Heuristic: If you have data and look for questions, you'll generate insight. If you have questions and look for data, you'll generate noise
Conversation
Replying to
Not quite right. Understand the sentiment, but the later sounds like the appropriate scientific testing of hypotheses.
1
1
Replying to
Science does not actually work that way (see Feyerabend, Against Method). Also...
2
1
Replying to
2/2 Science is the craft of "knowing what you know". Like with any craft there are multiple tools, and practitioners of varying talent.
1
Replying to
It sometimes does (minority of cases, especially the automatable ones). That's why I said heuristic, not iron-clad rule.
1
Replying to
Perhaps depends on whether the goal is scientific discovery or technology development.
1
Replying to
That would be nice but the biggest factor is governments treating science production *as* a technology. Goodhart's Law kicks in
Replying to
Too often, big science projects like big medicine or big DOD, become a career, where vested interests keep the project going regardless
3
1
Replying to
historically science breakthroughs have happened in funded but accidentally managerially neglected parts of large institutions
Replying to
Agreed! Agenda-driven "science" (i.e. much of gov't funded science) results large volumes of low quality work.
1
Replying to
Moreover, big science takes place over such long timescales (eg ITER) that no individual held accountable for bad strategic decisions.
1

