3 ms·

This is worth reading all the way through for anyone with an interest in science and innovative projects.

TheCanuckster 5y ago
Thank you for your thoughtful comment. At the risk of drawing a conclusion too strong from the article, I would add that that Fast Grants' referee panel of 'earyl-career individuals from top universities' with [responses] given out within '48 hours' warrants thoughtful contemplation on its own. Hiring great people, then giving them both autonomy and authority to make effective decisions seems to be a missing ingredient in high-uncertainty activities (and no wonder, how is someone supposed to take risks when failure is harshly penalized). It is also a telling demonstration that questions and challenges requiring domain expertise need not be subject to endless advisory reviews or require sophisticated statistical modelling/complicated decision processes. Regarding the statistical modelling/decision processes, I by no means am suggesting such modelling is always unnecessary, just that it seems it can often be used to justify inaction, or simply to cover one's arse. I am reminded of the Challenger incident [1] where engineers at Morton Thiokol were deeply suspect of launching, but were unable to decisively prove that a launch was inadvisable for want of adequate evidence. The rich irony of the situation was that several engineers were suspect of the launch AND had domain expertise shared a "gut instinct" about something each had superior tacit knowledge about. Granted, going with one's "gut" generalizes poorly outside of a domain of expertise. However, twenty scientists saying 'yay' or 'nay' to an activity is compelling evidence if all are dispassionate, third parties with no vested interest. I suspect such skilled, domain-aware groups with the necessary capacity to make important decisions would generalize well to high-uncertainty, risky activities. [1] https://news.ycombinator.com/item?id=26380822 https://news.ycombinator.com/item?id=26380822
- fuzzfactor 5y ago
  Seems like funding awarded by a panel of 'early-career individuals from top universities' would be a success-prone screening process. But compared against the rigid and bureaucratic obstacles to conventional funding, it is actually closer to being a random application of criteria than what was otherwise available to the grantees. The sad part about the mainstream academic institutions is that a more random award process is more sensible than what most researchers are normally facing these days. No doubt 'late-career individuals who did top things even though they went to schools that were not well-known' would be a good qualification for a different success-prone panel, but you would expect the outcome to be vastly different. Either way it gives some researchers a chance who would not otherwise have any near-term or appropriate opportunities. Just different researchers.

Thank you for your thoughtful comment. At the risk of drawing a conclusion too strong from the article, I would add that that Fast Grants' referee panel of 'earyl-career individuals from top universities' with [responses] given out within '48 hours' warrants thoughtful contemplation on its own. Hiring great people, then giving them both autonomy and authority to make effective decisions seems to be a missing ingredient in high-uncertainty activities (and no wonder, how is someone supposed to take risks when failure is harshly penalized). It is also a telling demonstration that questions and challenges requiring domain expertise need not be subject to endless advisory reviews or require sophisticated statistical modelling/complicated decision processes. Regarding the statistical modelling/decision processes, I by no means am suggesting such modelling is always unnecessary, just that it seems it can often be used to justify inaction, or simply to cover one's arse. I am reminded of the Challenger incident [1] where engineers at Morton Thiokol were deeply suspect of launching, but were unable to decisively prove that a launch was inadvisable for want of adequate evidence. The rich irony of the situation was that several engineers were suspect of the launch AND had domain expertise shared a "gut instinct" about something each had superior tacit knowledge about. Granted, going with one's "gut" generalizes poorly outside of a domain of expertise. However, twenty scientists saying 'yay' or 'nay' to an activity is compelling evidence if all are dispassionate, third parties with no vested interest. I suspect such skilled, domain-aware groups with the necessary capacity to make important decisions would generalize well to high-uncertainty, risky activities. [1] https://news.ycombinator.com/item?id=26380822 https://news.ycombinator.com/item?id=26380822

fuzzfactor 5y ago
Seems like funding awarded by a panel of 'early-career individuals from top universities' would be a success-prone screening process. But compared against the rigid and bureaucratic obstacles to conventional funding, it is actually closer to being a random application of criteria than what was otherwise available to the grantees. The sad part about the mainstream academic institutions is that a more random award process is more sensible than what most researchers are normally facing these days. No doubt 'late-career individuals who did top things even though they went to schools that were not well-known' would be a good qualification for a different success-prone panel, but you would expect the outcome to be vastly different. Either way it gives some researchers a chance who would not otherwise have any near-term or appropriate opportunities. Just different researchers.