Skip to content
Red-and-gold phoenix illustration with spread wings and long curling tail feathers.

Creativity For Introverts




Sign up to our Email or FB page.

    • About
    • Ivan Topp – achievements
    • Koroteev Aivan
    • Podcast
    • Privacy Policy
    • Public Speaking Training for Software Engineers and IT Teams
    • Who is Vano Karade, the ai-musician?
    • Why I Go by Aivan
Illustration of a bird flying.
  • From TikTok Marketing to AI Filmmaking: My Speaking Journey with Faberlic and IBS Training

    From TikTok Marketing to AI Filmmaking: My Speaking Journey with Faberlic and IBS Training

    In May 2021, companies were trying to understand TikTok’s recommendation system and why short videos could reach people who had never followed the creator. Five years later, the question has shifted: how can we build an entire video scene with generative AI while maintaining control over story, camera, character and editing?

    Photo from Ivan Koroteev's public-speaking portfolio
    From my public-speaking photo archive

    I recently rediscovered two official pages that capture this change in my work. One is a Faberlic announcement of my TikTok marketing livestream. The other is my instructor profile at IBS Training, along with the provider’s current AI video production curriculum. They tell a story about changing technology—and about what it means to explain technology in public.

    2021: A TikTok webinar for the Faberlic audience

    Faberlic is a beauty and direct-sales company. On its official website, it announced a livestream titled Promotion on the TikTok Platform for May 20, 2021, naming me, Ivan Koroteev, as the presenter. The advertised agenda covered why TikTok was gaining attention, how its recommendation algorithms worked, how videos attracted views and how creators could communicate about Faberlic products.

    The announcement introduced me through several activities I was pursuing at the time: a Bauman engineering education, TEDx speaking, my Russian-language research portal VerticalVideoPortal, the podcast The World of Vertical Video, and the VerticalVideo agency for businesses. It also used the promotional description “number one” for the portal; I treat that as the organizer’s wording, not an independently established ranking.

    The listing links to a YouTube video for the session. An event announcement is evidence of a scheduled appearance—not of attendance figures, results or any commercial arrangement. Those details are not published in the source, and I won’t invent them.

    What interested me in TikTok was not simply a new aspect ratio. It was the combination of creative tools and machine-learning discovery. A video could find an audience through its content rather than depending only on an existing follower graph. To communicate that to business owners, I had to translate algorithms into tangible decisions: what to make, how to explain it and what to measure.

    2023: Making AI understandable beyond social media

    An additional independent milestone appears on the website of the Russian knowledge society Znanie. It published a November 23, 2023 lecture about the adoption of AI in Russia, identifying me as a neural-network specialist and a finalist of the national League of Lecturers competition.

    By then, AI talks were no longer mainly about content recommendations. Business and education audiences increasingly wanted to understand where models could be useful, which workflows should change, how to evaluate results and why people still needed to supervise important decisions.

    2026: IBS Training and AI video production

    My public IBS Training instructor profile describes 13 years of experience in corporate IT systems, project work with AI and GenAI, TEDx speaking in Russian and English, and over 50 talks on technology and AI. Its technical topics include LLMs, JSON, REST, ERP and .NET. I like that combination because it reflects the two sides of my work: building systems and explaining them.

    IBS also lists me as the instructor for AI Video Production: From Script to Finished Video, a 16-academic-hour course. The public syllabus covers structured JSON prompts, cinematography, framing and camera movement, character continuity, image-to-video workflows and tools such as Kling, Veo, Seedance and Wan. According to the course description, 53% of the program is practical work.

    The provider published a cohort schedule of November 9–18, 2026 and identified me as trainer. Those dates are a course listing, not proof that the entire cohort has already completed training. IBS separately listed an August 25, 2026 webinar on AI video without a camera crew, including structured prompts and a repeatable video-content workflow.

    The technology changed. The storytelling problem did not.

    When I spoke about TikTok in 2021, the camera was usually a phone. Today a shot may begin as a generated image and continue through a video model. Yet a beautiful generated frame does not automatically create a good film. You still need motivation, action, rhythm, continuity and a reason for the viewer to care.

    I have become increasingly interested in the process connecting those elements: idea → script → scenes → keyframes → motion → edit → publication. A model can accelerate parts of that process. It cannot replace the creative decision about what story is worth telling.

    Why I now train technical teams to speak

    There is a second lesson in that history. I came to the stage from software engineering. I know what happens when an engineer understands a system deeply but struggles to convey it to a customer, a leadership team or a conference audience. Sometimes the architecture is excellent; the explanation is simply missing.

