Research Taste

Research taste is the interview’s term for the judgment that lets a researcher choose problems, run useful experiments, read the field, pivot, and present work coherently. In 133. 对谢赛宁的7小时马拉松访谈：世界模型、逃出硅谷、AMI Labs、两次拒绝Ilya、杨立昆、李飞飞和42, Xie Saining uses Kaiming He as the clearest example.

E242｜最快半年AI跑通自进化？与陈天桥首席科学家聊聊硅谷模型必争之地 turns research taste into a training target for Discovery Model systems. Du Shaolei argues that a scientific model needs to learn from top scientists which questions are fundamental, not merely which answers are fluent or publishable. Li Beibin adds that current models still have weaker taste than ordinary AI scientists, so human experts remain part of the Recursive Self-Improvement loop.

140. 对姚顺宇的4小时访谈：请允许我小疯一下！在Anthropic和Gemini训模型、技术预测、英雄主义已过去 adds a systems version through Yao Shunyu / 姚顺宇. His physics-to-AI path makes research taste less about lone brilliance and more about choosing objective, feedback-rich problems; designing experiments that rule out bugs and false assumptions; and taking responsibility for how local work affects the full training system.

138. 对罗福莉3.5小时访谈：AI范式已然巨变！OpenClaw、Agent范式很吃后训练、卡的分配、组织平权 adds Luo Fuli / 罗福莉’s acceleration version. When Open Claw-style agents can turn ideas into code and evaluations much faster, taste shifts toward selecting which ideas deserve cards, identifying whether a failure is real or infrastructural, and using parallel agents without drowning in shallow experiments.

Key Claims

• Good ideas rarely arrive from sitting still and thinking abstractly; they come from input, exploration, engineering, reading, and abstraction.
• Predicting experiment results before running them makes surprises meaningful: a wrong prediction can be a better signal than a small expected gain.
• Strong baselines and infrastructure matter because weak baselines can make shallow improvements look like research.
• Good research often pivots through difficulty and only later gets narrated as a straight line.
• Taste includes presentation, figures, websites, video, and storytelling, not only the technical trick.
• For scientific AI, taste includes knowing which hypotheses deserve verification and which questions are too shallow to spend scarce compute or expert time on.
• Expert preference data may matter disproportionately in post-training because the signal is about standards and problem choice rather than broad factual coverage.
• In large-scale model work, taste includes knowing when an experiment failed because the idea was wrong versus because the environment, data, token horizon, or implementation was flawed.
• Reliability can be part of research taste when the work affects a shared training system rather than only an individual paper.
• In agent-accelerated research, taste includes compute triage: deciding which generated ideas deserve Training Compute Allocation and which should be discarded quickly.

Connections

• Xie Saining and Kaiming He — source speaker and main exemplar.
• Representation Learning, Self-Supervised Learning, and Diffusion Transformers — research domains where the method is applied.
• Problem Definition In Research — adjacent ability to define what is worth solving.
• FAIR and NYU — institutional contexts where research culture is discussed.
• AI Organization Design — team structure can either preserve or suppress bottom-up research taste.
• Apodex, Du Shaolei, Li Beibin, and Discovery Model — source branch where research taste becomes part of training scientific AI.
• Yao Shunyu / 姚顺宇, Problem Definition In Research, Frontier Model Scaling, and AI Organization Design — systems and reliable-researcher branch added by episode 140.
• Luo Fuli / 罗福莉, Agent Post-Training, Training Compute Allocation, and AI Organization Design — agent-accelerated research branch added by episode 138.