Warsaw AI Safety Day 2026

September 8th. 8 pm. 7 days after i have returned home from last holiday trip, I have seen the most shocking news this year: OpenAI has posted the proof for the Navier-Stokes Conjecture. My insides were twisting as thought someone made them into spaghetti, ate the, and puked. I couldn't sleep. All i saw was the ever-ending abyss that is AI. My romanticized vision of a mathematical life filled with struggle over paper and pencil has shattered. My own intuition for life has dissapeared. I had only one thought: "We are in the End-times".

A new priority has appeared in front of my eyes - AI Safety. It was astoundding how convinced I became in a single second, that this foreign space to me requieres my intervention and attention. Over the following weeks, I have spend countless hours listening to podcasts, lectures, reportages, and tutorials on AI. But, it was the next day that I applied for a non-existant internship in the two-day-old Mathematical AI Safety Institute (MAISI) launched by a Fields-medal-laurate Jacob Tsimerman, applied to join AI Safety Polska, and in the following days decided to participate in Warsaw AI Safety Day 2026 - The first polish conference dedicated to solely this topic.

Looking in the mirror I do not understand myself sometimes - how can the same person, that cannot seem to decide on a single carrer choice and would rather, with all their heart, rot in bed, can be as spontanious in the matter as serious as that.

How serious is it thought?

Lęknąłem się, eż nic po mnie.
Ta mi rzecz barzo niemiła,
Iżeś mię tako postraszyła;
By była co przykrego przemowiła,
Zerwałaby się we mnie każda żyła;
Nagle by mię umorzyła
I duszę by wypdziła.
Proszę ciebie, ostęp mało,
Boć nie wiem, coć mi się zstało:
Mgleję wszytek i bladzieję,
Straciłem zdrowie i nadzieję.
Racz rzucić od siebie kosę,
Ać swoję głowę podniosę.

The Conference itself was very nice - the opening was great and big, around 300 people came. First keynote by Rishub Jain was engaging and extremely inviting. As a volonteer, the fist half of the conference till noon I was watching over the Room B which offered talks centered around AI Governence, out of three talks the stand out for me was Two kinds of "Loss of control": Frontier Risk and European Sovereignty given by Chloé Touzet. My feelings were a conbitation of exitement from new clear viewpoints presented and of the familiar parisian breeze of social justice. I really encourage anyone two follow her work and policy proposals. my Lunch break was preceeded by the premiere of the AI Safety Polska Film that is going to be presented tomorrow on their youtube page. In the afternoon i participated as a casual listener in introductions to Alignement, AIxBIO, Data Security, AI Verification.

Thank God, there are still people, that care about our existance with passion.

Thanks to the speekers, orginizers, my co-volonteers and everyone who participated with me in this conference!