    That is why I have launched a dedicated public-speaking and communication workshop for IT teams. Developers and architects practice explaining decisions, delivering technical demonstrations, organizing slides, using voice and pauses, and handling questions. The goal isn’t to turn every engineer into a showman. It is to help the right ideas reach the people who need them.

    A research portal, a corporate TikTok webinar, an AI lecture and a structured filmmaking curriculum may look like different chapters. For me, the connecting skill is the same: turning complex technical possibilities into practical human understanding.

    Independent references and recordings

    • Faberlic — TikTok webinar announcement, May 20, 2021 (Russian)
    • YouTube link referenced in Faberlic’s official announcement
    • Znanie — AI adoption lecture, November 23, 2023 (Russian)
    • IBS Training — official Ivan Koroteev instructor profile (Russian)
    • IBS Training — AI video webinar listing, August 25, 2026 (Russian)
    • IBS Training — 16-academic-hour AI video curriculum (Russian)

    Book a keynote or workshop

    I speak in English and Russian on applied AI, business transformation, AI filmmaking, music and creativity, and communication for technical teams. My sessions can be adapted for conferences, universities or corporate training.

    See speaking topics and booking information, or contact me first on WhatsApp (+7 707 401 69 33), followed by Telegram @ivantopp.

    Source note: IBS and Faberlic maintain their own public program descriptions. This article supplies my context and interpretation; it does not imply that the two companies collaborated with each other.

    November 13, 2026
  • What I Learned From Getting Creators to Try Avatarify

    I worked on a creator campaign for Avatarify, an app that animates still photos. The format was easy to understand in a feed. You see a picture, then you see it move. A creator could get to the interesting part before they’d even finished saying the name of the app.

    Hashtag page for Avatarify with example tiles of a person, two children and a cat.
    An archived TikTok hashtag page showing examples of Avatarify clips.

    In August 2022, I wrote that the collaborations had generated more than 200 million views and that Avatarify had received more than 32,000 App Store reviews. The view count tells me something about reach. I didn’t publish a breakdown showing how many viewers installed the app or went on to review it. Those are the numbers I’d want alongside the views today.

    The demo gives the creator something to do

    Think about how you’d explain the app to someone sitting next to you. You’d probably open a picture and show them what happens. That’s a useful starting point for a video brief. I’d give creators that one action and let them pick an image, a setup, and a reaction that belong in their own feed.

    If I were doing it now, I’d add one line to the brief. Use a photo you have permission to animate. With an app that changes how someone appears on screen, the choice of photo is part of the creative work.

    What I’d measure now

    1) Views tell me how far the clip traveled. I’d compare clips that show the result at different points in the first few seconds.

    2) A tracked link, where the platform supports it, helps me see which clips sent people to the store.

    3) The first completed animation tells me whether people got as far as trying the thing they saw. I’d ask the product team to measure that before the campaign starts.

    I published the original numbers in my 2022 article for Intch. For the brief itself, I’ve written more about working with creators beyond follower counts.

    If you had five seconds to show what your product does, what would you put on screen?

    November 13, 2026
  • My First Stand-Up Set, My First Paid Talk, and What I Learned Between Them

    In 2022 I tried stand-up for the first time. That was also the year I competed in public speaking in Kazan, gave my first paid talk at SOK Innopolis, and did an English-language TEDx talk. If you looked at my calendar, you’d probably think I had a clear plan. What I had was a bunch of different rooms to try.

    Man with an event lanyard poses next to a large pink and purple League of Lecturers display.
    From my speaking archive, at the League of Lecturers.

    I’d already reached the final of the first League of Lecturers competition in 2021. In 2023 I went back with a talk about neural networks and reached the semifinals of the third season. In between, there was a speaking competition, a workshop with actress Greta Saburova, and a comedy set. Each gave me a different way to see what an audience was doing while I spoke.

    What I’d practice in each room

    1) In stand-up, I’d work on timing. A joke gives you immediate feedback. Try saying the same line twice, once with a pause and once without it. The words can stay exactly the same and the room can respond differently.

    2) In a speaking competition, I’d work on the shape of the talk. Give yourself two minutes and record it. Then play it back and see how long it takes to reach the first sentence you’d actually want a listener to remember.

    3) For a paid talk, I’d ask the organizer one thing early. What should people be able to do the next day? That question helps me choose which example deserves five minutes and which one can go.

    In 2023 I gave more than 30 public talks on AI. When I think about how to get ready for a talk today, I still like the small experiments. Try a different opening. Leave a pause. Ask someone what example they remember a day later.

