Losing a business in a day and finding yourself in AI | Sergey Tsyptsyn

No Bullsh!t, September 24, 2025, 75 min, 1,406 views. The guest is Sergey Tsyptsyn, who tells his story: from a university researcher in the nineties, through selling computers and working with graphics workstations, to building a computer graphics school and then the largest 3D graphics conference in the region, which grew to thousands of participants.

Episode at a glance
Published
Length75 min
Views on YouTube1,406
Language of the episodeRussian
Where to watch and listenYouTube, Spotify, Telegram

Date, length and view count according to data from the episode page on YouTube, as of the page update date shown in the footer.

In short

In February 2022 all the conference's Western sponsors left the market at the same time and the business collapsed in a single day. Tsyptsyn goes through his main mistakes: excessive caution and turning down risky ideas, an unwillingness to think about the future, and failing to diversify assets. Out of that crisis came his current project: live meetups and a Telegram channel about artificial intelligence that grew organically to 43 thousand subscribers. The second half of the conversation is about generative AI, language models, image generation and where video is going.

Key ideas

  • Losing a business in one day teaches that you cannot tie a company to a narrow circle of sponsors: diversify income sources and assets rather than putting everything in one basket.
  • Organic growth of a Telegram channel, 43 thousand subscribers without buying traffic, gives a far more engaged audience than paid acquisition.
  • In generative video the hard part is not the picture but consistency of an object over time: every entity in the frame has its own timeline, which is why clips are still short.
  • The trend of August 2025 was the move from generating single images and clips to generating whole three-dimensional worlds you can walk into.
  • The mass arrival of AI tools does not replace directing and camera craft: without an understanding of story and editing, generated content stays a technical trick rather than a story.
  • Large studios spend years extending commercial software for their own needs and do not share those extensions, which is why real use of generative tools in film lags behind the demos.

The full transcript

The episode is recorded in Russian, and the full transcript of the conversation is on the Russian page of this episode. English and Spanish subtitles for the strongest episodes are rolling out on YouTube in waves.

Open the transcript