Recordings · Spring 2026

Talks from the first seminar

Recordings during the AFFINE Superintelligence Alignment Seminar 2026, posted on YouTube. Slides are linked where the speaker shared them.

Week 1

  • Writing: Getting Started

    Jonas Hallgren · 30 April 2026 · Recording

  • Effective Disagreements 2: Acknowledging Biography

    Camille Berger · 30 April 2026 · Recording · Slides

  • Effective Disagreements 1: Socratic Questioning

    Camille Berger · 1 May 2026 · online only · Recording · Slides

  • Variants of the Alignment Problem

    Kaarel Hänni · 1 May 2026 · Recording · Slides

Week 2

  • Agent Foundations: How and Why

    Noa “Ouro” Hölzer · 4 May 2026 · Recording · Slides

  • Inspection of Thinking as a Source of Ideas for Understanding Minds

    Kaarel Hänni · 4 May 2026 · Recording · Slides

  • Mental Mislabelling and How to Fix It

    Ryan Thomas · 4 May 2026 · Recording · Slides

  • The Many Faces of Goodhart’s Law

    Abram Demski · 4 May 2026 · Recording · Slides

  • Understanding Understanding (including Peer Tuition)

    plex · 5 May 2026 · Recording · Slides

  • The Trouble with Maximizers

    Noa “Ouro” Hölzer · 7 May 2026 · Recording · Slides

  • Jonas’ Research Agenda

    Jonas Hallgren · 7 May 2026 · Recording · Slides

  • Superintelligence Alignment: Principles, Attempts, Contributions

    plex · 7 May 2026 · Slides

  • Panel: How and Why to Not Burn Out

    Noa “Ouro” Hölzer, Pauliina Laine, plex, Julia Persson · 8 May 2026 · live-streamed only

  • Decision Theory for AI Safety

    Abram Demski · 8 May 2026 · Recording · Slides

  • Intro to Infra-Bayesianism

    Ben Auer · 9 May 2026 · fellow unconference · Recording · Slides

  • Solomonoff Induction and Inductive Biases of Neural Networks

    Lucius Bushnaq · 9 May 2026 · Slides

Week 3

  • AFFINE Tech Tree

    plex · 11 May 2026 · online only · Recording

  • Verification-Based Alignment Schemes

    Kaarel Hänni · 12 May 2026 · Recording · Slides

  • Panel: The Point of AFFINE

    Mateusz Bagiński, plex, Jonas Hallgren, Noa “Ouro” Hölzer, Julia Persson · 12 May 2026 · Recording

  • Overview of the Learning-Theoretic Agenda

    Vanessa Kosoy · 13 May 2026 · Recording

  • What Is the Alignment Problem?

    Mateusz Bagiński · 14 May 2026 · Recording

  • Deliberate Research Practice: Research Insights from Building Superhuman Poker AI

    Vojta Kovařík · 15 May 2026

  • Catastrophic Goodhart as a Blind Spot for Computer Science

    Vojta Kovařík · 15 May 2026

Week 4

  • Towards a Theory of Identity

    Richard Ngo · 18 May 2026

  • The Lethal Reality Hypothesis

    Ihor Kendiukhov · 18 May 2026

  • Biology and Habits of Open-Ended Research

    Jonas Hallgren · 19 May 2026

  • Safety by Design via Formally Enforcing Care for Human Empowerment?

    Jobst Heitzig · 19 May 2026 · Slides

  • On Model-Wise Thinking

    Kaarel Hänni · 19 May 2026

  • Reformulating Strength and Generality of Instrumental Convergence

    Tushita Jha (TJ) · 20 May 2026 · Slides

  • Value Inversion / Pessimization

    Julia Persson · 21 May 2026 · Slides

  • Flaws of Expected Utility Theory and Ways to Avoid Them

    Ihor Kendiukhov · 21 May 2026 · Slides

  • QACI: Question-Answer Counterfactual Interval

    Noa “Ouro” Hölzer · 21 May 2026