Sergey Tokarev: Where lies the future of AI

Sergey Tokarev Where lies the future of AI

In the past, the debate around artificial intelligence (AI) was about the time when AI develops to a such level so it can perform simple monotonous human work, yet people believed it will never be able to replace creative professions.

Three years ago, a breakthrough in AI showed the opposite result: technologies learned to work with text and images so well that it made specialists working in creative areas stressed.

Talking about AI replacing professions, Sergey Tokarev, Founding Partner at Roosh and Reface investor, emphasizes that high-quality professionals not only will not be replaced but at the moment they are extremely lucky to get another tool that can significantly ease their work.

“We are at the beginning of a new technology development cycle and the speed of adaptation to new conditions will directly affect your relevance in this cycle,” believes Tokarev.

Artificial intelligence helps to better systematize and analyze data and transform it into a much more concentrated and concise piece of information. Sergey Tokarev compares it to the development of the Internet: the development of AI can radically increase the speed of data transmission and processing.

“Look at GPT-4, Dalle-E, and Midjourney from this angle and you’ll see that AI is actually a complement to human intelligence, not a replacement. The artist does not just invent an image in their head, they describe it for themselves and, if necessary, can transfer this description to the text,” explains the Reface investor. 

Tokarev adds that this process can also work in the opposite direction: prompt engineering uses the description of the job that the AI is expected to do and includes the input, for example, as a question. Most people though are not yet able to effectively use new AI services precisely as they cannot create prompts.

Let’s say, you want to draw a green leprechaun. The obvious prompt is “Draw me a green leprechaun.” Yet, it is a bad prompt as there is a lack of information and the system will fill the gaps with assumptions so the result will be quite unpredictable. You should formulate your request in a special way considering context, all possible details, style, etc. A prompt generator is a new profession that emerging right now.

“Right now, you can open GPT-4, ask it to create a prompt for Midjourney, provide the chat with the technical specifications of the latest Midjourney model and examples of other people’s prompts, and ask it to generate the correct detailed query,” says Sergey Tokarev. “Try it, the result will be much more awesome than setting the prompt all by yourself. Your images will become photorealistic.”

Creative people should be the ones who run this phenomenal process. Now, you don’t need to know how to use every tool of graphic editors to see the quick implementation of your idea. You can almost instantly generate the details and build full-fledged work.

The amount of data that people have to deal with in their work and personal life is constantly growing. Finally, there is a tool that will help you work with them faster, and not spend plenty of time processing them.

Due to the data overload, people tend to work with it incredibly slowly. Thanks to AI, we are equalizing the amount of information available and productivity.

“We are currently working with an “infant technology” that can only solve basic tasks. In two years, the child will grow up and be able to produce much better results,” believes Tokarev.

Here are a few areas that AI can change for the better.

“When ChatGPT appeared, I immediately thought of using it as a search analytical engine. The main problem of search is the prevailing of advertisers as search results show a biased picture. Let’s say I want to find the best washing machine in 2023. What should AI do?”

Tokarev explains that the first thing AI does is search and select all sites with reviews of washing machines and create a table with all options. Then, it addresses Amazon, Ebay, and other retailers and collects reviews from real buyers.  And lastly, AI will combine these two tables and provides you with the result: “Probably, this model is the best choice.”

“This result will be less affected by advertising and at least 50 percent will be created based on real reviews. This is not absolute objectivity, but it is a better solution to many questions than just a list of links to various sites,” says Tokarev. 


The increasing amount of information, social connections, and data that we have to deal with at work overload our brains. We simply cannot process this data. Why not delegate part of the routine tasks to an assistant? Not just any assistant but a program that learns only from your data, is not connected to advertisers, does not transfer your data to any third party, and ensures you are the only user and owner of the information. The assistant reads your mail and writes in the messenger: “This email is not spam but from a person from your contacts. It is waiting for your reply for five days.”

“Such an assistant will perform every task you assign it with. Plus, it will be much easier to trust an artificial assistant who does not show any emotions,” says Sergey Tokarev.


Unfortunately, or fortunately, GPT seems to replace lawyers, especially international ones. No matter how good a person’s memory is, they cannot keep in their head a large amount of legal information about the peculiarities of the law in each country, while the algorithm easily analyzes tons of documents and finds legal grounds for the request.


Personalizing information for each student is an impossible task for a teacher. The algorithm will be able to adapt to the habits of its audience and provide information in the right way. Without losing the most important information, everyone will receive new knowledge most favorably.

Support services

This change is already on its way, and the development of AI will accelerate the trend. Algorithms can find answers to questions about the product or service much faster and even voice them as voice synthesis is no longer a problem. In a few years, it will be unclear who is talking to you: a real person or one created by AI.

Yet, what prevents the AI revolution from happening right now?

“Only an unprepared legal system hinders the development of AI. Today, we have laws that are different in each country, and we have individuals. So, it is difficult to answer who is responsible for AI: the author of the algorithm, a user of the algorithm, or the algorithm itself,” says Sergey Tokarev.

Who is to blame if an algorithm can help a dying person but doesn’t? What will happen if its actions cause harm? Under the jurisdiction of which country will liability be assessed?

There are no answers to these difficult questions yet. The world is based on a legal system that was developed centuries ago, but it will have to change and define AI regulations. Should an algorithm follow human moral standards? What is ethics and who should monitor its compliance?

The algorithm learns from answers and information from living people, from their ethics and morals. If we want to change AI, we have to start by changing people. 

“We need to label content and information generated by an algorithm, not a human, but we can’t allow an algorithm what humans are allowed to do. Decisions of society and regulators on this issue will radically affect the development of this technology and its adoption by businesses and people,” believes Tokarev. “AI should be used as a tool for innovation and process optimization. A key skill of a professional these days is to adapt to constant changes. If there are tasks in your career that can be performed by AI models, now is the time to be the first and start using them.”

Automating some of the manual work is a continuous process that has been increasing our efficiency since the Industrial Revolution.

Resistance to tech development turns us to the Luddites of the 21st century, who want to destroy machines instead of using them to create products and services of a qualitatively new level.

“Embracing changes is perhaps the most important task for your career right now,” says Sergey Tokarev.

Related Post