Ask HN: When human experts disagree, how should machines determine truth?

This question matters as fully automated or AI-assisted scientific loops become common.

For example, MDA-MB-435 was a cell line used to study breast cancer for decades, but its origin was disputed for years and eventually confirmed as melanoma.

Another open question, one layer above experiments: what is the greatest barrier to curing cancer?

Is it disease understanding? [0]

Or regulatory reform of clinical trials? [1]

Or clinical trial scarcity? [2]

Or science upstream of clinical trials, and other factors? [3, 4]

Many experts disagree.

If machines allocated resources and set priorities, how should they decide what is right?

If you want to dive into the challenges of curing cancer, I recommend reading all five sources, starting with [1]. It's a New York Times essay with insightful quotes from leading oncologists and vivid patient stories. Beautifully written, the piece captures why regulatory reform is critical and long overdue.

More importantly, experts are actively debating the essay on X and may answer your questions.

I intertwined ML with cancer research because both share surprising parallels. A simple one is that many ideas sound compelling in theory but die in reality. In cancer, findings may not translate to patients while ML methods may not transfer across models.

Cancer is a systemic problem, and HN attracts some of the best systems thinkers.

Hope this post sparks curiosity.

---

0. https://deepphenotype.substack.com/p/drug-discovery-has-no-magic-wands

1. https://www.nytimes.com/2026/09/04/opinion/clinical-trials-drugs-science.html

2. https://x.com/davidycli/status/2096315300841214423

3. https://www.drvinayprasad.com/p/why-dont-we-have-more-cancer-cures

4. https://x.com/LocasaleLab/status/2096353702684807416