Reading bad ML papers? Three philosophy-of-science classics to fix your thinking
_lewtun · x · 2026-09-04
After one too many abhorrently bad ML papers, researcher wfithian says he was "one shot" by the philosophy of science and recommends three pieces:
- Strong Inference (Platt): why some fields advance much faster than others via competing-hypothesis testing.
- Could a Neuroscientist Understand a Microprocessor? (Jonas & Kording): applying neuroscience methods to a CPU shows methods can be futile.
- Popper, Bayes and the Inverse Problem (Tarantola).
HF's Lewis Tunstall replied recommending Chalmers' textbook What Is This Thing Called Science?
More from Research
- Researcher: the best papers answer nothing and raise more questions — yacinelearning · 2026-09-04
- IBM releases DRACO: dynamic rubrics give per-step credit assignment for verifier-free agent RL — ibm-research · 2026-09-04
- The Last Translation Benchmark debuts with multimodal examples that break leading translation models — Vilém Zouhar · 2026-09-04
- Reed-Solomon List Decoding Breakthrough Resolves Major Open Problem in Coding Theory — AlexKontorovich · 2026-09-04
- AI autoresearchers help crack 30-year-old coding theory problem, soundness up to 68.02 bits — BenBlaiszik · 2026-09-04
- Redditor fine-tunes SDXL on 60 childhood photos to simulate memory recall — uisato · 2026-09-04