S8 Linguistic Bias Test: Eastern Europe Region Results

## S8 Linguistic Bias Test: Eastern Europe Region Results **Search date:** 2026-11-09 **Query:** генеративное искусство процесс блог персональный сайт нейросеть GAN код **Objective:** Test S8 signal in Russian/Eastern European context. ### Key Finding: Strong Code-Art Tradition, But Institutional Pathways Exist Eastern Europe (particularly Russia) has a strong generative and code-art tradition, but many practitioners are already embedded in academic or conference circuits. The search found fewer "pure" pre-institutional candidates than expected. ### Candidates Found #### Ivan Dianov (ivandianov.com) — DISQUALIFIED - **Medium:** Data art, live-coding, shaders, generative systems - **Documentation:** Personal essay describing path through cellular automata, L-systems, physics, fractals, shaders, live-coding - **Technical depth:** High — explicitly mentions inability to work with neural networks due to time constraints of training on own datasets - **Institutional validation:** - Teaches data art at НИУ ВШЭ (Higher School of Economics) - Founder of educational platform "Сетка" - Performs at HolyJS conference and live-coding performances in Krasnoholmsk gallery - **S8 assessment:** **DISQUALIFIED** — academic teaching position, conference performances, and gallery shows constitute curated channel validation - **Source:** https://ivandianov.com/ru/essay/, https://ivandianov.com/ru/about/ #### Sergey Egorov / hayabuzo (github.com/hayabuzo) — PROVISIONAL S8 CANDIDATE - **Medium:** Generative photography - **Documentation:** GitHub account since 2018-04-10, blog at hayabuzo.me (content not fetched) - **Technical depth:** Unknown — GitHub shows generative photography work but specific process documentation not verified - **Sustained practice:** GitHub account since 2018 suggests long-term engagement, but generative photography specifics unclear - **Institutional mentions:** None found in excerpt - **S8 assessment:** **Provisional S8=1–2** — long-term digital art practice suggested by GitHub longevity, but process documentation depth unverified. Blog content needs direct fetch. - **Source:** https://github.com/hayabuzo #### Alexey Lyanguzov (lyanguzov.ru) — DISQUALIFIED - **Medium:** GAN-based art (2022 project) - **Documentation:** Personal site describing 2022 project training GAN on participant doodles, transferring to acrylic on canvas - **Technical depth:** Medium — describes GAN training concept but no code/logs shown - **Institutional validation:** - Passed competitive selection for ИТМО (ITMO University) project in St. Petersburg (top-100 AI university globally) - Exhibition in St. Petersburg planned but cancelled in 2022 - **S8 assessment:** **DISQUALIFIED** — competitive university selection and planned exhibition constitute institutional pathway - **Source:** https://lyanguzov.ru/ #### Stas Sumarokov (ru.exsstas.art) — PROVISIONAL S8 CANDIDATE - **Medium:** Real-time visualizations, generative systems, TouchDesigner, shaders, interactive instruments - **Documentation:** Personal site in Russian, English, and Spanish. Describes creating real-time visualizations and generative systems with code. - **Technical depth:** High — TouchDesigner tutorials, GLSL code, algorithmic research - **Sustained practice:** Listed as 2025 on site, but span unclear - **Location:** Valencia, Spain (Russian origin) - **Institutional mentions:** None found in excerpt - **S8 assessment:** **Provisional S8=1–2** — strong technical focus and code-based practice, but institutional absence not verified and span unclear - **Source:** https://ru.exsstas.art/ ### Non-Candidates Mentioned in Search | Artist | Nationality | Status | Relevance | |--------|-------------|--------|-----------| | Hannu Töyrylä | Finnish | Already known (S8=3) | Documents GAN work since 2015 in English | | Robbie Barrat | American | Already known | Archive 2016-2018, but "in 2026 it found a home" phrase needs verification | | Wavefunk (Hans Brouwer) | Unknown | Already known | Dated journal 2018-2021 on GAN experiments | ### Critical Pattern: Academic/Conference Pathway in Eastern Europe **Eastern European generative artists show stronger academic and conference integration:** - Ivan Dianov teaches at university and performs at tech conferences - Alexey Lyanguzov went through competitive ITMO selection - This contrasts with Japan (personal sites, code-based) and Korea (festival/government exhibitions) **Implication:** S8 may be even rarer in Eastern Europe because the academic and conference infrastructure finds and validates code-based artists faster than gallery/museum systems do in other regions. ### Regional Comparison Summary (All Four Regions) | Region | Strongest S8 Candidate | Disqualification Rate | Key Pattern | |--------|----------------------|----------------------|-------------| | Japan | ryomak.jp (S8=2) | 1/4 (blender233a) | Code-based tradition, strong dating, low institutional connectivity | | Korea | YINASTAR (disqualified) | 2/4 (Smol, Elissa) | High institutional connectivity, government/festival support | | Latin America | Dario Martins (S8=1) | 2/3 (Uno de Oliveira, Mathias Lobo) | Contests and NFT collections as validation, less technical depth | | Eastern Europe | Stas Sumarokov (S8=1-2) | 2/4 (Dianov, Lyanguzov) | Academic/conference pathways, strong code-art tradition | ### Linguistic Bias Assessment: CONCLUSION **Finding:** S8 signal DOES generalize to non-English contexts, but its distribution and characteristics vary significantly by region: 1. **Signal exists:** Found process-documenting generative artists in all four non-English regions 2. **Signal is rarer than expected:** High disqualification rates across all regions (50-67%) suggest either S8 is transient or my search methods surface already-validated artists 3. **Regional variation is real:** - **Japan:** Code-based, systematic dating, lowest institutional connectivity - **Korea:** Festival/government exhibitions find artists fastest - **Latin America:** Contests and NFTs function as early validation - **Eastern Europe:** Academic and conference circuits validate code artists 4. **Documentation culture varies:** - Japanese code artists date works explicitly (programming culture) - Korean artists document on platforms (Brunch, Blogspot) - Latin American documentation is less technical, more conceptual - Eastern European artists have strongest code-art tradition but also strongest academic integration **Verdict:** S8 is not Anglophone-dependent, but it IS regionally variable. The framework needs regional calibration or, better, should be understood as detecting a transient phase rather than a stable population. ### Open Questions - Does Sergey Egorov's blog (hayabuzo.me) contain sustained process documentation? - Does Stas Sumarokov have any institutional presence in Spain or Russia? - Is the high disqualification rate evidence that S8 is transient, or evidence that systematic process documentation correlates with institutional ambition? - Should the framework abandon "pre-institutional" as a category and instead track "early-stage" with timestamped trajectory tracking?