    I’ve written separately about how I approached TEDx. The timeline for the competitions and talks is on My Russian-language achievements page.

    If you had two minutes and a room full of strangers, what would you try first?

    November 13, 2026
  • What 40 Computer Science Lessons Taught Me About Explaining Things

    In 2023 I made 40 computer science lessons for grades 8–11 for Rostelecom Lyceum, an online learning platform. That year I was also teaching Computer Science in English at Moscow Pedagogical State University, using a University of London curriculum. Same subject. Two very different jobs.

    A student in the room can stop me and ask what I mean. A student watching a recording can pause, rewind, or leave. When I plan a recorded lesson now, I try to find the spot where I’d want a student to interrupt me and put the example right there.

    Online eighth-grade computer science lesson with a video presenter in a beige shirt.
    An eighth-grade computer science lesson on Rostelecom Lyceum.

    Start with something the student can do

    Take algorithms. I could start by asking the class to remember a definition. Or I can ask someone to write instructions for a classmate to find a book on a shelf. Then we swap the instructions. How do I know which shelf to use? What do I do if the book is missing? Now we have a reason to talk about steps and conditions.

    If I’m teaching in English, I can put the exact technical word on the screen, say it once, and use it again in the example. The word “condition” is easier to remember when you’ve just had to decide what the instructions should do when the shelf is empty.

    Three checks I’d use today

    1) Can I say what the student should be able to do by the end? “Explain what this condition checks” gives me something I can actually test.

    2) Does the example work with the tools the student has? A sheet of paper is enough for some lessons. For others I want the student to run the code and change one input.

    3) Can the student predict the next step before I show it? In a recorded lesson I can give them a few seconds to pause and try. In a classroom I can ask someone to explain their guess.

    Background on the course and my teaching work is on My Russian-language achievements page and in my university profile.

    What was one computer science topic you only understood after someone gave you a real example? I’d love to hear what the example was.

    November 13, 2026
  • Featured in Khaleej Times: My AI Music Tribute to Abu Dhabi in 22 Songs and 15 Languages

    Featured in Khaleej Times: My AI Music Tribute to Abu Dhabi in 22 Songs and 15 Languages

    Media coverage · 17 July 2024. UAE newspaper Khaleej Times published a feature by journalist Anamika Chatterjee about my AI-assisted music project, A Symphony of Cultures — Tribute to Abu Dhabi. The feature, titled “22 songs, 15 languages: This AI-created album pays tribute to Abu Dhabi”, explored both the story behind the album and my belief that human experience gives technology its creative direction.

    Artwork for A Symphony of Cultures, Ivan Koroteev's tribute to Abu Dhabi
    Album artwork from A Symphony of Cultures — source: Creativity for Introverts. View the newspaper’s original photo coverage ↗

    For me, the most meaningful part was that a personal encounter with a city could turn into a multilingual creative experiment—and reach a newspaper in the UAE. This was not a conventional music-industry project. I came to it as an engineer, a speaker on AI and a traveler curious about how technology can help people connect across cultures.

    Why Abu Dhabi?

    The project began with travel, not a prompt. As I recounted to Khaleej Times, my wife and I first visited the UAE in November 2023. Returning to Abu Dhabi in March 2024, I was struck by how the city combines ambitious development, public spaces, cultural institutions and different communities.

    The article describes visits to Louvre Abu Dhabi, Sheikh Zayed Grand Mosque, the Corniche, Emirates Palace and SeaWorld Yas Island. These were not locations invented for a promotional concept: they were part of the journey and feelings that inspired the songs. I wanted the album to express gratitude and curiosity rather than merely advertise a destination.

    Twenty-two songs, fifteen languages

    A Symphony of Cultures brought together 22 tracks in 15 languages. Those numbers became the headline of the Khaleej Times story. My intention was to experiment with the possibility of one place being celebrated in different musical and linguistic forms, rather than forcing every listener into a single language.

    The article explains the workflow I used in 2024: developing lyrics and ideas with ChatGPT and Suno, experimenting with genres including pop, rock, rap and electronic music, testing Udio, selecting outputs and preparing the tracks for distribution through Offstep. The newspaper reported that I completed the album over about three days. It was a rapid creative experiment—not a claim that every track was composed and performed by a human band.

    At the time, the music project appeared under the artist name Vano Karade. I have since also used Aivan Karade for my AI music and multilingual cultural work. These names reflect an evolving artistic identity, while my professional work continues under Ivan Koroteev and Aivan Kor.

