4 ms·

Reproducibility seems like a social problem of people publishing shitty science for citations. As for the rate of innovation, discovery is certainly an NP-Hard

• lambdatronics 6y ago
  It's not just 'shitty' science -- it's good ideas that didn't pan out, which also get repackaged as 'breakthroughs.' You can propose a good experiment on a relevant/interesting topic, execute on it well, and still get results that aren't sexy -- that's not failure, it's life. Unfortunately, you can't get published in high-profile journals unless your results are sexy, but you need to do that if you want to keep getting funded. So, instead of just publishing honestly, you find some way to 'spin' your research. For instance, a paper on perovskite solar cells might claim that they have the best-in-class efficiency of any inorganic perovskite cell, but you have to buy the paper to find out that they sacrificed longevity to do it -- it only lasts 20hrs. Meanwhile, mixed organic/inorganic perovskites have even better efficiency AND better lifetimes simultaneously -- but the article can still claim a breakthrough b/c they qualified their claim with "of any inorganic cell." Why is that relevant? It probably isn't.