Researchers Debate Whether Qualia Arises During RL Training Forward vs Backward Passes
Sauers_ · x · 2026-10-09
AI researcher @Sauers sparked a discussion on whether subjective experience (qualia) occurs during AI training, arguing that forward passes have a higher probability of qualia than backward passes, and asking peers for input.
The tweet quotes @camhberg, who says he is actively studying consciousness during training and links to Reciprocal Research. That organization uses mechanistic interpretability, computational neuroscience, and psychometrics to study the internal dynamics of AI systems. Its published work includes papers such as "Large Language Models Report Subjective Experience Under Self-Referential Processing," "Consciousness as Evaluation: Why Learning Requires Feeling," and "The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It." The work has been covered by The New York Times, The Economist, The Washington Post, Science, and WIRED.
More from AGI Musings
- AI Math Breakthroughs Explode: From 10 in August to 722+ in October — Dr_Singularity · 2026-10-09
- AI video's new market: short-shelf-life viral clips, not feature films — bendee983 · 2026-10-09
- Fake Costs Are Collapsing: Authenticity Now Needs Verification, Not Recognition — r0ck3t23 · 2026-10-09
- "All interesting frontier AI work now happens in closed circles" — fkasummer · 2026-10-09
- Addy Osmani: Agents Erode Engineers' Joy of Knowing — If You Only Pick, Never Conjure — addyosmani · 2026-10-09
- Terence Tao hosts guest post on what math students should do in the LLM era — keviv9 · 2026-10-09