outdated. They might become an unpleasant person.

«That’s an interesting question, whether people get stuck because of brain aging or because they’re entrenched in their system. I don’t see a problem [with people living much longer], I’d like my parents to live longer, great scientists to live longer, and generally for people to live longer.»

Do you take anything against aging yourself?

«Nothing at all. (Laughter in the audience). I’m waiting for serious pharmacology. I think it will appear within the next 20 years. If during these 20 years you don’t get the first disease that marks aging (cancer, type 2 diabetes, etc.), there will be very good chances of living until the moment when life extension outpaces life reduction.»

Fun fact: cartilage grows continuously. Even if aging stops, your nose and ears will continue to grow. You’ll need another medicine to stop this growth.»

How likely are we to live to 100?

«Most people here will die at 100+. Unless we all die at once (if nuclear war doesn’t happen).»

Where Melnichek keeps his money

«Bonds aren’t for me, they’re low-yield. In the stock market, September is supposedly the worst month of the year. And then there will be some political upheavals—it’s still unclear who will come to power in the US.

These are weekly term deposits, so the maximum time to unlock the money is one week.»