Hacker News new | past | comments | ask | show | jobs | submit login

> the base assumption simply has to be that the field as a whole moves forward. Otherwise it becomes very difficult - maybe even impossible - to have any sort of evidence-based discussion.

It seems like this "assumption" needs to be qualified with time scales for disambiguating trends from "facts" .. otherwise this is equivalent to saying that academic consensus can never be wrong




> otherwise this is equivalent to saying that academic consensus can never be wrong

No it is not. I said that the assumption must be that the consensus becomes less wrong over time, not that it is never wrong. Those are two very different statements.


Over how much time? Does "consensus" never make a wrong turn?


Well, academic consensus usually takes a long time to build. And obviously it can take wrong turns (and there are plenty of examples of that historically, often due to work based on incomplete or wrong data - though arguably even if the academic consensus has been wrong on an absolute basis, it has often been "correct" in light of the available evidence).

But nevertheless, it should be obvious today that (probabilistically) the best understanding of the world - for laymen - is achieved by trusting the academic consensus. Anything else quickly veers into anti-intellectual, conspirationist mumbo-jumbo.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: