Macedonia has announced it is getting its first national center for artificial intelligence - with a name that wasn't chosen by chance: „Vezilka" (the Embroiderer). Digital Transformation Minister Stefan Andonovski said the goal is to build a Macedonian language model and a homegrown digital base, rather than depending entirely on foreign tools that poorly recognize our language and context.

The idea has a logic that's hard to dispute. Today's large language models treat Macedonian as a small, secondary language - they make case errors, mix up meanings, don't grasp the cultural context. The center, according to the announcements, will work on digitizing the Macedonian language corpus and developing a model intended for institutions, education, science, and citizens. The partners are MANU (the Academy of Sciences and Arts), the National and University Library, the Parliament, the Official Gazette, and the institutes for the Macedonian language.

The figures are concrete: the project costs around 6.5 million euros - 1.5 million from the state, 1 million from the faculties, and the rest from the European Union. A functional language model is expected by the end of 2026 or the beginning of 2027, with the intention of offering it as open-source software for domestic institutions and companies.

It sounds good on paper - and here it's precisely on paper that projects sound best. The question isn't whether the idea is good (it is), but whether a real tool will be built that institutions use every day, or another ceremonially opened structure that no one will mention in two years. The vezilka of folk costume embroiders with patience and with a thread that lasts decades. The same will be asked of this digital Vezilka - not just to open, but to endure.