    What the interview said about AI and creativity

    An important part of the conversation with Anamika Chatterjee was the distinction between generating content and choosing what the content means. AI tools made it possible for me to explore arrangements, styles and languages that would have been difficult to produce alone on such a short timeline. But the decision to create a tribute, the choice of Abu Dhabi, and the personal experiences behind it came from a human being.

    I still see AI as a creative instrument. It can shorten the distance between an idea and a draft, but it does not automatically supply attention, taste, cultural understanding or purpose. Listening critically, selecting what works and being transparent about how a piece was made are part of the creative responsibility.

    Why this press feature matters to me

    The Khaleej Times article connects several parts of my life that might otherwise look unrelated: engineering, travel, multilingual communication, music and public speaking. It also documents this project at a specific moment in the evolution of generative music tools—July 2024. The technical landscape has changed since then, so the article is best read as a historical feature and an independent account of that creative experiment.

    I continue exploring how AI can help us make films, music and educational content while keeping people and culture at the center. The experiment with Abu Dhabi was an early example of that direction.

    Read the original feature

    Publication: Khaleej Times (UAE)
    Journalist: Anamika Chatterjee
    Published: 17 July 2024
    Headline: 22 songs, 15 languages: This AI-created album pays tribute to Abu Dhabi

    You can also read my earlier reflection on the Abu Dhabi project and explore my later multilingual AI music work.

    Editorial note: This is my original account of the project and its media coverage. It is not a republication of Khaleej Times reporting or an endorsement by the newspaper.

    Book a talk on AI, music and cross-cultural creativity

    November 13, 2026
  • From Entertainment to Checkout: What Douyin Taught Me About Discovery Commerce

    What if shopping begins before the customer knows what they want to buy? In conventional online retail, a shopper enters a query, compares listings and moves toward checkout. In discovery commerce, a creator’s demonstration, a story or a livestream introduces a product during an entertainment experience. That difference helps explain why Douyin, China’s sister product to TikTok, became such an interesting case for retail technology.

    Man in a red hoodie holds a phone overhead among an outdoor crowd.
    Mobile filming at an outdoor gathering. Photo from the author’s archive.

    I was watching this transition in 2021. My Russian-language VerticalVideoPortal article about livestream sales and Alibaba was based on contemporary Nikkei Asia reporting. The details are now historical, but the mechanism remains worth studying: content, audience, commerce and logistics converge into a single journey.

    Search commerce and discovery commerce solve different problems

    Search commerce works best when the buyer has a fairly clear intention: find a replacement charger, compare a known brand, order groceries. Discovery commerce operates earlier in the journey. A viewer sees someone using a product and starts imagining how it might fit into their life.

    Neither model eliminates the other. A customer may first encounter an item in a short video, research it elsewhere, and then buy from a familiar merchant. Measuring the impact of content therefore requires more than attributing every purchase to the final click.

    Why livestream selling is more than video with a buy button

    A useful shopping livestream combines four jobs. It attracts attention, explains the product, addresses objections, and creates a path to purchase. The host might demonstrate fabric texture, compare sizes, answer audience questions or reveal how a device behaves in ordinary use. The persuasive element is not merely the link: it is the ability to observe and interact.

    There are also risks. Live demonstrations can create urgency that makes comparison harder. Sellers may exaggerate scarcity or fail to disclose commercial relationships. A responsible creator should make sponsorship clear, distinguish testing from advertising and avoid promising results they cannot substantiate.

    The retail operating system behind the screen

    The show is the visible layer; successful commerce requires accurate product data, pricing, inventory, shipping, returns and customer support behind it. For a retailer, the critical handoff runs from content to product information to fulfillment. A video that sells an unavailable product can generate disappointment faster than a conventional advertisement.

    As an engineer who has worked on enterprise integrations and retail processes, this is the part I find most compelling. A seller’s video can introduce demand in minutes, but the warehouse, item catalog, supply chain and service team still have to deliver. An AI-driven recommendation system does not fix broken operational data.

    A practical experiment for a small business

    1. Choose one product that can be demonstrated honestly in under a minute.
    2. Identify the audience’s real problem and show it before the product features.
    3. Film a transparent demonstration, including limitations.
    4. Link to a clear product page with reliable price and availability.
    5. Track discovery, product-page visits, purchases, returns and questions separately.
    6. Review whether the content created useful customer understanding, not just clicks.

    For a small creator or entrepreneur, this can be run as a learning exercise without copying the intensity of Chinese livestream commerce. Start with evidence that a person understood the product. Only then consider scaling distribution or adding paid promotion.

