More part-time and/or amateur scientists would be a good thing. This is more difficult today because there are fewer projects that one person, or even a handful of people can do on their own.

The canonical examples of 'big science' are the humane genome project, particle physics and atmospheric prediction. All three rely on massive international investment in infrastructure, the coordinated contributions of many specialists, and research programs with very long timelines, and where progress is mostly incremental (another bug sequenced, another 0.1 improvement in anomaly correlation, another dB of evidence in favour of some micro-theory).

That's not to say there are no problems left that a genius in a garage can't attack, just that it seems to me they are fewer than back in Lord Kelvin's day, and that the big problems that most of agree we want to solve require massive cooperation: the only effective system we have yet devised for this is via national science agencies.