Media Goal

Goal: To help aspiring and working level designers apply neural networks at specific stages of level creation: concept, blockout, content population, and testing.

Six-Month Goals: 1,000 Telegram subscribers (about 170 new per month) 3 regular content columns 2 publications per week

Themes


Segment: Game Development and Professional Education Narrow Topic: AI in Level Designer Work

Why this topic is relevant:

In academic literature, the application of machine learning for level generation is termed PCGML (Procedural Content Generation via Machine Learning). Authors of a 2025 study (arXiv: 2508.09860) note that existing systems often fail to account for human design principles, which limits their practical utility in real-world production processes.

Tools we cover in our sections (based on gamedesignskills.com reviews):Promethean AI: creating game environments and procedural levels Unreal Engine PCG: automatic environment population Unity ML-Agents: training AI agents for level testing

Competitors and Analysis

Критерии сравнения: площадки, подписчики, охват поста, вовлечённость, наличие ИИ-контента, формат.

0
1/2

Indirect Competitors: «Game Design, ” „GameDev: Game Development“ (Tproger). Potential Competitors: AI tool and engine developers (Promethean AI, Unity, Epic Games) who may release educational content.

Posts about AI (manual channel search, queries for „AI, ” „neural network“): „Level Design Talk“: 0; „What a Level!“: 3; „Game Design“: 0.

Original size 1672x941

In established channels, there is no regular feature dedicated to AI in level design. Only one channel has been found with three posts on this topic. Direct competitors are small (3,883 and 6,053 subscribers). The indirect competitor has the most subscribers (11,200). The channel with the smaller audience receives more views per post: 1,207 with 3,883 subscribers versus 627 with 6,053. This suggests that post quality is more important than channel size. «Vot eto urovel!» gained 320 subscribers in a month, while the others gained less than 20. It is the only one of the three showing noticeable growth. What we will adopt: the combination of the channel, chat, and VKontakte used by «Vot eto urovel!». What differentiates us: making AI the primary, rather than episodic, theme, and creating repeatable features.

Conclusions:

  1. There is no regular segment about AI in level design in the established channels. Only one channel has found three posts.
  2. Direct competitors are small (3,883 and 6,053 subscribers). The indirect competitor has the most subscribers (11,200).
  3. The channel with the smaller audience gets more views per post: 1,207 with 3,883 subscribers versus 627 with 6,053. This means post quality is more important than channel size.
  4. «Vot Eto Uroven!» gained 320 subscribers in a month, while the others gained fewer than 20. It is the only one growing noticeably.
  5. What we will borrow: the combination of a channel, a chat, and VKontakte that «Vot Eto Uroven!» uses.
  6. How we will differentiate: making AI the main, not episodic, topic, and implementing recurring segments.

Distribution Channel

Primary: Telegram. This is where all analyzed competitors operate, and Telegram hosts professional discussions and community chats for this topic.

Secondary: VKontakte (following the example of «This is the level!» which has a VK community) and DTF (Gamedev section) for in-depth analysis.

Logic: A concise format on Telegram, detailed analysis on DTF. We adapt the content to the platform rather than duplicating it.

Target Audience.

Original size 1672x941

Model personas describe typical representatives of these segments.

Context: Reading professional Telegram channels and chats, watching breakdowns on DTF.

Value for everyone: specific techniques that can be applied the very next day.

Unique Selling Proposition

  1. Focus only on level design. Not general game development or general AI, but the specific stages of working on a level.

  2. Practice over hype. Every piece of material includes a repeatable step: a prompt, a configuration setting, or a checklist.

  3. Honest discussion of limitations. We show what AI does poorly and analyze unsuccessful experiments.

ИИ для левел-дизайнера
Project created at 04.10.2026
We use cookies to improve the operation of the website and to enhance its usability. More detailed information on the use of cookies can be fo...
Show more