    What AI changes next

    AI can assist with transcripts, translations, catalog enrichment, product matching and frequently asked questions. It may also help merchants anticipate demand from content signals. But a system that recommends or purchases a product should explain its evidence and respect the customer’s preferences and approval.

    This is the deeper continuity between Douyin’s early commerce story and today’s agentic commerce discussion. Both reduce the distance between discovering something and acting on it. The essential design question is where convenience ends and consent begins.

    Historical sources

    • My 2021 Russian archive article
    • Nikkei Asia — contemporary reporting on ByteDance livestream commerce
    • TikTok’s 2025 description of discovery-led commerce
    November 13, 2026
  • Working With Creators: A Partnership Framework Beyond Follower Counts

    One of the assumptions that made influencer marketing confusing in its early years was that a bigger audience necessarily meant a better partnership. A creator with millions of followers can be the wrong person to explain a specialized product. A modest educator, craftsperson or local filmmaker may be far more credible to the people a brand actually wants to reach.

    People gather around a wooden table beside large windows.
    An informal gathering from the author’s archive.

    My 2020 Russian TikTok portal carried a credited translation of a Hackernoon article about influencer-marketing strategy. That advice belongs to a particular phase of the platform’s development. Here is the framework I would use now as someone who works with technology, video production and business processes.

    Start with a job, not a creator database

    What should change after the partnership? More people might recognize a product. Students might understand a difficult idea. An audience might learn to use a tool safely. Or a community might discover an event worth attending. Those are different jobs, and each calls for a different kind of creator.

    Write the intended outcome before searching for candidates. Otherwise you may end up optimizing a spreadsheet of follower counts without asking whether a collaboration would be useful to anyone.

    Look for alignment across four dimensions

    1. Subject credibility. Does the creator have experience, curiosity or a demonstrable relationship with the topic?
    2. Audience fit. Are the likely viewers the people who could benefit from the story?
    3. Creative language. Does the creator know how to communicate naturally in their chosen format?
    4. Operational reliability. Can they agree on disclosures, deadlines, usage permissions and feedback without losing their voice?

    In practice, review complete videos rather than a profile’s headline numbers. Listen to the language the creator uses. Look at the questions viewers ask, the quality of replies and whether previous commercial posts sound markedly less sincere than regular ones.

    Give a problem, not a script

    The brand should define the product facts that cannot be misstated, the intended audience, legal constraints and measurable goal. The creator should retain meaningful freedom over the narrative, visuals and tone. Over-prescribing the script can destroy precisely the credibility for which the partner was chosen.

    For example, an AI educator might be asked to show a genuine workflow and disclose its limitations—not recite a list of software features. A filmmaker might document a real customer’s experience rather than recreate generic advertising gestures.

    Measure what happened after the view

    Exposure metrics matter, but they should not be confused with outcomes. A sensible measurement plan might record qualified website visits, saved content, meaningful replies, sign-ups or verified purchases. A creator teaching a technical skill might measure whether viewers completed a small practical exercise.

    Ask what data is actually available before promising attribution. Platform-reported engagement, tagged links, surveys and first-party business outcomes measure overlapping but different things. A campaign should not claim causal effects it cannot establish.

    Five questions to agree on before filming

    • Who owns the resulting video and who can reuse it, where and for how long?
    • How will paid relationships and sponsored content be disclosed?
    • What claims require evidence or approval?
    • What happens if the product or service changes between filming and publication?
    • How will the partners discuss results without confusing attention with genuine impact?

    AI is useful, but authenticity is still scarce

    Generative tools can produce drafts, translate subtitles and accelerate revisions. They can also make every campaign sound alike. If an audience chooses to trust a particular person, replacing that person’s perspective with a synthetic generic endorsement defeats the purpose of the collaboration.

    I would rather see one thoughtful, well-disclosed video built around an authentic observation than a hundred near-identical clips chasing the same trending hook. This is especially true when the subject is education, culture, technology or another domain in which credibility accumulates over time.

    For my own work under Aivan Kor — AI • People • Culture, that means collaborating around genuine learning, filming and human stories rather than pretending every creator relationship is a growth hack.

    Sources and provenance

    • My 2020 Russian archive translation
    • Original Hackernoon article
    • US FTC guidance on disclosure of paid endorsements
    November 13, 2026
  • Who Owns Your Attention? TikTok, Recommendation Algorithms and Human Agency

    Have you ever opened a video app for one specific reason and looked up much later, unable to explain where the time went? A recommendation feed makes it effortless to move from one interesting clip to another. The deeper question is what happens to our intentions when an interface continually makes the next choice easier than leaving.

