Sasha Shlemov, with Drinkins, personal AI assistant · Slides and materials
Meridian Learning Foundation asked me to write something about the talk. I made 26 slides, so let us try the short version.
On 25 September, I gave a talk called Thinking with AI to a room of young students.
I did not try to persuade them that AI is good or bad. We opened the hood: next-token prediction, instructions, control tokens, context, tools, and the boundary between the model that talks and the ordinary software that acts.
Then we used AI to investigate AI. The students played yes-and-no with the question “Does AI help students learn?”, built the strongest cases on both sides, and looked for an observation that could distinguish them.
The same chat interface can answer, teach, challenge, fact-check, brainstorm, or merely validate. If you do not choose what you actually want, the model will choose a plausible default for you.
The slides and two reusable procedures from the talk are here:
Thinking with AI: slides and take-home materials
Meridian deserves a separate note. It is a genuinely inclusive and safe space. Everyone, including me, was able to speak freely and openly, ask difficult questions, and disagree.
Personal thanks to Oksana Stanevich for inviting me and making the talk happen, and to everyone in the room for the discussion.
Meridian: official site · LinkedIn