· Announcement
Speaker Series Winter 2026/27: meet our eight speakers
From Google DeepMind, Anthropic, Microsoft and the Max Planck Institute, to EPFL, Imperial College London, UCL and Oxford – eight researchers come to Berlin this winter. The first talk is on 3 November, and registration is open now.
The lineup is out! This winter, eight researchers from leading AI labs and universities come to Berlin for the BLISS Speaker Series. The topics range from reasoning in LLMs and AI safety to cognition in brains and AI, sovereign AI and the agentic web.
Updated 27 October: Pavel Izmailov’s talk has moved from 10 November to 2 February 2027. Jonas Geiping’s talk on 17 November takes place at the TU Berlin Mathematics Building, Str. des 17. Juni 136, rather than room C130. Badr AlKhamissi joins us from EPFL. The schedule below reflects these updates.
- When: Tuesdays at 7 pm, starting 3 November 2026
- Where: TU Berlin, room C130, except 17 November at the Mathematics Building (Str. des 17. Juni 136). Check your event registration page for room details.
- Entry: free, registration on Meetup or Luma required
As always, the discussion continues afterwards over free drinks and pizza.
- 3 November: Badr AlKhamissi (EPFL): Human-Like Artificial Intelligence, register on Meetup or Luma
- 17 November: Jonas Geiping (Max Planck Institute for Intelligent Systems): Safety and Security of LLM Reasoning, register on Meetup or Luma
- 1 December: Jonathan Richard Schwarz (Imperial College London): Sovereign AI from a Perspective of Model Building and AI Governance, register on Meetup or Luma
- 15 December: Ida Momennejad (Microsoft Research): Algorithmic Geometry of Flexible Cognition in Brains and AI, register on Meetup or Luma
- 12 January: Robert McCarthy (UCL): The Importance of Model Specs and Constitutions for AI Safety, register on Meetup or Luma
- 26 January: Nenad Tomašev (Google DeepMind): Agentic Web: Opportunities and Challenges, register on Meetup or Luma
- 2 February: Pavel Izmailov (Anthropic): Understanding Reasoning from Pretraining to Post-Training, register on Meetup or Luma
- 9 February: Uljad Berdica (University of Oxford, Google DeepMind): From Clean Slates to Many Worlds: Rethinking RL Algorithm Discovery, register on Meetup or Luma
Seats are limited and the doors close at 7:15 pm, so register early and come on time. All dates are also on the Speaker Series page. See you there!