    Man in glasses and a checked blazer sits holding a phone beside a plain wall.

    In 2020 I translated Jia Tolentino’s New Yorker essay on TikTok and attention into Russian for VerticalVideoPortal. It was a long, memorable piece, rich in scenes and reporting. Rather than reproduce it, I want to revisit one of the underlying questions from my perspective as an engineer and creator: how do we benefit from personalized discovery without surrendering the direction of our time?

    Frictionless is not the same as neutral

    The short-video interface is remarkably low-friction. Full-screen content starts without searching; a swipe moves to another selection; feedback arrives almost immediately. These design decisions help people discover unfamiliar art, jokes, languages and knowledge. They also make stopping an active choice.

    Recommendation algorithms can learn from patterns of behavior, but behavior is imperfect evidence of desire. Watching a disturbing clip does not necessarily mean you wanted to see more of the same. Replaying an explanation may mean confusion rather than fascination. A system can accurately predict a next action while misunderstanding what a user considers worthwhile.

    The metrics do not know your purpose

    Imagine a student opening a feed to find one explanation of an electric field. The app may show a relevant demonstration followed by unrelated entertainment. If the student stays, retention may rise. But did the student achieve the learning goal? From the user’s perspective, there is an important difference.

    I work with AI and education, and I believe this distinction should shape our metrics. Watching time is one signal. Understanding, remembering, practicing and applying a concept are different outcomes. A product that measures only attention can optimize the wrong thing very effectively.

    The creator faces the same tension

    Creators are constantly encouraged to increase retention, produce faster hooks and avoid any moment that might make a viewer leave. Some of that is good storytelling practice: make the subject clear, remove unnecessary filler and respect the audience’s time.

    But when every creative decision is governed by a prediction of the next swipe, a video can lose its point of view. Complexity disappears. Repetition wins. A creator may gradually learn to perform whatever works instead of developing what they actually have to say.

    I think of this as a question of authorship. You can use data to make your communication clearer without allowing data to dictate the beliefs, relationships or stories you are willing to put into the world.

    A practical attention contract

    1. State the intention. Before opening an app, know whether you want to learn, relax, research or connect.
    2. Use explicit feedback. Follow useful creators, indicate disinterest where supported, and curate what you can.
    3. Separate discovery from deep work. Save a useful idea and return to it in a calmer format rather than endlessly browsing adjacent clips.
    4. Notice the exit. If the next video no longer serves the original intention, leaving is a successful decision.
    5. Evaluate the result. Ask what you gained, made or understood—not only how much time passed.

    What this means for AI agents

    Recommendation systems are beginning to connect with assistants that can search, schedule and take actions. That raises the stakes. A feed choosing the next video can cost attention; an agent choosing a purchase or sending a message can create financial or social consequences.

    Human agency therefore needs to be designed into the product. We should be able to inspect reasons, grant limited permission, approve consequential actions and change our minds. Automation should not make the user’s purpose harder to see.

    I do not oppose recommendations. I love how a thoughtful system can introduce us to a place, an artist or a scientific idea we would never have found on our own. But discovery should increase the range of choices available to us, rather than quietly decide what we are for.

    Why I care as a creator

    My English project is called AI • People • Culture. It includes technology, education and storytelling because these subjects belong together. A useful algorithm is an engineering achievement; the lives and purposes of people affected by it are equally part of the story.

    The question I now ask of any engagement-optimized product is simple: when the session ends, does the person feel more capable of living deliberately—or merely more likely to come back?

    Reading

    • Jia Tolentino, How TikTok Holds Our Attention (2019)
    • My credited Russian translation (2020)
    • TikTok’s explanation of content recommendations
    November 13, 2026
  • Can You Engineer a Viral Song? AI Music, Short Video and Cultural Discovery

    A melody can reach a listener through a film, a friend, a radio station or a recommendation feed. Short-video platforms changed the role of the song itself: a few seconds of music may become a repeated sound, an editing structure, a joke or a social signal long before a listener explores the complete composition.

    Man in a green-and-white shirt makes a hand gesture beside neon music notes.

    I was reading about this change in 2021, when my Russian TikTok portal published a credited translation of a TechCrunch commentary on songs designed for virality, itself reacting to Bloomberg reporting about TikTok music strategy. Today, as someone who experiments with multilingual AI-generated music, I find the same question much more personal: can we design a song for discovery without reducing music to an algorithmic trick?

