Lila Sciences' Open-Endedness Team releases blog + 4 preprints, flags 'creative paradox'
tw_killian · x · 2026-11-10
Lila Sciences' Open-Endedness Team led by Kenneth Stanley has published its first batch of fundamental AI research: a vision blog plus 4 arXiv preprints outlining a broad agenda for open-endedness.
Key points:
- Identifies a "creative paradox": directly reinforcing creative outputs can cap creative potential, and proposes training methods designed to sidestep it;
- Explores the value of intentionally naive LLMs that see certain domains with fresh eyes;
- Each preprint opens its own novel research direction.
More from Research
- BVSR: a creativity framework inspired by blind variation for AI scientists — lupantech · 2026-11-11
- Professor uses Claude to audit reproducibility of five years of his journal papers — stanfordnlp · 2026-11-11
- Microsoft's SIRIS benchmark: frontier LLMs struggle to judge when simulator flaws invalidate designs — _akhaliq · 2026-11-11
- Geodesic's NovaAtom-Lite beats AlphaFold 3 (79.07% vs 47.90%) on antibody-antigen folding — QuanquanGu · 2026-11-11
- Macrodata's MacroHands hits 13.1mm stereo error and 80% detection under occlusion, beating all open baselines — gui_penedo · 2026-11-11
- Niantic Spatial adds object interaction in Gaussian Splats for embodied AI real2sim training — Scobleizer · 2026-11-11