/sci/ - Science & Math

I'm a bit conflicted about consensus. On the one hand, consensus means that the probability of something being true is rather high, so consensus should be the goal.
On the other hand, consensus carries the risk of neglect. The stronger the consensus, the bigger the burden of proof. Take Mandlbaur for example. To convince anyone that angular momentum is not conserved, he'd need to put in a lot of effort to make anyone even listen to him. Ok, he doesn't do that and people ignore him justifiably, but my point is that stronger consensus makes it hard to discover/publish earlier misconceptions.
Is that a good or bad? Should we build on consensus to discover new things at the risk of building upon false premises? Or should we reinvent the wheel continuously, but be sure that everything is correct, at the cost of progress?