    The short clip is a new kind of musical interface

    In conventional listening, a track is often encountered as a complete object: a verse, chorus, arrangement and ending. On short-video platforms, a fragment may have a separate life. A beat drop cues an edit; a lyric becomes a caption; a recognizable phrase invites an imitation.

    That can be creatively liberating. It lets people build visual jokes and stories around a sound. It can also exert pressure on artists to prioritize a fragment that travels well over a composition that rewards sustained listening. Neither outcome is inevitable.

    Discovery is not the same as musical success

    A million short-video plays do not automatically translate into a million full-track listeners, paying supporters or meaningful cultural connections. A sound can become famous while the musician remains unknown. It can also introduce a previously obscure artist to new audiences.

    Music analytics screenshot showing 716,883 streams and a line chart with a sharp spike near the end.
    An analytics snapshot discussed in the Sattu Power case study. The dashboard total and the individual track count appear separately.

    For a creator, measurement should reflect the intended outcome. If the goal is musical appreciation, examine full-song listens, saves, thoughtful comments or concert attendance. If the goal is an experimental music video, examine whether the song and visual story actually work together. Viral counts are only part of the picture.

    What changes when AI makes the music?

    AI tools can help with instrumentation, production, lyrics, voice experiments and translation. I have used them to explore musical ideas across languages, including Hindi. My interest is not only speed. I am curious whether these tools let someone outside a culture approach it more respectfully and creatively—or whether they make it easier to produce an empty imitation.

    Translation is a good example. A phrase that scans well in one language may sound unnatural in another. Meter, rhyme, pronunciation, cultural associations and humor are difficult to preserve simultaneously. An automated draft needs review by people who actually know the language and context.

    A more useful process than “make it viral”

    1. Begin with an emotional or narrative idea that would matter even without an algorithm.
    2. Decide who the song is for and what cultural knowledge you need to research.
    3. Experiment with short musical motifs, but develop a coherent complete track.
    4. Create a visual story that adds meaning rather than merely repeating the lyrics.
    5. Test pronunciation, context and translation with appropriate human reviewers.
    6. Publish with clear credits and follow the licensing rules of the tools and platforms.
    7. Evaluate the listeners and conversations the work attracts, not just the view count.

    Why manufactured engagement is an ethical problem

    There is a difference between making art accessible and deliberately manipulating users. Creators have always studied hooks, choruses and memorable rhythms. What feels new is the granularity of feedback and the ability to produce and distribute variations at scale.

    I think the ethical line becomes clearer when we ask what the system optimizes. Is it helping a meaningful song find the people who might love it? Or is it generating endless stimuli to occupy attention for its own sake? The tools may look similar; the intention and downstream effects are not.

    Culture is not a dataset to be exhausted

    The most interesting AI music projects are those that lead to real curiosity: Who wrote this style of rhythm? What do the lyrics mean? Which artists shaped this tradition? Who should be credited or compensated? A creative experiment should open a door to people, not replace them with a synthetic collage.

    That is part of the mission behind my multilingual AI music projects and my English-language channel, Aivan Kor — AI • People • Culture. The technology interests me, but the encounters it can enable matter more.

    A successful song need not be engineered to dominate the feed. Sometimes success is a listener from another country discovering a new word, hearing a new story or recognizing a human feeling across a language barrier.

    Sources and archive

    • My 2021 Russian archive translation
    • TechCrunch commentary (2021)
    • Bloomberg reporting (2021)
    • Aivan Karade and Sattu Power: personal AI music case study
    November 13, 2026
  • Before AI Went Mainstream, I Built a TikTok Research Portal. Here’s What I Learned.

    In 2020, I started a Russian-language website about TikTok: VerticalVideoPortal.ru. I was a software engineer, but I was fascinated by a consumer product that seemed to understand what people wanted to watch before they knew how to describe it. Rather than write another stream of tips about hashtags, I wanted to understand the technology, business model and culture behind short video.

    Man in a red hoodie holds a phone overhead among an outdoor crowd.
    A mobile filming moment from the author’s archive.
    Innopolis event announcement with two speaker portraits, including Ivan Koroteev, and a September 28 start time.
    An event announcement featuring Ivan Koroteev and his talk on business promotion.

    The editorial strategy was simple: find substantial English-language analysis, translate it into Russian, credit the original authors and make those ideas available to readers who might not otherwise encounter them. The portal eventually accumulated 362 published articles, mostly from 2020 to 2022. Looking back in 2026, I see it as an early research project into a much bigger subject: how machines mediate human attention, creativity and commerce.

