The government has purposely set up the process in an adversarial manner (you must prove why you uniquely are positioned to make scientific advances here, bette

The government has purposely set up the process in an adversarial manner (you must prove why you uniquely are positioned to make scientific advances here, better than anyone else), so it's not surprising that the process plays out in an adversarial manner. The government also encourages institutions to be entrepreneurial by doing things like patenting their results, or keeping them as trade secrets to hand to spinoff companies, so academia can prove its worth in the startup economy. Is this the best way to make scientific advances? Probably not. But it's a systemic problem. There is at least one study (I believe Canadian, can't find a link currently) which found that it would be more efficient just to give every researcher in Canada some baseline funding, rather than spending all the money they currently do on the adversarial funding process. That was actually somewhat closer to the traditional CS funding model in the U.S., though largely accidentally, through military funding. Why did someone like Richard Stallman (not an academic) get employed for decades in the MIT AI Lab? Because DARPA gave non-micromanaged block grants to the MIT AI Lab to fund a "center of excellence" for years at a time, which they were free to use to employ staff as they saw fit. Nowadays most grants are 3 years long and very specific, so it's all about positioning yourself for the next 3-year grant, if you want to keep the lights on and your staff paid.