Warsaw Software Architecture Group
Past meetup
25.03.2026
March 25, 2026, 17:30–21:00

WSAG Meetup #3

Join us for the third Warsaw Software Architecture Group meetup! This time we'll discuss AI ethics and how to learn with LLMs without incurring cognitive debt.

Polsko-Japońska Akademia Technik Komputerowych, Koszykowa 86

Polsko-Japońska Akademia Technik Komputerowych

About This Event

We're thrilled to host the third WSAG meetup at the Polish-Japanese Academy of Information Technology! This event will feature two presentations: AI ethics and responsibility beyond local optima, and the art of critical thinking in the era of LLMs.

This is a great opportunity to meet fellow software architects, share knowledge, and continue building our local architecture community.

Agenda

  1. 17:30

    Welcome & Networking

    Registration and networking

  2. 18:00
    Tomasz Zgliczyński-Cuber

    Where Smart Algorithms Get Stupid - On Ethics in AI

    Tomasz Zgliczyński-Cuber

    in Polish30 min
    Responsibility means going beyond the local optimum. AI excels at local optimization, so if you don't want to be replaced, here's a simple solution: AI won't replace those who look broader and think about what happens beyond the horizon.
  3. 19:05
    Tomasz Ducin

    The Forgotten Art of Thinking. How to Learn with LLMs Without Increasing Cognitive Debt?

    Tomasz Ducin

    in Polish30 min
    There is research showing that using LLMs leads to incurring "cognitive debt" - people intellectually take shortcuts. An LLM generates a wall of text, the operator may not even read it - and a false impression is created that "the job is done." This applies to research (not verifying sources - e.g., an LLM citing non-existent studies - ouch, politicians!), coding ("I don't read what I push to production") - but also - and this is our focus - learning.

    What if LLMs are a tool that simultaneously elevates to great heights - and terrifyingly dumbs down (and everything in between)? LLMs are "just a tool" - great, great. But how do you use it in practice to maximize those "heights" and minimize the dumbing down?

    When generating text/code becomes cheap, understanding and decision-making become relatively more expensive.

    We'll reach for the fundamentals of cognitive science and dust off the foundations of "critical thinking," then apply "learning with LLMs" in practice - using diverse techniques/methods/tools to build your own understanding (and mental models) about a given topic, verification methods, methods of "going deeper" (not just "there are always many levels deep" - but also - where to stop) - and many other practical tips.
  4. 20:00

    Closing & Networking

    Final Q&A and networking

Location