Wow, there is a recent “Metascience 101” podcast, a crash course in the debates + issues + ideas of the modern metascience movement in the USA Metascience = science on science: how to manage, improve, optimize scientific process. IFP did many inside talks & now go public. There were 9 episodes, that's summary of Episode 1: Introduction: Who - Caleb Watney, co-founder of Institute for Progress (by Erik Schmidt, to compete with rise of China) - 40 ppl think tank on innovation policy in Washington, lobby simplified high skill immigration + deregulate infra construction + metascience - Dylan Matthews, journalist at VOX - Heidi Williams, economist/NBER Science is important - U.S. federal gov is the world’s largest funder of basic research, thus impacts entire humankind. National Science Foundation (NSF) grants $10 bln/year (and 3x more by uniiversities) + National Institute of Health (NIH) for applied biomedical R&D $45 bln/year - Private sector also, but they under-provide research relative to the level that we might want as a society, thus we try to incentivize by R&D tax credit, by patents (yet lots of basic research isn't even patentable). Intricate scheme - We should spend a lot more Pharma - If a drug discovered today, published in Nature, it can be 12-16 years before sales - trials to show safety & efficacy. Statistical tests show if we had a tech with shorter clinical trials, we would get more innovation (e.g. cancer drugs will save more lives) From PhD to independence - We're lucky that lots of foreign students want to come to U.S. universities/seen as a premier research environment. After PhD you become postdoc (often complicated for internationals cause limited for certain types), apply for grants for your lab (may take 6 months for NSF, 18 months for NIH) so you plan a lot/salaries of ppl in your lab - Lengthening of trainings - You fundraise your salary in addition to research costs, so people look what funders are interested in, and not on own best ideas - Historically, scientists were funded by patronage so they could work on own directions. For project-based grants it can be very hard to pivot research from one way to another Team science - Ben Jones: structures we use to fund science are static last 100 years, despite rise of team science (pic 1900-2010), one of the most important changes in our lifetime, Einstein solo couple of summers vs. Large Hadron Collider - Nobel Prize is a very individual recognition, how we evaluate people's work is not for team science, such disconnect Immigration - USA 4% of the global population. To be a scientific superpower we need maximize agglomeration effects - a bunch of really smart ppl together are more productive, enormous returns - Bohr/Denmark, Rutherford/New Zealand etc. Solutions to improve funding - Scientists submit proposals for promising avenues, create a budget (rough), then a panel of peers grade across a number of dimensions: promising? social impact? probability? enough ppl's track record? So, the average opinion of peers for grant funding agencies - Consensus-oriented review processes might bias against high-risk, high-reward research Experiments - Hughes Medical Institute or Arc Institute - person-based funding, lots of freedom - ’Fast grant’-style programs, like Patrick Collison and Tyler Cowen did with two-week/two-day turnarounds in the pandemics - Like DARPA (Defense Advanced Research Projects Agency) - now ARPA-E for energy, ARPA-H for health. To give a really wide scope & autonomy to particular program managers who can then try to push grants or technologists & engineers to work on a very specific problem. Portfolio approach Doing metascience - In 10 years via systematic studies to build evidence and get metrics, iteratively, for different funding mechanisms. Learn what works, address specific problems, A/B tests like in biz - Basically, apply core scientific approach to the institutions that fund and incentivize science