AI scientists need a new research stack, not just papers written for humans
AnneliesGamble · x · 2026-07-22
This article argues that scientific publishing was built for humans, not AI researchers. It frames the paper as both a storytelling device and an engineering bottleneck, then uses Amber Liu’s paper, The Last Human-Written Paper: Agent-Native Research Artifacts, to argue for new research infrastructure that lets AI scientists accumulate knowledge and collaborate effectively.
Core argument
- The modern paper system emerged in the 1660s to make discoveries public, citable, and scrutinizable.
- But the PDF/arXiv/conference stack is optimized for human reading and human review.
- That creates two costs:
- Storytelling tax: polished papers hide the failed hypotheses, pivots, and tedious experimentation.
- Engineering tax: papers omit the exact setup details needed to reproduce the work.
Why it matters for AI scientists
The piece says AI agents will struggle to compound knowledge or coordinate with each other unless the research stack is redesigned for machine-speed work, not just human speed.
Related event: Research Paradigm Shift: Academic Papers Inadequate for AI Scientists(3 posts)→
More from AGI Musings
- François Fleuret: Only Two Long-Term Futures — No Super AI, or Staying Fully Human With It — francoisfleuret · 2026-09-11
- IG reel debunking the 'winning the AI race against China' fallacy hits 500k likes — louisvarge · 2026-09-11
- Post-AI World Leaves No Room for Learning on the Job — rachittshah · 2026-09-11
- Researcher questions AI safety eval firm, citing 'blatantly sloppy' security and monitoring — Kyrannio · 2026-09-11
- AI researcher memes agent-swarm tinkering with He Jiankui's embryo-editing quote — dejavucoder · 2026-09-11
- nabla_theta: happy to be wrong if the AI utopia arrives with little ex ante risk — nabla_theta · 2026-09-11