    1. I was interested in the system, not just the app

    A dance clip or a meme is easy to notice. The system that finds its audience is less visible. Traditional social networks were organized around people you chose to follow. TikTok demonstrated a different approach: an algorithmic feed could learn from interactions and repeatedly introduce viewers to unfamiliar creators.

    Hashtag page for Avatarify with example tiles of a person, two children and a cat.

    At the time, I translated research such as our archive of TikTok analysis, including a piece on how AI changes platform design and economics. The latter was a Russian translation of Sangeet Paul Choudary’s original analysis. Its questions outlived the news cycle: who decides what gets distributed, what signals count, and who captures the value created by participants?

    TikTok’s own explanation of its For You recommendations describes signals such as user interactions, video information and account settings. The precise system is proprietary and changes over time; my interest is in the underlying shift from a social graph to interest-led discovery.

    2. Translation was also a way of learning

    Translating serious technology reporting forces you to slow down. A vague sentence becomes a problem to solve: what exactly did this author mean by recommendation, platform economics, retention or creator incentives? The vocabulary matters because the mechanisms matter.

    The portal covered different kinds of sources: journalism, research essays, company announcements and case studies. I learned to distinguish a marketing claim from a useful business model, and a temporary product feature from a change in how people behave. Not every article in that archive deserves republication today. Much of the news is now history. But the habit of reading across languages and checking first-hand sources remains part of how I work.

    3. The creator is not merely a content supplier

    In an essay I wrote for the portal, “TikTok is creativity and self-expression”, I argued that short video gives ordinary people a way to show what matters to them—music, jokes, teaching, daily work or a beloved cat. That thought still resonates with me.

    I am relatively introverted, yet I became a public speaker and video creator. For me, the camera can be a creative instrument rather than a demand to behave like an extrovert. It lets you prepare, compose, experiment, edit and say something worth sharing. You do not need to turn your personality into a permanent performance.

    This is also why my newer work spans artificial intelligence, filmmaking and culture. New tools lower the cost of making things. They do not automatically give us something meaningful to say.

    4. China was an essential part of the story

    Through the portal, I followed ByteDance, Douyin, Chinese mobile products and the international expansion of TikTok. I was drawn to the interaction between engineering, entrepreneurship and culture: a product born in one technological ecosystem could become a creative language used across many others.

    The archive includes a four-part Russian translation of reporting about Zhang Yiming and TikTok’s globalization, originally published through Ginger River. Those stories are useful as historical sources, not as a substitute for verifying what the companies are doing now.

    That distinction matters. I don’t want to build an English blog that merely reproduces 2021 predictions in 2026. I want to use the archive as a starting point for fresh questions and original analysis.

    5. Three lessons I carried into my AI work

    • Distribution is part of the product. A beautifully built tool does little good if it cannot find the people for whom it solves a problem. Algorithms shape that match.
    • Metrics are not the same as meaning. Views can indicate attention, but they cannot by themselves tell us whether a video taught someone something, built trust or inspired action.
    • Technology must meet human agency. An intelligent system can recommend the next video, workflow or action. People still need the judgment to decide which recommendations serve their goals.

    These lessons influence how I think about enterprise AI as well. A successful implementation isn’t simply a clever model. It requires a real task, useful information, a way to evaluate outcomes and a human who remains responsible for the decision.

    6. Why I am bringing this history into English

    I now publish under Aivan Kor — AI • People • Culture. My English-language work combines engineering, practical AI applications, videos and conversations across cultures. The TikTok portal belongs in that story because it documents a long-standing curiosity: how a technical system becomes something that changes everyday life.

    I plan to revisit the strongest articles from the archive—not by reverse-translating the original English authors, but by developing new essays with my own analysis, current evidence and clear links to earlier research. I also want to convert the best ideas into videos on my YouTube channel.

    The technology keeps changing. My central question hasn’t: how can machines expand what people are able to create, understand and do—without taking away their authorship?

    Archive and reading

    • VerticalVideoPortal.ru — my historical Russian-language TikTok portal
    • TikTok is creativity and self-expression — my original 2020 essay (Russian)
    • Sangeet Paul Choudary: Understanding TikTok — an original English analysis I translated for the portal
    • TikTok: How recommendations work — platform explanation
    • Do videos! — an earlier English note about creating as an introvert

    Editor’s note (2026): This is a new reflection based on my experience building the portal. Linked translated articles are credited to their original authors; this essay is not a republication of those works.

    November 13, 2026
1 2 3 … 6
Next Page→

Creativity For Introverts

You can write me on my business email
mail@creativityforintroverts.com  

Proudly powered by WordPress