Subvolts take Poland builds its own AI, proving local control and language adaptation beat global giants. Dec 6, 2025 - Developers
- Learn about training efficient, language-specific models and technical strategies for AI adaptation.
- Business
- Understand how national AI initiatives can create cost advantages and ensure data privacy for local markets.
This episode offers a practical look at national AI development strategies and the challenges of building independent AI capabilities outside major labs.
Sovereign AI in Poland: Language Adaptation, Local Control & Cost Advantages with Marek Kozlowski
Audio streams from Erik Torenberg, Nathan Labenz. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Marek Kozlowski discusses Poland's PLLuM project on The Cognitive Revolution podcast. He explains how training smaller, local AI models can achieve AI sovereignty. This approach addresses English bias in frontier models, ensures data privacy, and offers cost advantages, aligning with EU regulations and promoting national technological control.
Imagine training a smart assistant that only speaks Polish and understands Polish culture. This episode explains how countries like Poland can build their own AI, rather than relying on big foreign companies. This keeps data private and makes AI more useful for local needs.
This episode is for AI developers and policymakers interested in national AI strategies. Kozlowski details practical steps for "language adaptation" of base models, offering a counterpoint to the massive scale of frontier models. It connects to current discussions around AI governance, regional AI development, and the challenges of global AI dominance, showing how countries can build independent AI capabilities.
- HostNathan Labenz
- GuestMarek KozlowskiHead of the AI Lab, Poland's National Information Processing Institute
- Episode Title: Sovereign AI in Poland: Language Adaptation, Local Control & Cost Advantages with Marek Kozlowski
- Show: The Cognitive Revolution
- Guest: Marek Kozlowski, Head of the AI Lab at Poland's National Information Processing Institute
- Project Discussed: PLLuM (Polish Large Language Models)
- Key Strategy: Training small, locally-adapted models for specific languages and cultures
Marek Kozlowski, Head of the AI Lab at Poland's National Information Processing Institute, discusses project PLLuM (Polish Large Language Models). PSA for AI builders: Interested in alignment, governance, or AI safety? Learn more about the MATS Summer 2026 Fellowship and submit your name to be notified when applications open: https://matsprogram.org/s26-tcr. He shares how countries like Poland can achieve AI sovereignty by training small, locally-adapted models for specific languages and cultures, ensuring control, privacy, and cost advantages. The conversation delves into challenges like frontier models' English bias, EU regulations, and technical strategies like "Language Adaptation" on base models. Discover how transparently created, locally-controlled AI offers a viable path for nations to maintain their technological destiny.
LINKS:
Sponsors:
Google AI Studio:
Google AI Studio features a revamped coding experience to turn your ideas into reality faster than ever. Describe your app and Gemini will automatically wire up the right models and APIs for you at https://ai.studio/build
Agents of Scale:
Agents of Scale is a podcast from Zapier CEO Wade Foster, featuring conversations with C-suite leaders who are leading AI transformation. Subscribe to the show wherever you get your podcasts
Framer:
Framer is the all-in-one platform that unifies design, content management, and publishing on a single canvas, now enhanced with powerful AI features. Start creating for free and get a free month of Framer Pro with code COGNITIVE at https://framer.com/design
Tasklet:
Tasklet is an AI agent that automates your work 24/7; just describe what you want in plain English and it gets the job done. Try it for free and use code COGREV for 50% off your first month at https://tasklet.ai
Shopify:
Shopify powers millions of businesses worldwide, handling 10% of U.S. e-commerce. With hundreds of templates, AI tools for product descriptions, and seamless marketing campaign creation, it's like having a design studio and marketing team in one. Start your $1/month trial today at https://shopify.com/cognitive
PRODUCED BY:
CHAPTERS:
(00:00) Sponsor: Google AI Studio
(00:31) About the Episode
(03:17) Sovereign AI in Poland
(04:41) The Case for Localization
(13:38) The PLUME Project's Mission (Part 1)
(20:25) Sponsors: Agents of Scale | Framer
(22:47) The PLUME Project's Mission (Part 2)
(22:47) Defining Polish AI Values
(35:32) Sourcing and Curating Data (Part 1)
(35:38) Sponsors: Tasklet | Shopify
(38:46) Sourcing and Curating Data (Part 2)
(44:40) Small Models, Big Advantage
(58:21) Training and Domain Ad
- Sovereign AI
- Marek Kozlowski
- AI that a country controls itself, not relying on foreign companies.
- Language Adaptation
- Marek Kozlowski
- Adjusting a general AI model to better understand and use a specific language.
- Frontier models
- Marek Kozlowski
- The most advanced, large-scale AI models developed by leading tech companies.
1 h 27AI:AM: Where's the Steering Wheel? My Agent Set the Password, Hidden Valence, Proofs Find Linux Bugs
WithCameron Berg, Everett Hildenbrandt, Grigore Rosu, Phil Maher
2 h 192 h 021 h 28AI:AM: A Level We Shouldn't Pass? Notes from The Curve + Tokens vs. Salaries & Is SaaS Cooked?
WithThomas Sohmers, swyx, Evan Miyazono, Edward Hu
Questions people ask
- What is the PLLuM project?
- PLLuM stands for Polish Large Language Models. It is a project focused on developing AI models tailored for Poland's specific language and cultural needs.
- How can countries achieve AI sovereignty?
- Countries can achieve AI sovereignty by training smaller, locally adapted AI models. This ensures control, privacy, and cost advantages, and addresses issues like the English bias found in many frontier models.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Erik Torenberg, Nathan Labenz.